Compare commits

...
Author SHA1 Message Date
autofix-ci[bot]andGitHub 756ce8a10e [autofix.ci] apply automated fixes 2026-07-28 05:33:30 +00:00
-LAN- e07de297bf fix(workflow): migrate sys files to user input 2026-07-28 13:29:09 +08:00
yyhandGitHub 1e5e47b889 fix: validate plugin installation scope (#39669) 2026-07-28 04:13:20 +00:00
Asuka MinatoGitHubautofix-ci[bot] <114827586+autofix-ci[bot]@users.noreply.github.com>Byron.wang
dd28b0d165 test: move OAuth server service coverage to unit tests (#38931)
Co-authored-by: autofix-ci[bot] <114827586+autofix-ci[bot]@users.noreply.github.com>
Co-authored-by: Byron.wang <byron@dify.ai>
2026-07-28 04:09:32 +00:00
JoelandGitHub 49e74e2f58 fix: prevent the model selector footer from covering options (#39668) 2026-07-28 04:02:28 +00:00
e6e5d761c2 fix: prevent Safari from clipping the settings close button (#39664)
Co-authored-by: yyh <92089059+lyzno1@users.noreply.github.com>
2026-07-28 04:01:12 +00:00
Asuka MinatoandGitHub e1ce808567 test: move data source controller coverage to unit tests (#38924) 2026-07-28 03:51:22 +00:00
Asuka MinatoandGitHub 3c7ad816d9 test: use SQLite sessions in commands (#39080) 2026-07-28 03:29:02 +00:00
Asuka MinatoandGitHub f44dd343da test: use sqlite3 session in test_plugin_service (#38727) 2026-07-28 03:27:58 +00:00
Asuka MinatoandGitHub 2d9b2d50f3 test: use sqlite3 session in test_wraps (#38770) 2026-07-28 03:27:22 +00:00
Asuka MinatoandGitHub 003e0f9614 test: move message cleanup coverage to unit tests (#38932) 2026-07-28 03:26:43 +00:00
Asuka MinatoGitHubautofix-ci[bot] <114827586+autofix-ci[bot]@users.noreply.github.com>Byron.wang
da0979b373 test: use SQLite sessions in services core (#39090)
Co-authored-by: autofix-ci[bot] <114827586+autofix-ci[bot]@users.noreply.github.com>
Co-authored-by: Byron.wang <byron@dify.ai>
2026-07-28 03:20:27 +00:00
Asuka MinatoGitHubautofix-ci[bot] <114827586+autofix-ci[bot]@users.noreply.github.com>Byron.wang
81fb57639e test: use SQLite sessions in controllers service api (#39098)
Co-authored-by: autofix-ci[bot] <114827586+autofix-ci[bot]@users.noreply.github.com>
Co-authored-by: Byron.wang <byron@dify.ai>
2026-07-28 03:19:17 +00:00
Asuka MinatoandGitHub 1e61078e93 test: use SQLite sessions in services core (#39088) 2026-07-28 03:18:32 +00:00
Escape0707andGitHub b597bb1b17 test: separate human input unit and database paths (#39655) 2026-07-28 03:18:01 +00:00
Asuka MinatoGitHubautofix-ci[bot] <114827586+autofix-ci[bot]@users.noreply.github.com>Byron.wang
65ead05dfc test: use SQLite sessions in services plugin (#39084)
Co-authored-by: autofix-ci[bot] <114827586+autofix-ci[bot]@users.noreply.github.com>
Co-authored-by: Byron.wang <byron@dify.ai>
2026-07-28 03:16:19 +00:00
6f8ed69ee1 fix: fix mcp output_schema is optional (#39453)
Co-authored-by: yunlu.wen <yunlu.wen@dify.ai>
2026-07-28 02:31:18 +00:00
Yunlu WenGitHubautofix-ci[bot] <114827586+autofix-ci[bot]@users.noreply.github.com>
59b879d2df chore: bump version to 1.16.1 (#39653)
Co-authored-by: autofix-ci[bot] <114827586+autofix-ci[bot]@users.noreply.github.com>
2026-07-28 01:09:25 +00:00
Xiyuan ChenandGitHub a57b0b9b58 feat(plugin): allow disabling the tenant plugin model providers cache (#39632) 2026-07-27 23:53:59 +00:00
Xiyuan ChenandGitHub d8506efed6 fix(cli): decouple release script tests from the live compat window (#39658) 2026-07-27 23:50:05 +00:00
JingyiandGitHub b25b28cc76 fix(workflow): align block icon vector sizes (#39657) 2026-07-27 23:43:03 +00:00
yyhGitHubautofix-ci[bot] <114827586+autofix-ci[bot]@users.noreply.github.com>
f68cabe226 fix(web): align tour trigger DOM order (#39654)
Co-authored-by: autofix-ci[bot] <114827586+autofix-ci[bot]@users.noreply.github.com>
2026-07-27 13:00:16 +00:00
Asuka MinatoGitHubautofix-ci[bot] <114827586+autofix-ci[bot]@users.noreply.github.com>Byron.wang
f73f83c5d6 test: use SQLite sessions in core app (#39075)
Co-authored-by: autofix-ci[bot] <114827586+autofix-ci[bot]@users.noreply.github.com>
Co-authored-by: Byron.wang <byron@dify.ai>
2026-07-27 10:45:16 +00:00
Asuka MinatoGitHubautofix-ci[bot] <114827586+autofix-ci[bot]@users.noreply.github.com>Byron.wang
29801d6a65 test: use SQLite sessions in core tools (#39077)
Co-authored-by: autofix-ci[bot] <114827586+autofix-ci[bot]@users.noreply.github.com>
Co-authored-by: Byron.wang <byron@dify.ai>
2026-07-27 10:45:11 +00:00
Asuka MinatoGitHubautofix-ci[bot] <114827586+autofix-ci[bot]@users.noreply.github.com>Byron.wang
75016e8bfe test: use SQLite sessions in core ops (#39072)
Co-authored-by: autofix-ci[bot] <114827586+autofix-ci[bot]@users.noreply.github.com>
Co-authored-by: Byron.wang <byron@dify.ai>
2026-07-27 10:41:17 +00:00
yyhandGitHub b6dea7b2ba fix(workflow): preserve latest collaboration session (#39646) 2026-07-27 10:37:45 +00:00
yyhandGitHub b2b1cd7e97 chore: update workspace dependencies (#39641) 2026-07-27 10:27:31 +00:00
Asuka MinatoGitHubautofix-ci[bot] <114827586+autofix-ci[bot]@users.noreply.github.com>Byron.wang
ca63e01d45 test: use SQLite sessions in controllers service api (#39069)
Co-authored-by: autofix-ci[bot] <114827586+autofix-ci[bot]@users.noreply.github.com>
Co-authored-by: Byron.wang <byron@dify.ai>
2026-07-27 10:26:23 +00:00
yyhandGitHub 271b6e1f5c fix(ui): unify combobox trigger focus rings (#39643) 2026-07-27 10:21:32 +00:00
Asuka MinatoGitHubautofix-ci[bot] <114827586+autofix-ci[bot]@users.noreply.github.com>Byron.wang
609ddf9e01 test: use SQLite sessions in core memory (#39066)
Co-authored-by: autofix-ci[bot] <114827586+autofix-ci[bot]@users.noreply.github.com>
Co-authored-by: Byron.wang <byron@dify.ai>
2026-07-27 10:15:08 +00:00
Asuka MinatoGitHubautofix-ci[bot] <114827586+autofix-ci[bot]@users.noreply.github.com>Byron.wang
9fd1fea58c test: use SQLite sessions in services core (#39057)
Co-authored-by: autofix-ci[bot] <114827586+autofix-ci[bot]@users.noreply.github.com>
Co-authored-by: Byron.wang <byron@dify.ai>
2026-07-27 09:36:44 +00:00
JoelandGitHub 8de3b4d033 fix: prevent the agent build draft UI from flashing during apply (#39642) 2026-07-27 09:32:52 +00:00
dependabot[bot]GitHubdependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
d9c038daf2 chore(deps): bump the storage group across 1 directory with 3 updates (#39604)
Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
2026-07-27 09:24:58 +00:00
Asuka MinatoGitHubautofix-ci[bot] <114827586+autofix-ci[bot]@users.noreply.github.com>Byron.wang
c9c057fd04 test: use SQLite sessions in rag retrieval (#39049)
Co-authored-by: autofix-ci[bot] <114827586+autofix-ci[bot]@users.noreply.github.com>
Co-authored-by: Byron.wang <byron@dify.ai>
2026-07-27 09:12:30 +00:00
dependabot[bot]GitHubdependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
85189be53d chore: bump pymdown-extensions from 10.21.2 to 11.0 in /dify-agent (#39618)
Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
2026-07-27 09:10:22 +00:00
Asuka MinatoGitHubautofix-ci[bot] <114827586+autofix-ci[bot]@users.noreply.github.com>
01e736aaf7 test: use sqlite3 session in test_extract_thread_messages (#38739)
Co-authored-by: autofix-ci[bot] <114827586+autofix-ci[bot]@users.noreply.github.com>
2026-07-27 09:00:47 +00:00
dependabot[bot]GitHubdependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
a28969f564 chore(deps): bump golang.org/x/net from 0.53.0 to 0.55.0 in /dify-agent-runtime (#39025)
Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
2026-07-27 08:58:40 +00:00
Asuka MinatoandGitHub 42f9610ba5 test: use sqlite3 session in test_api (#38733) 2026-07-27 08:57:51 +00:00
dependabot[bot]GitHubdependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
1a7ffbe5ee chore: bump gitpython from 3.1.52 to 3.1.54 in /api (#39619)
Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
2026-07-27 08:57:44 +00:00
JoelandGitHub 481354ca70 refactor: make agent composer save and publish state consistent (#39637) 2026-07-27 08:43:45 +00:00
3c80857ea3 feat: add bearer auth to agent backend (#39622)
Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>
2026-07-27 08:37:30 +00:00
Xiyuan ChenandGitHub 94702efbc2 fix: gate service API, MCP and trigger surfaces on enterprise license (#39635) 2026-07-27 08:23:15 +00:00
Asuka MinatoandGitHub 755f7b0e8b test: use sqlite3 session in test_workflow_comment_service (#38700) 2026-07-27 08:22:45 +00:00
Asuka MinatoandGitHub a1c9564b30 test: use sqlite3 session in test_llm_generator_missing (#38742) 2026-07-27 08:22:16 +00:00
Asuka MinatoandGitHub 97acd9c70b test: use sqlite3 session in test_utils (#38729) 2026-07-27 08:21:39 +00:00
Asuka MinatoandGitHub 71601bf76c test: use sqlite3 session in test_draft_var_loader_simple (#38702) 2026-07-27 08:20:18 +00:00
Asuka MinatoandGitHub 460efbf285 test: use sqlite3 session in test_workflow_run_service (#38701) 2026-07-27 08:19:32 +00:00
a752f43b8e test: use pristine file-backed SQLite fixtures (#39624)
Co-authored-by: Asuka Minato <i@asukaminato.eu.org>
2026-07-27 08:07:05 +00:00
非法操作GitHubautofix-ci[bot] <114827586+autofix-ci[bot]@users.noreply.github.com>
b3774bfe1c perf: add lightweight recent apps endpoint (#39625)
Co-authored-by: autofix-ci[bot] <114827586+autofix-ci[bot]@users.noreply.github.com>
2026-07-27 08:00:39 +00:00
yyhandGitHub 4313d23889 perf(web): lazy-load Home creation modals (#39634) 2026-07-27 07:48:35 +00:00
林玮 (Jade Lin)andGitHub 26a43db6a4 fix(api): wait for workflow worker cleanup before task completion (#39614) 2026-07-27 07:34:25 +00:00
yyhandGitHub e5d40336b3 fix(workflow): lazy-load Loro collaboration runtime (#39631) 2026-07-27 07:15:13 +00:00
237 changed files with 12155 additions and 8788 deletions
+3
View File
@@ -666,6 +666,7 @@ PLUGIN_REMOTE_INSTALL_PORT=5003
PLUGIN_REMOTE_INSTALL_HOST=localhost
PLUGIN_MAX_PACKAGE_SIZE=15728640
PLUGIN_MODEL_SCHEMA_CACHE_TTL=3600
PLUGIN_MODEL_PROVIDERS_CACHE_ENABLED=true
PLUGIN_MODEL_PROVIDERS_CACHE_TTL=86400
# Comma-separated marketplace plugin IDs whose latest versions are installed for newly registered users.
# Example: langgenius/openai,langgenius/gemini
@@ -677,6 +678,8 @@ INNER_API_KEY_FOR_PLUGIN=QaHbTe77CtuXmsfyhR7+vRjI/+XbV1AaFy691iy+kGDv2Jvy0/eAh8Y
# Dify Agent backend
AGENT_BACKEND_BASE_URL=http://localhost:5050
# Bearer token sent to the Agent backend /runs API. Must match DIFY_AGENT_API_TOKEN on the server side.
AGENT_BACKEND_API_TOKEN=dify-agent-run-token-for-dev-only
AGENT_BACKEND_STREAM_READ_TIMEOUT_SECONDS=30
AGENT_BACKEND_STREAM_MAX_RECONNECTS=3
AGENT_BACKEND_RUN_TIMEOUT_SECONDS=1200
+59 -30
View File
@@ -1,10 +1,13 @@
import logging
import time
from collections.abc import Callable
from typing import NamedTuple
import socketio
from flask import request
from opentelemetry.trace import get_current_span
from opentelemetry.trace.span import INVALID_SPAN_ID, INVALID_TRACE_ID
from werkzeug.exceptions import Forbidden, HTTPException, ServiceUnavailable
from configs import dify_config
from contexts.wrapper import RecyclableContextVar
@@ -42,6 +45,53 @@ _CONSOLE_EXEMPT_PREFIXES = (
"/console/api/activate/check",
)
_WEBAPP_EXEMPT_PREFIXES = ("/api/system-features",)
_INVALID_LICENSE_STATUSES = (LicenseStatus.INACTIVE, LicenseStatus.EXPIRED, LicenseStatus.LOST)
def _session_surface_error(license_status: LicenseStatus | None) -> HTTPException:
if license_status is None:
return UnauthorizedAndForceLogout("Unable to verify enterprise license. Please contact your administrator.")
return UnauthorizedAndForceLogout(f"Enterprise license is {license_status}. Please contact your administrator.")
def _bearer_surface_error(license_status: LicenseStatus | None) -> HTTPException:
"""Token-authed: forcing a logout is meaningless and license state must not leak."""
return Forbidden(description="license_required")
def _retryable_surface_error(license_status: LicenseStatus | None) -> HTTPException:
"""Webhook senders retry on 5xx but treat 4xx as permanent, disabling the subscription."""
return ServiceUnavailable(description="license_required")
class _LicenseGatedSurface(NamedTuple):
prefix: str
exempt_prefixes: tuple[str, ...]
build_error: Callable[[LicenseStatus | None], HTTPException]
# /files (plugin-daemon data plane), /inner/api (enterprise control plane) and /health
# stay ungated: blocking them breaks workflow execution or license recovery itself.
_LICENSE_GATED_SURFACES = (
_LicenseGatedSurface("/console/api/", _CONSOLE_EXEMPT_PREFIXES, _session_surface_error),
_LicenseGatedSurface("/api/", _WEBAPP_EXEMPT_PREFIXES, _session_surface_error),
_LicenseGatedSurface("/v1", (), _bearer_surface_error),
_LicenseGatedSurface("/mcp", (), _bearer_surface_error),
_LicenseGatedSurface("/triggers", (), _retryable_surface_error),
)
def _match_license_gated_surface(path: str) -> _LicenseGatedSurface | None:
for surface in _LICENSE_GATED_SURFACES:
if not path.startswith(surface.prefix):
continue
if any(path.startswith(exempt) for exempt in surface.exempt_prefixes):
return None
return surface
return None
# ----------------------------
# Application Factory Function
@@ -62,38 +112,17 @@ def create_flask_app_with_configs() -> DifyApp:
init_request_context()
RecyclableContextVar.increment_thread_recycles()
# Enterprise license validation for API endpoints (both console and webapp)
# When license expires, block all API access except bootstrap endpoints needed
# for the frontend to load the license expiration page without infinite reloads.
if dify_config.ENTERPRISE_ENABLED:
is_console_api = request.path.startswith("/console/api/")
is_webapp_api = request.path.startswith("/api/")
surface = _match_license_gated_surface(request.path)
if surface is not None:
try:
license_status = EnterpriseService.get_cached_license_status()
except Exception:
logger.exception("Failed to check enterprise license status")
license_status = None
if is_console_api or is_webapp_api:
if is_console_api:
is_exempt = any(request.path.startswith(p) for p in _CONSOLE_EXEMPT_PREFIXES)
else: # webapp API
is_exempt = request.path.startswith("/api/system-features")
if not is_exempt:
try:
# Check license status (cached — see EnterpriseService for TTL details)
license_status = EnterpriseService.get_cached_license_status()
if license_status in (LicenseStatus.INACTIVE, LicenseStatus.EXPIRED, LicenseStatus.LOST):
raise UnauthorizedAndForceLogout(
f"Enterprise license is {license_status}. Please contact your administrator."
)
if license_status is None:
raise UnauthorizedAndForceLogout(
"Unable to verify enterprise license. Please contact your administrator."
)
except UnauthorizedAndForceLogout:
raise
except Exception:
logger.exception("Failed to check enterprise license status")
raise UnauthorizedAndForceLogout(
"Unable to verify enterprise license. Please contact your administrator."
)
if license_status is None or license_status in _INVALID_LICENSE_STATUSES:
raise surface.build_error(license_status)
# add after request hook for injecting trace headers from OpenTelemetry span context
# Only adds headers when OTEL is enabled and has valid context
+5 -1
View File
@@ -11,6 +11,7 @@ from clients.agent_backend.fake_client import FakeAgentBackendRunClient, FakeAge
def create_agent_backend_run_client(
*,
base_url: str | None = None,
api_token: str | None = None,
use_fake: bool = False,
fake_scenario: str | FakeAgentBackendScenario = FakeAgentBackendScenario.SUCCESS,
stream_read_timeout_seconds: float = 30,
@@ -22,8 +23,11 @@ def create_agent_backend_run_client(
return FakeAgentBackendRunClient(scenario=FakeAgentBackendScenario(fake_scenario))
if base_url is None:
raise ValueError("base_url is required when creating a real Agent backend client")
headers: dict[str, str] = {}
if api_token:
headers["Authorization"] = f"Bearer {api_token}"
return DifyAgentBackendRunClient(
Client(base_url=base_url, stream_timeout=stream_read_timeout_seconds),
Client(base_url=base_url, stream_timeout=stream_read_timeout_seconds, headers=headers),
stream_max_reconnects=stream_max_reconnects,
stream_timeout_seconds=stream_run_timeout_seconds,
)
+5 -6
View File
@@ -3,6 +3,7 @@ CLI command modules extracted from `commands.py`.
"""
from .account import create_tenant, reset_email, reset_password
from .app_maintenance import convert_to_agent_apps, fix_app_site_missing
from .data_migrate import data_migrate, legacy_model_types
from .data_migration import (
export_migration_data,
@@ -10,6 +11,7 @@ from .data_migration import (
import_migration_data,
migration_data_wizard,
)
from .database import upgrade_db
from .plugin import (
backfill_plugin_auto_upgrade,
extract_plugins,
@@ -36,12 +38,6 @@ from .retention import (
restore_workflow_runs,
)
from .storage import clear_orphaned_file_records, file_usage, migrate_oss, remove_orphaned_files_on_storage
from .system import (
convert_to_agent_apps,
fix_app_site_missing,
reset_encrypt_key_pair,
upgrade_db,
)
from .vector import (
add_qdrant_index,
migrate_annotation_vector_database,
@@ -49,6 +45,8 @@ from .vector import (
old_metadata_migration,
vdb_migrate,
)
from .workflow_migration import migrate_legacy_sys_files_workflows
from .workspace import reset_encrypt_key_pair
__all__ = [
"add_qdrant_index",
@@ -80,6 +78,7 @@ __all__ = [
"migrate_data_for_plugin",
"migrate_dataset_permissions_to_rbac",
"migrate_knowledge_vector_database",
"migrate_legacy_sys_files_workflows",
"migrate_member_roles_to_rbac",
"migrate_oss",
"migration_data_wizard",
@@ -1,86 +1,27 @@
"""App data maintenance CLI commands."""
import logging
import click
import sqlalchemy as sa
from sqlalchemy import delete, select, update
from sqlalchemy.orm import sessionmaker
from sqlalchemy import select, update
from configs import dify_config
from enums.deployment_edition import DeploymentEdition
from events.app_event import app_was_created
from extensions.ext_database import db
from extensions.ext_redis import redis_client
from libs.db_migration_lock import DbMigrationAutoRenewLock
from libs.rsa import generate_key_pair
from models import Tenant
from models.model import App, AppMode, Conversation
from models.provider import Provider, ProviderModel
from models.tools import ApiToolProvider, BuiltinToolProvider, MCPToolProvider
logger = logging.getLogger(__name__)
DB_UPGRADE_LOCK_TTL_SECONDS = 60
@click.command(
"reset-encrypt-key-pair",
help="Reset the asymmetric key pair of workspace for encrypt LLM credentials. "
"After the reset, all LLM credentials and tool provider credentials "
"(builtin / API / MCP) will be purged, requiring re-entry. "
"Only support SELF_HOSTED mode.",
)
@click.confirmation_option(
prompt=click.style(
"Are you sure you want to reset encrypt key pair? "
"This will also purge builtin / API / MCP tool provider records for every tenant. "
"This operation cannot be rolled back!",
fg="red",
)
)
def reset_encrypt_key_pair():
"""
Reset the encrypted key pair of workspace for encrypt LLM credentials.
After the reset, all LLM credentials will become invalid, requiring re-entry.
Only support SELF_HOSTED mode.
"""
if dify_config.DEPLOYMENT_EDITION == DeploymentEdition.CLOUD:
click.echo(click.style("This command is only for SELF_HOSTED installations.", fg="red"))
return
with sessionmaker(db.engine, expire_on_commit=False).begin() as session:
tenants = session.scalars(select(Tenant)).all()
for tenant in tenants:
if not tenant:
click.echo(click.style("No workspaces found. Run /install first.", fg="red"))
return
tenant.encrypt_public_key = generate_key_pair(tenant.id)
session.execute(delete(Provider).where(Provider.provider_type == "custom", Provider.tenant_id == tenant.id))
session.execute(delete(ProviderModel).where(ProviderModel.tenant_id == tenant.id))
# Purge tool provider records that hold credentials encrypted under the
# tenant key. Leaving them in place causes /console/api/workspaces/current/
# tool-providers to 500 because decryption fails on stale ciphertext (#35396).
session.execute(delete(BuiltinToolProvider).where(BuiltinToolProvider.tenant_id == tenant.id))
session.execute(delete(ApiToolProvider).where(ApiToolProvider.tenant_id == tenant.id))
session.execute(delete(MCPToolProvider).where(MCPToolProvider.tenant_id == tenant.id))
click.echo(
click.style(
f"Congratulations! The asymmetric key pair of workspace {tenant.id} has been reset.",
fg="green",
)
)
@click.command("convert-to-agent-apps", help="Convert Agent Assistant to Agent App.")
def convert_to_agent_apps():
def convert_to_agent_apps() -> None:
"""
Convert Agent Assistant to Agent App.
"""
click.echo(click.style("Starting convert to agent apps.", fg="green"))
proceeded_app_ids = []
proceeded_app_ids: list[str] = []
while True:
# fetch first 1000 apps
@@ -133,48 +74,14 @@ def convert_to_agent_apps():
click.echo(click.style(f"Conversion complete. Converted {len(proceeded_app_ids)} agent apps.", fg="green"))
@click.command("upgrade-db", help="Upgrade the database")
def upgrade_db():
click.echo("Preparing database migration...")
lock = DbMigrationAutoRenewLock(
redis_client=redis_client,
name="db_upgrade_lock",
ttl_seconds=DB_UPGRADE_LOCK_TTL_SECONDS,
logger=logger,
log_context="db_migration",
)
if lock.acquire(blocking=False):
migration_succeeded = False
try:
click.echo(click.style("Starting database migration.", fg="green"))
# run db migration
import flask_migrate
flask_migrate.upgrade()
migration_succeeded = True
click.echo(click.style("Database migration successful!", fg="green"))
except Exception as e:
logger.exception("Failed to execute database migration")
click.echo(click.style(f"Database migration failed: {e}", fg="red"))
raise SystemExit(1)
finally:
status = "successful" if migration_succeeded else "failed"
lock.release_safely(status=status)
else:
click.echo("Database migration skipped")
@click.command("fix-app-site-missing", help="Fix app related site missing issue.")
def fix_app_site_missing():
def fix_app_site_missing() -> None:
"""
Fix app related site missing issue.
"""
click.echo(click.style("Starting fix for missing app-related sites.", fg="green"))
failed_app_ids = []
failed_app_ids: list[str] = []
while True:
sql = """select apps.id as id from apps left join sites on sites.app_id=apps.id
where sites.id is null limit 1000"""
+45
View File
@@ -0,0 +1,45 @@
"""Database schema migration CLI commands."""
import logging
import click
from extensions.ext_redis import redis_client
from libs.db_migration_lock import DbMigrationAutoRenewLock
logger = logging.getLogger(__name__)
DB_UPGRADE_LOCK_TTL_SECONDS = 60
@click.command("upgrade-db", help="Upgrade the database")
def upgrade_db() -> None:
click.echo("Preparing database migration...")
lock = DbMigrationAutoRenewLock(
redis_client=redis_client,
name="db_upgrade_lock",
ttl_seconds=DB_UPGRADE_LOCK_TTL_SECONDS,
logger=logger,
log_context="db_migration",
)
if lock.acquire(blocking=False):
migration_succeeded = False
try:
click.echo(click.style("Starting database migration.", fg="green"))
import flask_migrate
flask_migrate.upgrade()
migration_succeeded = True
click.echo(click.style("Database migration successful!", fg="green"))
except Exception as e:
logger.exception("Failed to execute database migration")
click.echo(click.style(f"Database migration failed: {e}", fg="red"))
raise SystemExit(1)
finally:
status = "successful" if migration_succeeded else "failed"
lock.release_safely(status=status)
else:
click.echo("Database migration skipped")
+173
View File
@@ -0,0 +1,173 @@
"""Workflow data migration CLI commands.
TODO: Remove the legacy system file workflow migration command after the production migration is complete.
"""
import logging
from dataclasses import dataclass
import click
from sqlalchemy import select
from sqlalchemy.orm import Session, load_only, sessionmaker
from extensions.ext_database import db
from models.workflow import Workflow, WorkflowType
logger = logging.getLogger(__name__)
@dataclass
class LegacySysFilesWorkflowMigrationStats:
scanned: int = 0
migrated: int = 0
failed: int = 0
batches: int = 0
last_id: str | None = None
def _build_legacy_sys_files_workflow_query(
*,
start_after_id: str | None,
batch_size: int,
tenant_id: str | None,
app_id: str | None,
):
# Workflow IDs are UUID4, so this is not chronological pagination. The migration only needs a stable total
# order that matches the resume cursor; ordering by the same primary-key column used in the `id > cursor`
# predicate lets each batch continue deterministically without offset scans.
stmt = (
select(Workflow)
.options(load_only(Workflow.id, Workflow.type, Workflow.graph))
.where(Workflow.type.in_((WorkflowType.WORKFLOW, WorkflowType.CHAT)))
.order_by(Workflow.id)
.limit(batch_size)
)
if start_after_id:
stmt = stmt.where(Workflow.id > start_after_id)
if tenant_id:
stmt = stmt.where(Workflow.tenant_id == tenant_id)
if app_id:
stmt = stmt.where(Workflow.app_id == app_id)
return stmt
def _migrate_legacy_sys_files_workflow_batch(
*,
session: Session,
start_after_id: str | None,
batch_size: int,
tenant_id: str | None,
app_id: str | None,
dry_run: bool,
) -> LegacySysFilesWorkflowMigrationStats:
stats = LegacySysFilesWorkflowMigrationStats()
workflows = session.scalars(
_build_legacy_sys_files_workflow_query(
start_after_id=start_after_id,
batch_size=batch_size,
tenant_id=tenant_id,
app_id=app_id,
)
).all()
for workflow in workflows:
stats.scanned += 1
stats.last_id = workflow.id
try:
if workflow.migrate_legacy_sys_files_graph_in_place():
stats.migrated += 1
except Exception:
stats.failed += 1
logger.exception("Failed to migrate legacy sys.files workflow, workflow_id=%s", workflow.id)
if dry_run:
session.rollback()
else:
session.commit()
return stats
def run_legacy_sys_files_workflow_migration(
*,
batch_size: int,
limit: int | None,
start_after_id: str | None,
tenant_id: str | None,
app_id: str | None,
dry_run: bool,
) -> LegacySysFilesWorkflowMigrationStats:
"""Scan Workflow and Advanced Chat graphs in keyset-paginated batches."""
if batch_size <= 0:
raise click.UsageError("--batch-size must be greater than 0")
if limit is not None and limit <= 0:
raise click.UsageError("--limit must be greater than 0 when provided")
session_maker = sessionmaker(db.engine, expire_on_commit=False)
total = LegacySysFilesWorkflowMigrationStats(last_id=start_after_id)
next_start_after_id = start_after_id
while limit is None or total.scanned < limit:
remaining = None if limit is None else limit - total.scanned
current_batch_size = batch_size if remaining is None else min(batch_size, remaining)
if current_batch_size <= 0:
break
with session_maker() as session:
batch_stats = _migrate_legacy_sys_files_workflow_batch(
session=session,
start_after_id=next_start_after_id,
batch_size=current_batch_size,
tenant_id=tenant_id,
app_id=app_id,
dry_run=dry_run,
)
if batch_stats.scanned == 0:
break
total.scanned += batch_stats.scanned
total.migrated += batch_stats.migrated
total.failed += batch_stats.failed
total.batches += 1
total.last_id = batch_stats.last_id
next_start_after_id = batch_stats.last_id
if batch_stats.scanned < current_batch_size:
break
return total
@click.command(
"migrate-legacy-sys-files-workflows",
help="Migrate Workflow and Advanced Chat graphs that still reference deprecated sys.files.",
)
@click.option("--batch-size", default=1000, show_default=True, type=int, help="Number of workflows to scan per batch.")
@click.option("--limit", default=None, type=int, help="Maximum number of workflows to scan in this run.")
@click.option("--start-after-id", default=None, help="Resume scanning after this workflow ID.")
@click.option("--tenant-id", default=None, help="Limit migration to one tenant.")
@click.option("--app-id", default=None, help="Limit migration to one app.")
@click.option("--dry-run", is_flag=True, default=False, help="Scan and report without saving changes.")
def migrate_legacy_sys_files_workflows(
batch_size: int,
limit: int | None,
start_after_id: str | None,
tenant_id: str | None,
app_id: str | None,
dry_run: bool,
) -> None:
stats = run_legacy_sys_files_workflow_migration(
batch_size=batch_size,
limit=limit,
start_after_id=start_after_id,
tenant_id=tenant_id,
app_id=app_id,
dry_run=dry_run,
)
click.echo(
"Legacy sys.files workflow migration finished: "
f"scanned={stats.scanned} migrated={stats.migrated} failed={stats.failed} "
f"batches={stats.batches} last_id={stats.last_id or ''}"
)
if dry_run:
click.echo("Dry run only: no workflow graph changes were saved.")
+64
View File
@@ -0,0 +1,64 @@
"""Workspace maintenance CLI commands."""
import click
from sqlalchemy import delete, select
from sqlalchemy.orm import sessionmaker
from configs import dify_config
from enums.deployment_edition import DeploymentEdition
from extensions.ext_database import db
from libs.rsa import generate_key_pair
from models import Tenant
from models.provider import Provider, ProviderModel
from models.tools import ApiToolProvider, BuiltinToolProvider, MCPToolProvider
@click.command(
"reset-encrypt-key-pair",
help="Reset the asymmetric key pair of workspace for encrypt LLM credentials. "
"After the reset, all LLM credentials and tool provider credentials "
"(builtin / API / MCP) will be purged, requiring re-entry. "
"Only support SELF_HOSTED mode.",
)
@click.confirmation_option(
prompt=click.style(
"Are you sure you want to reset encrypt key pair? "
"This will also purge builtin / API / MCP tool provider records for every tenant. "
"This operation cannot be rolled back!",
fg="red",
)
)
def reset_encrypt_key_pair() -> None:
"""
Reset the encrypted key pair of workspace for encrypt LLM credentials.
After the reset, all LLM credentials will become invalid, requiring re-entry.
Only support SELF_HOSTED mode.
"""
if dify_config.DEPLOYMENT_EDITION == DeploymentEdition.CLOUD:
click.echo(click.style("This command is only for SELF_HOSTED installations.", fg="red"))
return
with sessionmaker(db.engine, expire_on_commit=False).begin() as session:
tenants = session.scalars(select(Tenant)).all()
for tenant in tenants:
if not tenant:
click.echo(click.style("No workspaces found. Run /install first.", fg="red"))
return
tenant.encrypt_public_key = generate_key_pair(tenant.id)
session.execute(delete(Provider).where(Provider.provider_type == "custom", Provider.tenant_id == tenant.id))
session.execute(delete(ProviderModel).where(ProviderModel.tenant_id == tenant.id))
# Purge tool provider records that hold credentials encrypted under the
# tenant key. Leaving them in place causes /console/api/workspaces/current/
# tool-providers to 500 because decryption fails on stale ciphertext (#35396).
session.execute(delete(BuiltinToolProvider).where(BuiltinToolProvider.tenant_id == tenant.id))
session.execute(delete(ApiToolProvider).where(ApiToolProvider.tenant_id == tenant.id))
session.execute(delete(MCPToolProvider).where(MCPToolProvider.tenant_id == tenant.id))
click.echo(
click.style(
f"Congratulations! The asymmetric key pair of workspace {tenant.id} has been reset.",
fg="green",
)
)
@@ -12,6 +12,11 @@ class AgentBackendConfig(BaseSettings):
default=None,
)
AGENT_BACKEND_API_TOKEN: str | None = Field(
description="Bearer token for authenticating with the Agent backend /runs API.",
default=None,
)
AGENT_BACKEND_USE_FAKE: bool = Field(
description="Use the deterministic in-process fake Agent backend client.",
default=False,
+6
View File
@@ -266,6 +266,12 @@ class PluginConfig(BaseSettings):
default=60 * 60,
)
PLUGIN_MODEL_PROVIDERS_CACHE_ENABLED: bool = Field(
description="Whether tenant plugin model providers are cached in Redis. Disable when plugins are installed "
"by a system other than this one, which cannot invalidate the cache when a tenant's plugins change.",
default=True,
)
PLUGIN_MODEL_PROVIDERS_CACHE_TTL: PositiveInt = Field(
description="TTL in seconds for caching tenant plugin model providers in Redis",
default=60 * 60 * 24,
+77
View File
@@ -58,6 +58,7 @@ from services.app_service import (
AppResponseView,
AppService,
CreateAppParams,
RecentAppMode,
StarredAppListParams,
)
from services.enterprise import rbac_service as enterprise_rbac_service
@@ -139,6 +140,10 @@ class AppListBaseQuery(BaseModel):
raise ValueError("Invalid UUID format in creator_ids.") from exc
class RecentAppListQuery(BaseModel):
limit: int = Field(default=8, ge=1, le=8, description="Number of recently modified apps to return (1-8)")
class AppListQuery(AppListBaseQuery):
pass
@@ -411,6 +416,33 @@ class AppPartial(AppResponseModel):
return to_timestamp(value)
class RecentAppResponse(ResponseModel):
id: str
name: str
icon_type: IconType | None = None
icon: str | None = None
icon_background: str | None = None
mode: RecentAppMode
author_name: str | None = None
updated_at: int
permission_keys: list[str] = Field(default_factory=list)
maintainer: str | None = None
@computed_field(return_type=str | None) # type: ignore[prop-decorator]
@property
def icon_url(self) -> str | None:
return build_icon_url(self.icon_type, self.icon)
@field_validator("updated_at", mode="before")
@classmethod
def _normalize_timestamp(cls, value: datetime | int) -> int:
return to_timestamp(value)
class RecentAppListResponse(ResponseModel):
data: list[RecentAppResponse]
class AppDetail(AppResponseModel):
id: str
name: str
@@ -575,6 +607,8 @@ register_schema_models(
register_response_schema_models(
console_ns,
AppPartial,
RecentAppResponse,
RecentAppListResponse,
AppDetailWithSite,
AppPagination,
)
@@ -699,6 +733,49 @@ class AppListApi(Resource):
return app_detail.model_dump(mode="json"), 201
@console_ns.route("/apps/recent")
class RecentAppListApi(Resource):
@console_ns.doc("list_recent_apps")
@console_ns.doc(description="Get recently modified apps for the home Continue Work section")
@console_ns.doc(params=query_params_from_model(RecentAppListQuery))
@console_ns.response(200, "Success", console_ns.models[RecentAppListResponse.__name__])
@setup_required
@login_required
@account_initialization_required
@enterprise_license_required
@with_session(write=False)
@with_current_user_id
@with_current_tenant_id
def get(self, current_tenant_id: str, current_user_id: str, session: Session):
"""Return the lightweight app cards needed by the Explore home page."""
args = query_params_from_request(RecentAppListQuery)
params = AppListParams(limit=args.limit)
permissions = enterprise_rbac_service.RBACService.MyPermissions.get(
current_tenant_id,
current_user_id,
session=session,
)
if dify_config.RBAC_ENABLED:
access_filter = resolve_app_access_filter(
current_tenant_id,
current_user_id,
session=session,
permissions=permissions,
)
access_filter.apply_to_params(params)
recent_apps = AppService().get_recent_apps(current_user_id, current_tenant_id, params, session)
permission_keys_map = permissions.app.permission_keys_by_resource_ids([app.id for app in recent_apps])
response_items = [
RecentAppResponse.model_validate(app, from_attributes=True).model_copy(
update={"permission_keys": permission_keys_map.get(app.id, [])}
)
for app in recent_apps
]
return dump_response(RecentAppListResponse, {"data": response_items}), 200
@console_ns.route("/apps/starred")
class StarredAppListApi(Resource):
@console_ns.doc("list_starred_apps")
-2
View File
@@ -23,7 +23,6 @@ from .knowledge import retrieval as _knowledge_retrieval
from .plugin import agent_config as _agent_config
from .plugin import agent_drive as _agent_drive
from .plugin import plugin as _plugin
from .workspace import plugin_model_providers as _plugin_model_providers
from .workspace import workspace as _workspace
api.add_namespace(inner_api_ns)
@@ -36,7 +35,6 @@ __all__ = [
"_knowledge_retrieval",
"_mail",
"_plugin",
"_plugin_model_providers",
"_runtime_credentials",
"_workspace",
"api",
@@ -1,39 +0,0 @@
from flask_restx import Resource
from pydantic import BaseModel, ConfigDict, Field
from controllers.common.schema import register_schema_model
from controllers.console.wraps import setup_required
from controllers.inner_api import inner_api_ns
from controllers.inner_api.wraps import enterprise_inner_api_only
from core.plugin.plugin_service import PluginService
class InvalidatePluginModelProvidersCachePayload(BaseModel):
model_config = ConfigDict(extra="forbid")
tenant_ids: list[str] = Field(default_factory=list, description="Workspace ids whose cache should be invalidated")
register_schema_model(inner_api_ns, InvalidatePluginModelProvidersCachePayload)
@inner_api_ns.route("/enterprise/workspace/plugin-model-providers/invalidate")
class EnterprisePluginModelProvidersCacheInvalidate(Resource):
@setup_required
@enterprise_inner_api_only
@inner_api_ns.doc(
"enterprise_invalidate_plugin_model_providers_cache",
responses={
200: "Cache invalidated",
400: "Invalid request",
401: "Unauthorized - invalid API key",
},
)
@inner_api_ns.expect(inner_api_ns.models[InvalidatePluginModelProvidersCachePayload.__name__])
def post(self):
args = InvalidatePluginModelProvidersCachePayload.model_validate(inner_api_ns.payload or {})
for tenant_id in args.tenant_ids:
PluginService.invalidate_plugin_model_providers_cache(tenant_id)
return {"result": "success"}, 200
@@ -26,6 +26,10 @@ from controllers.service_api.app.error import (
ProviderQuotaExceededError,
WorkflowVersionExecutionNotAllowedError,
)
from controllers.service_api.app.legacy_system_files import (
attach_legacy_system_file_warning_for_service_api,
normalize_legacy_system_file_args_for_service_api,
)
from controllers.service_api.schema import (
InputFileList,
expect_user_json,
@@ -390,6 +394,7 @@ class ChatApi(Resource):
args["external_trace_id"] = external_trace_id
streaming = _resolve_agent_app_streaming(app_mode=app_mode, response_mode=payload.response_mode)
legacy_system_file_compat = None
try:
# Eagerly validate conversation to avoid hanging on invalid conversation_id
@@ -401,6 +406,14 @@ class ChatApi(Resource):
session=session,
)
if app_mode == AppMode.ADVANCED_CHAT:
args, legacy_system_file_compat = normalize_legacy_system_file_args_for_service_api(
session=session,
app_model=app_model,
args=args,
raw_payload=service_api_ns.payload,
workflow_id=args.get("workflow_id"),
)
response = AppGenerateService.generate(
session=session,
app_model=app_model,
@@ -409,6 +422,7 @@ class ChatApi(Resource):
invoke_from=InvokeFrom.SERVICE_API,
streaming=streaming,
)
response = attach_legacy_system_file_warning_for_service_api(response, legacy_system_file_compat)
# response-contract:ignore compact_generate_response
return helper.compact_generate_response(response)
@@ -0,0 +1,72 @@
"""Temporary Service API adapter for the deprecated workflow file input."""
from collections.abc import Generator, Mapping
from typing import Any
from sqlalchemy.orm import Session
from core.app.entities.app_invoke_entities import InvokeFrom
from core.app.features.rate_limiting.rate_limit import RateLimitGenerator
from core.workflow.legacy_system_files import (
LegacySysFilesCompatVariable,
attach_legacy_sys_files_warning,
normalize_legacy_sys_files_args,
)
from models.model import App
from services.app_generate_service import AppGenerateService
type ServiceAPIGenerateResponse = Mapping[str, Any] | Generator[str, None, None] | RateLimitGenerator
def normalize_legacy_system_file_args_for_service_api(
*,
session: Session,
app_model: App,
args: dict[str, Any],
raw_payload: Mapping[str, Any] | None,
workflow_id: str | None = None,
) -> tuple[dict[str, Any], LegacySysFilesCompatVariable | None]:
# TODO: Remove this hidden Service API compatibility path after all persisted workflows are migrated.
args_with_hidden_system = _copy_hidden_system_files_arg(args=args, raw_payload=raw_payload)
if not _has_legacy_file_arg(args_with_hidden_system):
return args, None
workflow = AppGenerateService.get_workflow(
app_model,
InvokeFrom.SERVICE_API,
workflow_id,
session=session,
)
return normalize_legacy_sys_files_args(graph=workflow.graph_dict, args=args_with_hidden_system)
def attach_legacy_system_file_warning_for_service_api(
response: ServiceAPIGenerateResponse,
compat_variable: LegacySysFilesCompatVariable | None,
) -> ServiceAPIGenerateResponse:
# TODO: Remove this warning once Service API clients no longer need the legacy migration notice.
if compat_variable is None:
return response
return attach_legacy_sys_files_warning(response, compat_variable)
def _copy_hidden_system_files_arg(
*,
args: dict[str, Any],
raw_payload: Mapping[str, Any] | None,
) -> dict[str, Any]:
system = raw_payload.get("system") if isinstance(raw_payload, Mapping) else None
if not isinstance(system, Mapping) or "files" not in system or system["files"] is None:
return args
copied_args = dict(args)
copied_args["system"] = {"files": system["files"]}
return copied_args
def _has_legacy_file_arg(args: Mapping[str, Any]) -> bool:
if args.get("files") is not None:
return True
system = args.get("system")
return isinstance(system, Mapping) and system.get("files") is not None
@@ -30,6 +30,10 @@ from controllers.service_api.app.error import (
ProviderQuotaExceededError,
WorkflowVersionExecutionNotAllowedError,
)
from controllers.service_api.app.legacy_system_files import (
attach_legacy_system_file_warning_for_service_api,
normalize_legacy_system_file_args_for_service_api,
)
from controllers.service_api.schema import (
expect_user_json,
expect_with_user,
@@ -344,6 +348,12 @@ class WorkflowRunApi(Resource):
streaming = payload.response_mode == "streaming"
try:
args, legacy_system_file_compat = normalize_legacy_system_file_args_for_service_api(
session=session,
app_model=app_model,
args=args,
raw_payload=service_api_ns.payload,
)
response = AppGenerateService.generate(
session=session,
app_model=app_model,
@@ -352,6 +362,7 @@ class WorkflowRunApi(Resource):
invoke_from=InvokeFrom.SERVICE_API,
streaming=streaming,
)
response = attach_legacy_system_file_warning_for_service_api(response, legacy_system_file_compat)
# response-contract:ignore compact_generate_response
return helper.compact_generate_response(response)
@@ -471,6 +482,13 @@ class WorkflowRunByIdApi(Resource):
streaming = payload.response_mode == "streaming"
try:
args, legacy_system_file_compat = normalize_legacy_system_file_args_for_service_api(
session=session,
app_model=app_model,
args=args,
raw_payload=service_api_ns.payload,
workflow_id=workflow_id,
)
response = AppGenerateService.generate(
session=session,
app_model=app_model,
@@ -479,6 +497,7 @@ class WorkflowRunByIdApi(Resource):
invoke_from=InvokeFrom.SERVICE_API,
streaming=streaming,
)
response = attach_legacy_system_file_warning_for_service_api(response, legacy_system_file_compat)
# response-contract:ignore compact_generate_response
return helper.compact_generate_response(response)
@@ -46,6 +46,7 @@ from core.ops.ops_trace_manager import TraceQueueManager
from core.prompt.utils.get_thread_messages_length import get_thread_messages_length
from core.repositories import DifyCoreRepositoryFactory
from core.repositories.factory import WorkflowExecutionRepository, WorkflowNodeExecutionRepository
from core.workflow.legacy_system_files import normalize_legacy_sys_files_args
from extensions.ext_database import db
from factories import file_factory
from graphon.filters import ResponseStreamFilter
@@ -147,6 +148,8 @@ class AdvancedChatAppGenerator(MessageBasedAppGenerator):
if not args.get("query"):
raise ValueError("query is required")
# TODO: Remove this compatibility normalization after all persisted workflows are migrated.
args, _ = normalize_legacy_sys_files_args(graph=workflow.graph_dict, args=args)
query = args["query"]
if not isinstance(query, str):
raise ValueError("query must be a string")
@@ -616,23 +619,34 @@ class AdvancedChatAppGenerator(MessageBasedAppGenerator):
message_snapshot = MessageSnapshot.from_message(message)
session.close()
# return response or stream generator
response = self._handle_advanced_chat_response(
application_generate_entity=application_generate_entity,
workflow=workflow_snapshot,
queue_manager=queue_manager,
conversation=conversation_snapshot,
message=message_snapshot,
user=user,
stream=stream,
draft_var_saver_factory=self._get_draft_var_saver_factory(
invoke_from,
account=user,
tenant_id=application_generate_entity.app_config.tenant_id,
),
)
try:
response = self._handle_advanced_chat_response(
application_generate_entity=application_generate_entity,
workflow=workflow_snapshot,
queue_manager=queue_manager,
conversation=conversation_snapshot,
message=message_snapshot,
user=user,
stream=stream,
draft_var_saver_factory=self._get_draft_var_saver_factory(
invoke_from,
account=user,
tenant_id=application_generate_entity.app_config.tenant_id,
),
)
converted_response = AdvancedChatAppGenerateResponseConverter.convert(
response=response,
invoke_from=invoke_from,
)
except BaseException:
self._join_worker_thread(worker_thread)
raise
return AdvancedChatAppGenerateResponseConverter.convert(response=response, invoke_from=invoke_from)
if isinstance(converted_response, Generator):
return self._wrap_stream_with_worker_thread_join(converted_response, worker_thread)
self._join_worker_thread(worker_thread)
return converted_response
def _generate_worker(
self,
@@ -538,6 +538,7 @@ class AgentAppGenerator(MessageBasedAppGenerator):
request_builder=AgentAppRuntimeRequestBuilder(credentials_provider=credentials_provider),
agent_backend_client=create_agent_backend_run_client(
base_url=dify_config.AGENT_BACKEND_BASE_URL,
api_token=dify_config.AGENT_BACKEND_API_TOKEN,
use_fake=dify_config.AGENT_BACKEND_USE_FAKE,
fake_scenario=dify_config.AGENT_BACKEND_FAKE_SCENARIO,
stream_read_timeout_seconds=dify_config.AGENT_BACKEND_STREAM_READ_TIMEOUT_SECONDS,
+29
View File
@@ -1,3 +1,5 @@
import logging
import threading
from collections.abc import Generator, Mapping, Sequence
from contextlib import AbstractContextManager, nullcontext
from typing import TYPE_CHECKING, Any, Union, final
@@ -23,6 +25,10 @@ from services.workflow_draft_variable_service import DraftVariableSaver as Draft
if TYPE_CHECKING:
from graphon.variables.input_entities import VariableEntity
logger = logging.getLogger(__name__)
_WORKER_THREAD_JOIN_TIMEOUT_SECONDS = 300
@final
class _DebuggerDraftVariableSaver:
@@ -64,6 +70,29 @@ class _DebuggerDraftVariableSaver:
class BaseAppGenerator:
_file_access_controller: DatabaseFileAccessController = DatabaseFileAccessController()
@staticmethod
def _join_worker_thread(worker_thread: threading.Thread) -> None:
# Bound the wait so a leaked app worker cannot occupy an execution slot indefinitely.
worker_thread.join(timeout=_WORKER_THREAD_JOIN_TIMEOUT_SECONDS)
if worker_thread.is_alive():
logger.warning(
"Possible app worker thread leak: thread_name=%s timeout_seconds=%s; "
"continuing without waiting further to avoid occupying an execution slot indefinitely",
worker_thread.name,
_WORKER_THREAD_JOIN_TIMEOUT_SECONDS,
)
@staticmethod
def _wrap_stream_with_worker_thread_join[ResponseT](
response_stream: Generator[ResponseT, None, None],
worker_thread: threading.Thread,
) -> Generator[ResponseT, None, None]:
"""Keep the producer owned by the response stream until both finish."""
try:
yield from response_stream
finally:
BaseAppGenerator._join_worker_thread(worker_thread)
@staticmethod
def _bind_file_access_scope(
*,
@@ -155,8 +155,10 @@ class WorkflowResponseConverter:
# TODO(@future-refactor): store system variables separately from user inputs so we don't
# need to flatten `sys.*` entries into the input payload just for rerun/export tooling.
if field_name == SystemVariableKey.CONVERSATION_ID:
# Conversation IDs are session-scoped; omitting them keeps workflow inputs
# reusable without pinning new runs to a prior conversation.
# Conversation IDs are session-scoped; omitting them keeps workflow inputs reusable.
continue
if field_name == SystemVariableKey.FILES:
# When files are exposed as an input, application inputs use the canonical `userinput.files` key.
continue
inputs[f"sys.{field_name}"] = value
handled = WorkflowEntry.handle_special_values(inputs)
@@ -351,17 +351,28 @@ class PipelineGenerator(BaseAppGenerator):
user,
tenant_id=pipeline.tenant_id,
)
# return response or stream generator
response = self._handle_response(
application_generate_entity=application_generate_entity,
workflow=workflow,
queue_manager=queue_manager,
user=user,
stream=streaming,
draft_var_saver_factory=draft_var_saver_factory,
)
try:
response = self._handle_response(
application_generate_entity=application_generate_entity,
workflow=workflow,
queue_manager=queue_manager,
user=user,
stream=streaming,
draft_var_saver_factory=draft_var_saver_factory,
)
converted_response = WorkflowAppGenerateResponseConverter.convert(
response=response,
invoke_from=invoke_from,
)
except BaseException:
self._join_worker_thread(worker_thread)
raise
return WorkflowAppGenerateResponseConverter.convert(response=response, invoke_from=invoke_from)
if isinstance(converted_response, Generator):
return self._wrap_stream_with_worker_thread_join(converted_response, worker_thread)
self._join_worker_thread(worker_thread)
return converted_response
def single_iteration_generate(
self,
+24 -10
View File
@@ -41,6 +41,7 @@ from core.helper.trace_id_helper import (
from core.ops.ops_trace_manager import TraceQueueManager
from core.repositories import DifyCoreRepositoryFactory
from core.repositories.factory import WorkflowExecutionRepository, WorkflowNodeExecutionRepository
from core.workflow.legacy_system_files import normalize_legacy_sys_files_args
from extensions.ext_database import db
from factories import file_factory
from graphon.filters import ResponseStreamFilter
@@ -164,6 +165,8 @@ class WorkflowAppGenerator(BaseAppGenerator):
pause_state_config: PauseStateLayerConfig | None = None,
) -> Mapping[str, Any] | Generator[Mapping[str, Any] | str, None, None]:
with self._bind_file_access_scope(tenant_id=app_model.tenant_id, user=user, invoke_from=invoke_from):
# TODO: Remove this compatibility normalization after all persisted workflows are migrated.
args, _ = normalize_legacy_sys_files_args(graph=workflow.graph_dict, args=args)
files: Sequence[Mapping[str, Any]] = args.get("files") or []
# parse files
@@ -405,17 +408,28 @@ class WorkflowAppGenerator(BaseAppGenerator):
tenant_id=app_model.tenant_id,
)
# return response or stream generator
response = self._handle_response(
application_generate_entity=application_generate_entity,
workflow=workflow,
queue_manager=queue_manager,
user=user,
draft_var_saver_factory=draft_var_saver_factory,
stream=streaming,
)
try:
response = self._handle_response(
application_generate_entity=application_generate_entity,
workflow=workflow,
queue_manager=queue_manager,
user=user,
draft_var_saver_factory=draft_var_saver_factory,
stream=streaming,
)
converted_response = WorkflowAppGenerateResponseConverter.convert(
response=response,
invoke_from=invoke_from,
)
except BaseException:
self._join_worker_thread(worker_thread)
raise
return WorkflowAppGenerateResponseConverter.convert(response=response, invoke_from=invoke_from)
if isinstance(converted_response, Generator):
return self._wrap_stream_with_worker_thread_join(converted_response, worker_thread)
self._join_worker_thread(worker_thread)
return converted_response
def single_iteration_generate(
self,
+38 -23
View File
@@ -66,7 +66,7 @@ from services.enterprise.plugin_manager_service import (
PreUninstallPluginRequest,
)
from services.errors.plugin import PluginInstallationForbiddenError
from services.feature_service import FeatureService, PluginInstallationScope
from services.feature_service import FeatureService, PluginInstallationPermissionModel, PluginInstallationScope
logger = logging.getLogger(__name__)
_provider_entities_adapter: TypeAdapter[list[ProviderEntity]] = TypeAdapter(list[ProviderEntity])
@@ -434,14 +434,18 @@ class PluginService:
exc_info=True,
)
@classmethod
def _fetch_plugin_model_providers_uncached(
cls, tenant_id: str, client: PluginModelClient | None
) -> tuple[ProviderEntity, ...]:
model_client = client or PluginModelClient()
return tuple(cls._to_provider_entity(provider) for provider in model_client.fetch_model_providers(tenant_id))
@classmethod
def _fetch_and_cache_plugin_model_providers(
cls, tenant_id: str, client: PluginModelClient | None, *, refresh_generation: int | None
) -> tuple[ProviderEntity, ...]:
model_client = client or PluginModelClient()
providers = tuple(
cls._to_provider_entity(provider) for provider in model_client.fetch_model_providers(tenant_id)
)
providers = cls._fetch_plugin_model_providers_uncached(tenant_id, client)
generation = cls._load_plugin_model_providers_generation(tenant_id)
if generation is not None and generation == refresh_generation:
cls._store_cached_plugin_model_providers(tenant_id, generation, providers)
@@ -471,6 +475,9 @@ class PluginService:
are intentionally owned by this service so tenant isolation and cache
expiry are handled in one place.
"""
if not dify_config.PLUGIN_MODEL_PROVIDERS_CACHE_ENABLED:
return cls._fetch_plugin_model_providers_uncached(tenant_id, client)
deadline = time.monotonic() + cls.PLUGIN_MODEL_PROVIDERS_LOCK_WAIT_TIMEOUT
while True:
@@ -597,22 +604,30 @@ class PluginService:
return result
@staticmethod
def _check_marketplace_only_permission():
def _check_marketplace_only_permission() -> None:
"""
Check if the marketplace only permission is enabled
"""
features = FeatureService.get_system_features()
if features.plugin_installation_permission.restrict_to_marketplace_only:
permission = PluginService._get_plugin_installation_permission()
if permission.restrict_to_marketplace_only:
raise PluginInstallationForbiddenError("Plugin installation is restricted to marketplace only")
@staticmethod
def _check_plugin_installation_scope(plugin_verification: PluginVerification | None):
def _get_plugin_installation_permission() -> PluginInstallationPermissionModel:
"""Resolve the validated policy and reject deny-all before any installation side effect."""
permission = FeatureService.get_plugin_installation_permission()
if permission.plugin_installation_scope == PluginInstallationScope.NONE:
raise PluginInstallationForbiddenError("Installing plugins is not allowed")
return permission
@staticmethod
def _check_plugin_installation_scope(plugin_verification: PluginVerification | None) -> None:
"""
Check the plugin installation scope
"""
features = FeatureService.get_system_features()
permission = PluginService._get_plugin_installation_permission()
match features.plugin_installation_permission.plugin_installation_scope:
match permission.plugin_installation_scope:
case PluginInstallationScope.OFFICIAL_ONLY:
if (
plugin_verification is None
@@ -627,10 +642,10 @@ class PluginService:
raise PluginInstallationForbiddenError(
"Plugin installation is restricted to official and specific partners"
)
case PluginInstallationScope.NONE:
raise PluginInstallationForbiddenError("Installing plugins is not allowed")
case PluginInstallationScope.ALL:
pass
case _:
raise PluginInstallationForbiddenError("Plugin installation policy is invalid")
@staticmethod
def get_debugging_key(tenant_id: str) -> str:
@@ -900,7 +915,7 @@ class PluginService:
# check if plugin pkg is already downloaded
manager = PluginInstaller()
features = FeatureService.get_system_features()
permission = PluginService._get_plugin_installation_permission()
try:
manager.fetch_plugin_manifest(tenant_id, new_plugin_unique_identifier)
@@ -912,7 +927,7 @@ class PluginService:
response = manager.upload_pkg(
tenant_id,
pkg,
verify_signature=features.plugin_installation_permission.restrict_to_marketplace_only,
verify_signature=permission.restrict_to_marketplace_only,
)
# check if the plugin is available to install
@@ -967,11 +982,11 @@ class PluginService:
"""
PluginService._check_marketplace_only_permission()
manager = PluginInstaller()
features = FeatureService.get_system_features()
permission = PluginService._get_plugin_installation_permission()
response = manager.upload_pkg(
tenant_id,
pkg,
verify_signature=features.plugin_installation_permission.restrict_to_marketplace_only,
verify_signature=permission.restrict_to_marketplace_only,
)
PluginService._check_plugin_installation_scope(response.verification)
@@ -989,13 +1004,13 @@ class PluginService:
pkg = download_with_size_limit(
f"https://github.com/{repo}/releases/download/{version}/{package}", dify_config.PLUGIN_MAX_PACKAGE_SIZE
)
features = FeatureService.get_system_features()
permission = PluginService._get_plugin_installation_permission()
manager = PluginInstaller()
response = manager.upload_pkg(
tenant_id,
pkg,
verify_signature=features.plugin_installation_permission.restrict_to_marketplace_only,
verify_signature=permission.restrict_to_marketplace_only,
)
PluginService._check_plugin_installation_scope(response.verification)
@@ -1069,7 +1084,7 @@ class PluginService:
if not dify_config.MARKETPLACE_ENABLED:
raise ValueError("marketplace is not enabled")
features = FeatureService.get_system_features()
permission = PluginService._get_plugin_installation_permission()
manager = PluginInstaller()
try:
@@ -1079,7 +1094,7 @@ class PluginService:
response = manager.upload_pkg(
tenant_id,
pkg,
verify_signature=features.plugin_installation_permission.restrict_to_marketplace_only,
verify_signature=permission.restrict_to_marketplace_only,
)
# check if the plugin is available to install
PluginService._check_plugin_installation_scope(response.verification)
@@ -1101,7 +1116,7 @@ class PluginService:
# collect actual plugin_unique_identifiers
actual_plugin_unique_identifiers = []
metas = []
features = FeatureService.get_system_features()
permission = PluginService._get_plugin_installation_permission()
# check if already downloaded
for plugin_unique_identifier in plugin_unique_identifiers:
@@ -1119,7 +1134,7 @@ class PluginService:
response = manager.upload_pkg(
tenant_id,
pkg,
verify_signature=features.plugin_installation_permission.restrict_to_marketplace_only,
verify_signature=permission.restrict_to_marketplace_only,
)
# check if the plugin is available to install
PluginService._check_plugin_installation_scope(response.verification)
@@ -2023,6 +2023,8 @@ class DatasetRetrieval:
redis_client.zremrangebyscore(key, 0, current_time - 60000)
request_count = redis_client.zcard(key)
if request_count > knowledge_rate_limit.limit:
# The rate-limit exception is raised after this block, so commit the audit row
# explicitly instead of relying on the Session context, which only closes it.
with session_factory.create_session() as session:
rate_limit_log = RateLimitLog(
tenant_id=tenant_id,
@@ -2030,6 +2032,7 @@ class DatasetRetrieval:
operation="knowledge",
)
session.add(rate_limit_log)
session.commit()
raise exc.RateLimitExceededError(
"you have reached the knowledge base request rate limit of your subscription."
)
+2
View File
@@ -107,6 +107,8 @@ class MCPTool(Tool):
if self.entity.output_schema and result.structuredContent:
for k, v in result.structuredContent.items():
yield self.create_variable_message(k, v)
elif result.structuredContent:
yield self.create_json_message(result.structuredContent)
def _process_text_content(self, content: TextContent) -> Generator[ToolInvokeMessage, None, None]:
"""Process text content and yield appropriate messages."""
@@ -42,9 +42,9 @@ _NODE_SNIPPETS: dict[str, str] = {
["local_file", "remote_url"]. Only when you include "custom" must you
also set ``allowed_file_extensions`` to a non-empty list like
[".epub", ".rtf"]; otherwise leave it [].
In Advanced-Chat mode ``sys.query`` and ``sys.files`` are automatic
system variables — downstream nodes may reference them; do NOT add
them to ``variables``.""",
In Advanced-Chat mode ``sys.query`` is automatic. ``userinput.files`` is
the automatic file-upload variable in both app modes. Downstream nodes
may reference these variables; do NOT add them to ``variables``.""",
"end": """\
- end (Workflow mode only):
{"outputs": [
@@ -168,8 +168,8 @@ _NODE_SNIPPETS: dict[str, str] = {
Single output variable ``text``: a string when ``is_array_file`` is false,
an array of strings (one per file) when it is true. ``variable_selector``
MUST point at a ``start`` variable declared with type "file" / "file-list"
(or ``sys.files`` in Advanced-Chat mode). That start variable MUST set a
non-empty ``allowed_file_types`` (use ["document"] for document text).""",
(or the automatic ``userinput.files`` variable). A declared start variable
MUST set a non-empty ``allowed_file_types`` (use ["document"] for document text).""",
"variable-aggregator": """\
- variable-aggregator (merge mutually-exclusive branches into one output):
{"output_type": "string", # VarType of the merged value — one of
@@ -91,21 +91,20 @@ def format_parallel_plan(
def format_mode_section(mode: str) -> str:
"""Tell each builder which app mode it is configuring for.
Matters most in advanced-chat, where ``sys.query`` / ``sys.files`` are the
sanctioned way to reference the user's message — without this the model
invents start-node variables that postprocess then materializes as
spurious form inputs.
``sys.query`` is available in advanced-chat, while ``userinput.files`` is
available in both app modes. Without this guidance the model invents
start-node variables that postprocess then materializes as spurious inputs.
"""
if mode == "advanced-chat":
return (
"# App mode\n\n"
"advanced-chat: the user's chat message is available as sys.query and uploaded files "
'as sys.files placeholder {{#sys.query#}}, selector ["sys", "query"]. Reference them '
"directly; do NOT invent start-node variables for the chat message.\n\n"
"as userinput.files. Use placeholder {{#sys.query#}} or selector "
'["userinput", "files"] directly; do NOT invent start-node variables for them.\n\n'
)
return (
"# App mode\n\n"
"workflow: there are NO automatic system variables; reference user input only through "
"workflow: uploaded files are available as userinput.files; all other user input must use "
"the start node's declared variables.\n\n"
)
@@ -96,10 +96,10 @@ minimum set of Dify workflow nodes needed to fulfil it, in execution order.
- "text-input" for short single-line values (URLs, names),
- "paragraph" for free-form multi-line text (descriptions, queries),
- "number" / "select" / "file" / "file-list" for the obvious cases.
In Advanced-Chat mode the ``sys.query`` / ``sys.files`` system
variables are automatic — downstream nodes may reference them without
a ``start_inputs`` entry. In Workflow mode there is NO automatic
variable; everything the user supplies must be in ``start_inputs``.
In Advanced-Chat mode ``sys.query`` is automatic. ``userinput.files`` is
automatic in both app modes. Downstream nodes may reference these values
without a ``start_inputs`` entry; every other user-supplied Workflow value
must be declared in ``start_inputs``.
11. Give every node a unique runtime-safe ``id`` using only letters, digits,
and underscores. In create mode use ``node1``, ``node2``, ... in node-list
order. In refine mode preserve the existing id for every retained node.
+7 -4
View File
@@ -1382,8 +1382,8 @@ class WorkflowGenerator:
multiple outputs remain untouched so validation fails closed instead
of guessing which value the workflow should consume.
For Advanced-Chat mode, ``sys.query`` and ``sys.files`` are always
treated as resolved without any declaration. Tool nodes' parameter
``sys.query`` in Advanced-Chat mode and ``userinput.files`` in either mode
are treated as resolved without declarations. Tool nodes' parameter
references aren't validated here because we don't know each tool's
schema the run time validates those.
"""
@@ -1398,9 +1398,12 @@ class WorkflowGenerator:
for node in nodes:
cls._collect_refs_in_data(node.get("data") or {}, refs)
automatic_refs = {("userinput", "files")}
if mode == "advanced-chat":
automatic_refs.add(("sys", "query"))
for node_id, var in refs:
# Advanced-Chat system variables are always resolved.
if mode == "advanced-chat" and node_id == "sys":
if (node_id, var) in automatic_refs:
continue
target = nodes_by_id.get(node_id)
if target is None:
+218
View File
@@ -0,0 +1,218 @@
"""Compatibility helpers for workflows that still reference deprecated `sys.files`.
TODO: Remove this module after all persisted Workflow and Advanced Chat graphs
have been migrated from the deprecated system file variable to `userinput.files`.
"""
from __future__ import annotations
import json
from collections.abc import Generator, Iterable, Mapping
from dataclasses import dataclass
from typing import Any
_LEGACY_SYSTEM_NODE_ID = "sys"
_USER_INPUT_NODE_ID = "userinput"
_LEGACY_FILES_VARIABLE = "files"
_USER_INPUT_FILE_SELECTOR = [_USER_INPUT_NODE_ID, _LEGACY_FILES_VARIABLE]
_USER_INPUT_FILE_INPUT_KEY = ".".join(_USER_INPUT_FILE_SELECTOR)
_LEGACY_FILES_TEMPLATE = "{{#sys.files#}}"
_USER_INPUT_FILES_TEMPLATE = "{{#userinput.files#}}"
@dataclass(frozen=True)
class LegacySysFilesCompatVariable:
node_id: str
variable_name: str
@dataclass(frozen=True)
class LegacySysFilesGraphMigrationResult:
graph: dict[str, Any]
changed: bool
def migrate_legacy_sys_files_graph_with_result(
graph: Mapping[str, Any],
) -> LegacySysFilesGraphMigrationResult:
"""Return the migrated graph and whether any legacy reference was rewritten."""
graph_copy = dict(graph)
nodes = graph_copy.get("nodes")
if not isinstance(nodes, list):
return LegacySysFilesGraphMigrationResult(graph=graph_copy, changed=False)
# Legacy references are stored in node data. Restricting both search and replacement to `nodes`
# avoids recursively scanning graph-level metadata and edges for every workflow load.
if not _contains_legacy_sys_files_reference(nodes):
return LegacySysFilesGraphMigrationResult(graph=graph_copy, changed=False)
graph_copy["nodes"] = _replace_legacy_sys_files_references(nodes)
return LegacySysFilesGraphMigrationResult(graph=graph_copy, changed=True)
def resolve_legacy_sys_files_compat_variable(graph: Mapping[str, Any]) -> LegacySysFilesCompatVariable | None:
"""Resolve the target variable used by the `sys.files` compatibility layer."""
nodes = graph.get("nodes")
if not isinstance(nodes, list):
return None
if not _contains_file_input_reference(nodes):
return None
return LegacySysFilesCompatVariable(node_id=_USER_INPUT_NODE_ID, variable_name=_LEGACY_FILES_VARIABLE)
def normalize_legacy_sys_files_args(
*,
graph: Mapping[str, Any],
args: Mapping[str, Any],
) -> tuple[dict[str, Any], LegacySysFilesCompatVariable | None]:
"""Map Service/Web API file arguments onto the `userinput.files` system alias.
The top-level `files` argument and hidden `system.files` payload both feed
the same runtime file collection. After graph references are migrated, the
file collection is exposed in the variable pool as `userinput.files`.
"""
normalized_args = dict(args)
files_from_input, input_files_used = _extract_userinput_files(args)
if input_files_used:
normalized_args["files"] = files_from_input
return normalized_args, None
compat_variable = resolve_legacy_sys_files_compat_variable(graph)
if compat_variable is None:
return normalized_args, None
files, legacy_files_used = _extract_legacy_files(args)
if not legacy_files_used:
return normalized_args, None
if normalized_args.get("files") is None:
normalized_args["files"] = files
raw_inputs = normalized_args.get("inputs")
inputs = dict(raw_inputs) if isinstance(raw_inputs, Mapping) else {}
inputs.setdefault(_USER_INPUT_FILE_INPUT_KEY, files)
normalized_args["inputs"] = inputs
return normalized_args, compat_variable
def attach_legacy_sys_files_warning(
response: Mapping[str, Any] | Iterable[Any],
compat_variable: LegacySysFilesCompatVariable,
) -> Mapping[str, Any] | Generator[str, None, None]:
warning = build_legacy_sys_files_warning(compat_variable)
if isinstance(response, Mapping):
response_with_warning = dict(response)
existing_warnings = response_with_warning.get("warnings")
warnings = list(existing_warnings) if isinstance(existing_warnings, list) else []
warnings.append(warning)
response_with_warning["warnings"] = warnings
return response_with_warning
def _with_warning() -> Generator[str, None, None]:
try:
yield f"data: {json.dumps({'event': 'warning', 'warning': warning})}\n\n"
yield from response
finally:
close = getattr(response, "close", None)
if callable(close):
close()
return _with_warning()
def build_legacy_sys_files_warning(compat_variable: LegacySysFilesCompatVariable) -> str:
variable_selector = ".".join((compat_variable.node_id, compat_variable.variable_name))
return (
"sys.files is deprecated. This workflow now reads files from "
f"`{variable_selector}`; update Service API calls to pass files in "
f"`inputs.{variable_selector}` instead of `system.files` or top-level `files`."
)
def _contains_legacy_sys_files_reference(value: Any) -> bool:
if _is_legacy_sys_files_selector(value):
return True
if isinstance(value, str):
return _LEGACY_FILES_TEMPLATE in value
if isinstance(value, Mapping):
return any(_contains_legacy_sys_files_reference(item) for item in value.values())
if isinstance(value, list):
return any(_contains_legacy_sys_files_reference(item) for item in value)
return False
def _contains_file_input_reference(value: Any) -> bool:
if _is_legacy_sys_files_selector(value) or _is_userinput_files_selector(value):
return True
if isinstance(value, str):
return _LEGACY_FILES_TEMPLATE in value or _USER_INPUT_FILES_TEMPLATE in value
if isinstance(value, Mapping):
return any(_contains_file_input_reference(item) for item in value.values())
if isinstance(value, list):
return any(_contains_file_input_reference(item) for item in value)
return False
def _replace_legacy_sys_files_references(value: Any) -> Any:
if _is_legacy_sys_files_selector(value):
return list(_USER_INPUT_FILE_SELECTOR)
if isinstance(value, str):
return value.replace(_LEGACY_FILES_TEMPLATE, _USER_INPUT_FILES_TEMPLATE)
if isinstance(value, Mapping):
return {key: _replace_legacy_sys_files_references(item) for key, item in value.items()}
if isinstance(value, list):
return [_replace_legacy_sys_files_references(item) for item in value]
return value
def _is_legacy_sys_files_selector(value: Any) -> bool:
return (
isinstance(value, list)
and len(value) == 2
and value[0] == _LEGACY_SYSTEM_NODE_ID
and value[1] == _LEGACY_FILES_VARIABLE
)
def _is_userinput_files_selector(value: Any) -> bool:
return isinstance(value, list) and value == _USER_INPUT_FILE_SELECTOR
def serialized_graph_may_contain_legacy_sys_files(serialized_graph: str) -> bool:
"""Cheaply reject stored graphs that cannot contain a legacy file reference."""
return _LEGACY_FILES_TEMPLATE in serialized_graph or ('"sys"' in serialized_graph and '"files"' in serialized_graph)
def _extract_legacy_files(args: Mapping[str, Any]) -> tuple[Any, bool]:
if "files" in args and args["files"] is not None:
return args["files"], True
system = args.get("system")
if isinstance(system, Mapping) and "files" in system and system["files"] is not None:
return system["files"], True
return None, False
def _extract_userinput_files(args: Mapping[str, Any]) -> tuple[Any, bool]:
inputs = args.get("inputs")
if isinstance(inputs, Mapping) and inputs.get(_USER_INPUT_FILE_INPUT_KEY) is not None:
return inputs[_USER_INPUT_FILE_INPUT_KEY], True
return None, False
+1
View File
@@ -497,6 +497,7 @@ class DifyNodeFactory(NodeFactory):
),
"agent_backend_client": create_agent_backend_run_client(
base_url=dify_config.AGENT_BACKEND_BASE_URL,
api_token=dify_config.AGENT_BACKEND_API_TOKEN,
use_fake=dify_config.AGENT_BACKEND_USE_FAKE,
fake_scenario=dify_config.AGENT_BACKEND_FAKE_SCENARIO,
stream_read_timeout_seconds=dify_config.AGENT_BACKEND_STREAM_READ_TIMEOUT_SECONDS,
@@ -368,7 +368,7 @@ class WorkflowAgentRuntimeRequestBuilder:
if uploaded_files is not None:
lines.append("- Uploaded workflow files:")
lines.append(f" - sys.files: {uploaded_files}")
lines.append(f" - userinput.files: {uploaded_files}")
if resolved_outputs:
lines.append("- Previous node outputs:")
+7
View File
@@ -16,6 +16,7 @@ from .variable_prefixes import (
ENVIRONMENT_VARIABLE_NODE_ID,
RAG_PIPELINE_VARIABLE_NODE_ID,
SYSTEM_VARIABLE_NODE_ID,
USER_INPUT_VARIABLE_NODE_ID,
)
@@ -118,6 +119,12 @@ def build_bootstrap_variables(
*(_with_selector(variable, ENVIRONMENT_VARIABLE_NODE_ID) for variable in environment_variables),
*(_with_selector(variable, CONVERSATION_VARIABLE_NODE_ID) for variable in conversation_variables),
]
# TODO: Stop emitting the legacy `sys.files` selector after stored graphs and Service API callers are migrated.
# `userinput.files` remains the canonical file-upload variable.
for variable in system_variables:
if variable.name == SystemVariableKey.FILES.value:
variables.append(_with_selector(variable, USER_INPUT_VARIABLE_NODE_ID))
break
rag_pipeline_variables_map: defaultdict[str, dict[str, Any]] = defaultdict(dict)
for rag_var in rag_pipeline_variables:
+1
View File
@@ -1,4 +1,5 @@
SYSTEM_VARIABLE_NODE_ID = "sys"
USER_INPUT_VARIABLE_NODE_ID = "userinput"
ENVIRONMENT_VARIABLE_NODE_ID = "env"
CONVERSATION_VARIABLE_NODE_ID = "conversation"
RAG_PIPELINE_VARIABLE_NODE_ID = "rag"
+2
View File
@@ -29,6 +29,7 @@ def init_app(app: DifyApp):
install_rag_pipeline_plugins,
migrate_data_for_plugin,
migrate_dataset_permissions_to_rbac,
migrate_legacy_sys_files_workflows,
migrate_member_roles_to_rbac,
migrate_oss,
migration_data_wizard,
@@ -57,6 +58,7 @@ def init_app(app: DifyApp):
data_migrate,
upgrade_db,
fix_app_site_missing,
migrate_legacy_sys_files_workflows,
migrate_data_for_plugin,
migrate_dataset_permissions_to_rbac,
migrate_member_roles_to_rbac,
+2 -2
View File
@@ -1117,14 +1117,14 @@ class ExporleBanner(TypeBase):
status: Mapped[BannerStatus] = mapped_column(
EnumText(BannerStatus, length=255),
nullable=False,
server_default=sa.text("'enabled'::character varying"),
server_default=sa.text("'enabled'"),
default=BannerStatus.ENABLED,
)
created_at: Mapped[datetime] = mapped_column(
sa.DateTime, nullable=False, server_default=func.current_timestamp(), init=False
)
language: Mapped[str] = mapped_column(
String(255), nullable=False, server_default=sa.text("'en-US'::character varying"), default="en-US"
String(255), nullable=False, server_default=sa.text("'en-US'"), default="en-US"
)
+39 -3
View File
@@ -25,6 +25,10 @@ from typing_extensions import deprecated
from core.trigger.constants import TRIGGER_PLUGIN_NODE_TYPE
from core.workflow.human_input_adapter import adapt_node_config_for_graph
from core.workflow.legacy_system_files import (
migrate_legacy_sys_files_graph_with_result,
serialized_graph_may_contain_legacy_sys_files,
)
from core.workflow.nodes.human_input.pause_reason import (
HumanInputRequired,
)
@@ -325,7 +329,39 @@ class Workflow(Base): # bug
# Currently, the following functions / methods would mutate the returned dict:
#
# - `_get_graph_and_variable_pool_for_single_node_run`.
return json.loads(self.graph) if self.graph else {}
if not self.graph:
return {}
graph = json.loads(self.graph)
if not self._supports_legacy_sys_files_compatibility() or not serialized_graph_may_contain_legacy_sys_files(
self.graph
):
return graph
# TODO: Remove this load-time compatibility rewrite after all persisted workflows are migrated.
return migrate_legacy_sys_files_graph_with_result(graph).graph
def migrate_legacy_sys_files_graph_in_place(self) -> bool:
if (
not self.graph
or not self._supports_legacy_sys_files_compatibility()
or not serialized_graph_may_contain_legacy_sys_files(self.graph)
):
return False
# TODO: Remove this in-place compatibility rewrite after all persisted workflows are migrated.
migration_result = migrate_legacy_sys_files_graph_with_result(json.loads(self.graph))
if migration_result.changed:
self.graph = json.dumps(migration_result.graph)
return migration_result.changed
def _supports_legacy_sys_files_compatibility(self) -> bool:
return self.type in {
WorkflowType.WORKFLOW,
WorkflowType.CHAT,
WorkflowType.WORKFLOW.value,
WorkflowType.CHAT.value,
}
def get_node_config_by_id(self, node_id: str) -> NodeConfigDict:
"""Extract a node configuration from the workflow graph by node ID.
@@ -487,7 +523,7 @@ class Workflow(Base): # bug
"memory":
{
"window": { "enabled": false, "size": 10 },
"query_prompt_template": "{{#sys.query#}}\n\n{{#sys.files#}}",
"query_prompt_template": "{{#sys.query#}}\n\n{{#userinput.files#}}",
"role_prefix": { "user": "", "assistant": "" },
},
"selected": false,
@@ -1520,7 +1556,7 @@ class ConversationVariable(TypeBase):
return variable_factory.build_conversation_variable_from_mapping(mapping)
# Only `sys.query` and `sys.files` could be modified.
# TODO: Remove file-system-variable editability after all persisted workflows are migrated.
_EDITABLE_SYSTEM_VARIABLE = frozenset(("query", "files"))
+39
View File
@@ -1672,6 +1672,23 @@ Create a new application
| 200 | Import confirmed | **application/json**: [Import](#import)<br> |
| 400 | Import failed | **application/json**: [Import](#import)<br> |
### [GET] /apps/recent
**Return the lightweight app cards needed by the Explore home page**
Get recently modified apps for the home Continue Work section
#### Parameters
| Name | Located in | Description | Required | Schema |
| ---- | ---------- | ----------- | -------- | ------ |
| limit | query | Number of recently modified apps to return (1-8) | No | integer, <br>**Default:** 8 |
#### Responses
| Code | Description | Schema |
| ---- | ----------- | ------ |
| 200 | Success | **application/json**: [RecentAppListResponse](#recentapplistresponse)<br> |
### [GET] /apps/starred
Get applications starred by the current account
@@ -21018,6 +21035,28 @@ Whitelist scopes accepted by RBAC app and dataset access config APIs.
| result | string | | Yes |
| updated_at | integer | | Yes |
#### RecentAppListResponse
| Name | Type | Description | Required |
| ---- | ---- | ----------- | -------- |
| data | [ [RecentAppResponse](#recentappresponse) ] | | Yes |
#### RecentAppResponse
| Name | Type | Description | Required |
| ---- | ---- | ----------- | -------- |
| author_name | string | | No |
| icon | string | | No |
| icon_background | string | | No |
| icon_type | [IconType](#icontype) | | No |
| icon_url | string | | Yes |
| id | string | | Yes |
| maintainer | string | | No |
| mode | string, <br>**Available values:** "advanced-chat", "agent-chat", "chat", "completion", "workflow" | *Enum:* `"advanced-chat"`, `"agent-chat"`, `"chat"`, `"completion"`, `"workflow"` | Yes |
| name | string | | Yes |
| permission_keys | [ string ] | | No |
| updated_at | integer | | Yes |
#### RecommendedAppDetailNullableResponse
| Name | Type | Description | Required |
+4 -4
View File
@@ -1,12 +1,12 @@
[project]
name = "dify-api"
version = "1.16.0"
version = "1.16.1"
requires-python = "~=3.12.0"
dependencies = [
# Legacy: mature and widely deployed
"bleach>=6.4.0,<7.0.0",
"boto3>=1.43.46,<2.0.0",
"boto3>=1.43.56,<2.0.0",
"celery>=5.6.3,<6.0.0",
"croniter>=6.2.2,<7.0.0",
"dify-agent",
@@ -193,10 +193,10 @@ dev = [
############################################################
storage = [
"azure-storage-blob>=12.30.0,<13.0.0",
"bce-python-sdk==0.9.72",
"bce-python-sdk==0.9.76",
"cos-python-sdk-v5>=1.9.44,<2.0.0",
"esdk-obs-python>=3.26.6,<4.0.0",
"google-cloud-storage>=3.12.1,<4.0.0",
"google-cloud-storage>=3.13.0,<4.0.0",
"opendal==0.46.0",
"oss2>=2.19.1,<3.0.0",
"supabase>=2.31.0,<3.0.0",
+12 -1
View File
@@ -488,7 +488,7 @@ class AppGenerateService:
)
@classmethod
def _get_workflow(
def get_workflow(
cls,
app_model: App,
invoke_from: InvokeFrom,
@@ -533,6 +533,17 @@ class AppGenerateService:
return workflow
@classmethod
def _get_workflow(
cls,
app_model: App,
invoke_from: InvokeFrom,
workflow_id: str | None = None,
*,
session: Session,
) -> Workflow:
return cls.get_workflow(app_model, invoke_from, workflow_id, session=session)
@classmethod
def get_response_generator(
cls,
+84
View File
@@ -1,6 +1,7 @@
import json
import logging
from collections.abc import Sequence
from dataclasses import dataclass
from datetime import datetime
from typing import Any, Literal, NotRequired, TypedDict, cast, override
@@ -41,6 +42,20 @@ from tasks.remove_app_and_related_data_task import remove_app_and_related_data_t
logger = logging.getLogger(__name__)
AppListSortBy = Literal["last_modified", "recently_created", "earliest_created"]
RecentAppMode = Literal[
AppMode.COMPLETION,
AppMode.WORKFLOW,
AppMode.CHAT,
AppMode.ADVANCED_CHAT,
AppMode.AGENT_CHAT,
]
RECENT_APP_MODES: tuple[RecentAppMode, ...] = (
AppMode.COMPLETION,
AppMode.WORKFLOW,
AppMode.CHAT,
AppMode.ADVANCED_CHAT,
AppMode.AGENT_CHAT,
)
class AppListBaseParams(BaseModel):
@@ -65,6 +80,19 @@ class StarredAppListParams(AppListBaseParams):
pass
@dataclass(frozen=True)
class RecentAppListItem:
id: str
name: str
icon_type: IconType | None
icon: str | None
icon_background: str | None
mode: RecentAppMode
author_name: str | None
updated_at: datetime
maintainer: str | None
class CreateAppParams(BaseModel):
name: str = Field(min_length=1)
description: str | None = None
@@ -323,6 +351,62 @@ class AppService:
return app_models
def get_recent_apps(
self,
user_id: str,
tenant_id: str,
params: AppListParams,
session: Session,
) -> list[RecentAppListItem]:
"""Return recently modified apps as one lightweight, non-paginated projection."""
filters = self._build_app_list_filters(user_id, tenant_id, params, session)
if not filters:
return []
stmt = (
sa.select(
App.id,
App.name,
App.icon_type,
App.icon,
App.icon_background,
App.mode,
Account.name.label("author_name"),
App.updated_at,
App.maintainer,
)
.outerjoin(Account, Account.id == App.created_by)
.where(*filters, App.mode.in_(RECENT_APP_MODES))
.order_by(App.updated_at.desc())
.limit(params.limit)
)
rows = session.execute(stmt).all()
return [
RecentAppListItem(
id=str(app_id),
name=name,
icon_type=icon_type,
icon=icon,
icon_background=icon_background,
mode=cast(RecentAppMode, mode),
author_name=author_name,
updated_at=updated_at,
maintainer=maintainer,
)
for (
app_id,
name,
icon_type,
icon,
icon_background,
mode,
author_name,
updated_at,
maintainer,
) in rows
]
def get_paginate_starred_apps(
self,
user_id: str,
+48 -9
View File
@@ -1,6 +1,8 @@
import logging
from collections.abc import Mapping
from enum import StrEnum
from pydantic import BaseModel, ConfigDict, Field
from pydantic import BaseModel, ConfigDict, Field, ValidationError
from configs import dify_config
from constants.dsl_version import CURRENT_APP_DSL_VERSION
@@ -10,6 +12,8 @@ from enums.hosted_provider import HostedTrialProvider
from services.billing_service import BillingInfo, BillingService
from services.enterprise.enterprise_service import EnterpriseService
logger = logging.getLogger(__name__)
class FeatureResponseModel(BaseModel):
model_config = ConfigDict(json_schema_serialization_defaults_required=True, protected_namespaces=())
@@ -131,6 +135,13 @@ class PluginInstallationPermissionModel(FeatureResponseModel):
restrict_to_marketplace_only: bool = False
class _EnterprisePluginInstallationPermission(BaseModel):
model_config = ConfigDict(extra="ignore")
plugin_installation_scope: PluginInstallationScope = Field(alias="pluginInstallationScope")
restrict_to_marketplace_only: bool = Field(alias="restrictToMarketplaceOnly", strict=True)
class FeatureModel(FeatureResponseModel):
billing: BillingModel = BillingModel()
education: EducationModel = EducationModel()
@@ -285,6 +296,14 @@ class FeatureService:
"""Return whether Enterprise plugin credential policies must be enforced."""
return dify_config.ENTERPRISE_ENABLED
@classmethod
def get_plugin_installation_permission(cls) -> PluginInstallationPermissionModel:
"""Resolve the validated deployment-wide plugin installation policy."""
if not dify_config.ENTERPRISE_ENABLED:
return PluginInstallationPermissionModel()
return cls._resolve_plugin_installation_permission(EnterpriseService.get_info())
@classmethod
def get_license(cls) -> LicenseModel:
"""Return full license detail. Enterprise-only; requires an authenticated caller.
@@ -452,6 +471,33 @@ class FeatureService:
)
return license_model
@classmethod
def _resolve_plugin_installation_permission(
cls, enterprise_info: Mapping[str, object]
) -> PluginInstallationPermissionModel:
if "PluginInstallationPermission" not in enterprise_info:
return PluginInstallationPermissionModel()
try:
permission = _EnterprisePluginInstallationPermission.model_validate(
enterprise_info["PluginInstallationPermission"]
)
except ValidationError as exc:
# Do not attach the exception because it may contain raw Enterprise configuration values.
logger.error( # noqa: TRY400
"Invalid Enterprise plugin installation permission; denying all plugin installations: %s",
exc.errors(include_input=False),
)
return PluginInstallationPermissionModel(
plugin_installation_scope=PluginInstallationScope.NONE,
restrict_to_marketplace_only=True,
)
return PluginInstallationPermissionModel(
plugin_installation_scope=permission.plugin_installation_scope,
restrict_to_marketplace_only=permission.restrict_to_marketplace_only,
)
@classmethod
def _fulfill_params_from_enterprise(cls, features: SystemFeatureModel):
enterprise_info = EnterpriseService.get_info()
@@ -499,11 +545,4 @@ class FeatureService:
status=LicenseStatus(license_info.get("status", LicenseStatus.INACTIVE))
)
if "PluginInstallationPermission" in enterprise_info:
plugin_installation_info = enterprise_info["PluginInstallationPermission"]
features.plugin_installation_permission.plugin_installation_scope = plugin_installation_info[
"pluginInstallationScope"
]
features.plugin_installation_permission.restrict_to_marketplace_only = plugin_installation_info[
"restrictToMarketplaceOnly"
]
features.plugin_installation_permission = cls._resolve_plugin_installation_permission(enterprise_info)
+2 -1
View File
@@ -18,6 +18,7 @@ from core.app.apps.completion.app_config_manager import CompletionAppConfigManag
from core.helper import encrypter
from core.prompt.simple_prompt_transform import SimplePromptTransform
from core.prompt.utils.prompt_template_parser import PromptTemplateParser
from core.workflow.variable_prefixes import USER_INPUT_VARIABLE_NODE_ID
from events.app_event import app_was_created
from graphon.file import FileUploadConfig
from graphon.model_runtime.entities.llm_entities import LLMMode
@@ -597,7 +598,7 @@ class WorkflowConverter:
},
"vision": {
"enabled": file_upload is not None,
"variable_selector": ["sys", "files"] if file_upload is not None else None,
"variable_selector": [USER_INPUT_VARIABLE_NODE_ID, "files"] if file_upload is not None else None,
"configs": {"detail": file_upload.image_config.detail}
if file_upload is not None and file_upload.image_config is not None
else None,
+55 -2
View File
@@ -6,8 +6,10 @@ from collections.abc import Callable, Generator, Mapping, Sequence
from dataclasses import dataclass
from typing import Any, cast
from sqlalchemy import exists, select
from sqlalchemy import exists, inspect, select, update
from sqlalchemy.exc import SQLAlchemyError
from sqlalchemy.orm import Session, sessionmaker
from sqlalchemy.orm.attributes import set_committed_value
from configs import dify_config
from core.app.apps.advanced_chat.app_config_manager import AdvancedChatAppConfigManager
@@ -143,6 +145,8 @@ from .human_input_delivery_test_service import (
from .workflow_draft_variable_service import DraftVariableSaver, DraftVarLoader, WorkflowDraftVariableService
from .workflow_restore import apply_published_workflow_snapshot_to_draft
logger = logging.getLogger(__name__)
_file_access_controller = DatabaseFileAccessController()
@@ -155,6 +159,7 @@ class WorkflowService:
"""Initialize WorkflowService with repository dependencies."""
if session_maker is None:
session_maker = sessionmaker(bind=db.engine, expire_on_commit=False)
self._session_maker = session_maker
self._node_execution_service_repo = DifyAPIRepositoryFactory.create_api_workflow_node_execution_repository(
session_maker
)
@@ -211,7 +216,7 @@ class WorkflowService:
)
# return draft workflow
return workflow
return self._persist_legacy_sys_files_migration_on_load(workflow)
def get_published_workflow_by_id(self, app_model: App, workflow_id: str, *, session: Session) -> Workflow | None:
"""
@@ -236,6 +241,7 @@ class WorkflowService:
f"Cannot use draft workflow version. Workflow ID: {workflow_id}. "
f"Please use a published workflow version or leave workflow_id empty."
)
self._persist_legacy_sys_files_migration_on_load(workflow)
return workflow
def get_published_workflow(self, app_model: App, *, session: Session) -> Workflow | None:
@@ -259,6 +265,53 @@ class WorkflowService:
.limit(1)
)
return self._persist_legacy_sys_files_migration_on_load(workflow)
def _persist_legacy_sys_files_migration_on_load(self, workflow: Workflow | None) -> Workflow | None:
"""Persist a load-time graph rewrite without joining or dirtying the caller's transaction."""
if workflow is None:
return None
if inspect(workflow, raiseerr=False) is None:
return workflow
# TODO: Remove this load-time persistence path after the historical workflow migration is complete.
original_graph = workflow.graph
if not workflow.migrate_legacy_sys_files_graph_in_place():
return workflow
migrated_graph = workflow.graph
try:
with self._session_maker.begin() as session:
result = session.execute(
update(Workflow)
.where(
Workflow.id == workflow.id,
Workflow.tenant_id == workflow.tenant_id,
Workflow.graph == original_graph,
)
.values(graph=migrated_graph)
)
if getattr(result, "rowcount", None) == 0:
logger.warning(
"Skipped persisting legacy sys.files workflow migration because the workflow changed "
"concurrently, "
"workflow_id=%s tenant_id=%s",
workflow.id,
workflow.tenant_id,
)
except SQLAlchemyError:
logger.warning(
"Failed to persist legacy sys.files workflow migration, workflow_id=%s tenant_id=%s",
workflow.id,
workflow.tenant_id,
exc_info=True,
)
finally:
# The conditional update owns persistence. Mark the caller's instance clean so its later flush cannot
# overwrite a concurrent workflow edit with the compatibility rewrite.
set_committed_value(workflow, "graph", migrated_graph)
return workflow
def get_accessible_app_ids(self, app_ids: Sequence[str], tenant_id: str, *, session: Session) -> set[str]:
@@ -22,6 +22,7 @@ def _create_agent_backend_client():
return None
return create_agent_backend_run_client(
base_url=dify_config.AGENT_BACKEND_BASE_URL,
api_token=dify_config.AGENT_BACKEND_API_TOKEN,
use_fake=dify_config.AGENT_BACKEND_USE_FAKE,
fake_scenario=dify_config.AGENT_BACKEND_FAKE_SCENARIO,
stream_read_timeout_seconds=dify_config.AGENT_BACKEND_STREAM_READ_TIMEOUT_SECONDS,
@@ -457,7 +457,7 @@ def _publish_streaming_response(
@shared_task(queue=WORKFLOW_BASED_APP_EXECUTION_QUEUE)
def workflow_based_app_execution_task(
payload: str,
) -> Generator[Mapping[str, Any] | str, None, None] | Mapping[str, Any] | None:
) -> Mapping[str, Any] | None:
exec_params = AppExecutionParams.model_validate_json(payload)
logger.info("workflow_based_app_execution_task run with params: %s", exec_params)
+1 -1
View File
@@ -61,7 +61,7 @@ workflow:
query_prompt_template: '{{#sys.query#}}
{{#sys.files#}}'
{{#userinput.files#}}'
window:
enabled: false
size: 10
@@ -162,7 +162,7 @@ workflow:
query_prompt_template: '{{#sys.query#}}
{{#sys.files#}}'
{{#userinput.files#}}'
role_prefix:
assistant: ''
user: ''
@@ -207,7 +207,7 @@ workflow:
query_prompt_template: '{{#sys.query#}}
{{#sys.files#}}'
{{#userinput.files#}}'
role_prefix:
assistant: ''
user: ''
+1 -1
View File
@@ -178,7 +178,7 @@ workflow:
query_prompt_template: '{{#sys.query#}}
{{#sys.files#}}'
{{#userinput.files#}}'
role_prefix:
assistant: ''
user: ''
@@ -1,494 +1,85 @@
"""Testcontainers integration tests for controllers.console.datasets.data_source endpoints."""
"""Integration coverage for Notion page bindings backed by persisted documents."""
from __future__ import annotations
import inspect
from collections.abc import Iterator
from datetime import UTC, datetime
from unittest.mock import MagicMock, PropertyMock, patch
from inspect import unwrap
from unittest.mock import MagicMock, patch
from uuid import uuid4
import pytest
from flask import Flask
from sqlalchemy.orm import Session
from werkzeug.exceptions import NotFound
from controllers.console.datasets import data_source
from controllers.console.datasets.data_source import (
DataSourceApi,
DataSourceNotionDatasetSyncApi,
DataSourceNotionDocumentSyncApi,
DataSourceNotionIndexingEstimateApi,
DataSourceNotionListApi,
DataSourceNotionPreviewApi,
)
from core.rag.index_processor.constant.index_type import IndexStructureType
from models import Account, DataSourceOauthBinding
from controllers.console.datasets.data_source import DataSourceNotionListApi
from models import Account
from models.dataset import Document
from models.enums import DataSourceType, DocumentCreatedFrom, IndexingStatus
@pytest.fixture
def current_user() -> Account:
account = Account(name="Test User", email="u1@example.com")
account.id = "u1"
return account
def test_notion_page_is_marked_bound_from_persisted_document(
flask_app_with_containers: Flask,
db_session_with_containers: Session,
) -> None:
tenant_id = str(uuid4())
dataset_id = str(uuid4())
account = Account(name="Test User", email="user@example.com")
account.id = str(uuid4())
document = Document(
tenant_id=tenant_id,
dataset_id=dataset_id,
position=1,
data_source_type=DataSourceType.NOTION_IMPORT,
data_source_info='{"notion_page_id": "page-1"}',
batch=f"batch-{uuid4()}",
name="Notion Page",
created_from=DocumentCreatedFrom.WEB,
created_by=str(uuid4()),
indexing_status=IndexingStatus.COMPLETED,
enabled=True,
)
db_session_with_containers.add(document)
db_session_with_containers.commit()
runtime = MagicMock(
get_online_document_pages=lambda **_kwargs: iter(
[
MagicMock(
result=[
MagicMock(
workspace_id="workspace-1",
workspace_name="Workspace",
workspace_icon=None,
pages=[
MagicMock(
page_id="page-1",
page_name="Page",
type="page",
parent_id="parent",
page_icon=None,
)
],
)
]
)
]
),
datasource_provider_type=lambda: None,
)
@pytest.fixture
def mock_engine() -> Iterator[None]:
with patch.object(
type(data_source.db),
"engine",
new_callable=PropertyMock,
return_value=MagicMock(),
with (
flask_app_with_containers.test_request_context(f"/?credential_id=c1&dataset_id={dataset_id}"),
patch(
"controllers.console.datasets.data_source.DatasourceProviderService.get_datasource_credentials",
return_value={"token": "token"},
),
patch(
"controllers.console.datasets.data_source.DatasetService.get_dataset",
return_value=MagicMock(data_source_type="notion_import"),
),
patch(
"core.datasource.datasource_manager.DatasourceManager.get_datasource_runtime",
return_value=runtime,
),
):
yield
class TestDataSourceApi:
@pytest.fixture
def app(self, flask_app_with_containers: Flask) -> Flask:
return flask_app_with_containers
def test_get_success(self, app: Flask) -> None:
api = DataSourceApi()
method = inspect.unwrap(api.get)
binding = DataSourceOauthBinding(
tenant_id="tenant-1",
access_token="token",
provider="notion",
source_info={
"workspace_name": "Workspace",
"workspace_id": "workspace-1",
"workspace_icon": None,
"total": 1,
"pages": [
{
"page_id": "page-1",
"page_name": "Page",
"page_icon": {"type": "emoji", "emoji": "P", "url": None},
"parent_id": "parent-1",
"type": "page",
}
],
},
)
binding.id = "b1"
binding.created_at = datetime(2026, 5, 25, 1, 2, 3, tzinfo=UTC)
binding.disabled = False
with (
app.test_request_context("/"),
patch(
"controllers.console.datasets.data_source.db.session.scalars",
return_value=MagicMock(all=lambda: [binding]),
),
):
response, status = method(api, "tenant-1")
assert status == 200
assert response["data"][0] == {
"id": "b1",
"provider": "notion",
"created_at": 1779670923,
"is_bound": True,
"disabled": False,
"source_info": {
"workspace_name": "Workspace",
"workspace_id": "workspace-1",
"workspace_icon": None,
"pages": [
{
"page_name": "Page",
"page_id": "page-1",
"page_icon": {"type": "emoji", "url": None, "emoji": "P"},
"parent_id": "parent-1",
"type": "page",
}
],
"total": 1,
},
"link": "http://localhost/console/api/oauth/data-source/notion",
}
def test_get_no_bindings(self, app: Flask) -> None:
api = DataSourceApi()
method = inspect.unwrap(api.get)
with (
app.test_request_context("/"),
patch(
"controllers.console.datasets.data_source.db.session.scalars",
return_value=MagicMock(all=lambda: []),
),
):
response, status = method(api, "tenant-1")
assert status == 200
assert response["data"] == []
def test_patch_enable_binding(self, app: Flask) -> None:
api = DataSourceApi()
method = inspect.unwrap(api.patch)
binding = MagicMock(id="b1", disabled=True)
session = MagicMock()
session.scalar.return_value = binding
with app.test_request_context("/"):
response, status = method(api, session, "tenant-1", "b1", "enable")
assert status == 200
assert binding.disabled is False
def test_patch_disable_binding(self, app: Flask) -> None:
api = DataSourceApi()
method = inspect.unwrap(api.patch)
binding = MagicMock(id="b1", disabled=False)
session = MagicMock()
session.scalar.return_value = binding
with app.test_request_context("/"):
response, status = method(api, session, "tenant-1", "b1", "disable")
assert status == 200
assert binding.disabled is True
def test_patch_binding_not_found(self, app: Flask) -> None:
api = DataSourceApi()
method = inspect.unwrap(api.patch)
session = MagicMock()
session.scalar.return_value = None
with app.test_request_context("/"):
with pytest.raises(NotFound):
method(api, session, "tenant-1", "b1", "enable")
def test_patch_enable_already_enabled(self, app: Flask) -> None:
api = DataSourceApi()
method = inspect.unwrap(api.patch)
binding = MagicMock(id="b1", disabled=False)
session = MagicMock()
session.scalar.return_value = binding
with app.test_request_context("/"):
with pytest.raises(ValueError):
method(api, session, "tenant-1", "b1", "enable")
def test_patch_disable_already_disabled(self, app: Flask) -> None:
api = DataSourceApi()
method = inspect.unwrap(api.patch)
binding = MagicMock(id="b1", disabled=True)
session = MagicMock()
session.scalar.return_value = binding
with app.test_request_context("/"):
with pytest.raises(ValueError):
method(api, session, "tenant-1", "b1", "disable")
class TestDataSourceNotionListApi:
@pytest.fixture
def app(self, flask_app_with_containers: Flask) -> Flask:
return flask_app_with_containers
def test_get_credential_not_found(self, app: Flask, current_user: Account) -> None:
api = DataSourceNotionListApi()
method = inspect.unwrap(api.get)
with (
app.test_request_context("/?credential_id=c1"),
patch(
"controllers.console.datasets.data_source.DatasourceProviderService.get_datasource_credentials",
return_value=None,
),
):
with pytest.raises(NotFound):
method(api, MagicMock(), "tenant-1", current_user)
def test_get_success_no_dataset_id(self, app: Flask, current_user: Account, mock_engine: None) -> None:
api = DataSourceNotionListApi()
method = inspect.unwrap(api.get)
page = MagicMock(
page_id="p1",
page_name="Page 1",
type="page",
parent_id="parent",
page_icon=None,
response, status = unwrap(DataSourceNotionListApi().get)(
DataSourceNotionListApi(), db_session_with_containers, tenant_id, account
)
online_document_message = MagicMock(
result=[
MagicMock(
workspace_id="w1",
workspace_name="My Workspace",
workspace_icon="icon",
pages=[page],
)
]
)
with (
app.test_request_context("/?credential_id=c1"),
patch(
"controllers.console.datasets.data_source.DatasourceProviderService.get_datasource_credentials",
return_value={"token": "t"},
),
patch(
"core.datasource.datasource_manager.DatasourceManager.get_datasource_runtime",
return_value=MagicMock(
get_online_document_pages=lambda **kw: iter([online_document_message]),
datasource_provider_type=lambda: None,
),
),
):
response, status = method(api, MagicMock(), "tenant-1", current_user)
assert status == 200
def test_get_success_with_dataset_id(
self, app: Flask, current_user: Account, mock_engine: None, db_session_with_containers: Session
) -> None:
api = DataSourceNotionListApi()
method = inspect.unwrap(api.get)
tenant_id = str(uuid4())
dataset_id = str(uuid4())
page = MagicMock(
page_id="p1",
page_name="Page 1",
type="page",
parent_id="parent",
page_icon=None,
)
online_document_message = MagicMock(
result=[
MagicMock(
workspace_id="w1",
workspace_name="My Workspace",
workspace_icon="icon",
pages=[page],
)
]
)
dataset = MagicMock(data_source_type="notion_import")
document = Document(
tenant_id=tenant_id,
dataset_id=dataset_id,
position=1,
data_source_type=DataSourceType.NOTION_IMPORT,
data_source_info='{"notion_page_id": "p1"}',
batch=f"batch-{uuid4()}",
name="Notion Page",
created_from=DocumentCreatedFrom.WEB,
created_by=str(uuid4()),
indexing_status=IndexingStatus.COMPLETED,
enabled=True,
)
db_session_with_containers.add(document)
db_session_with_containers.commit()
with (
app.test_request_context(f"/?credential_id=c1&dataset_id={dataset_id}"),
patch(
"controllers.console.datasets.data_source.DatasourceProviderService.get_datasource_credentials",
return_value={"token": "t"},
),
patch(
"controllers.console.datasets.data_source.DatasetService.get_dataset",
return_value=dataset,
),
patch(
"core.datasource.datasource_manager.DatasourceManager.get_datasource_runtime",
return_value=MagicMock(
get_online_document_pages=lambda **kw: iter([online_document_message]),
datasource_provider_type=lambda: None,
),
),
):
response, status = method(api, db_session_with_containers, tenant_id, current_user)
assert status == 200
def test_get_invalid_dataset_type(self, app: Flask, current_user: Account, mock_engine: None) -> None:
api = DataSourceNotionListApi()
method = inspect.unwrap(api.get)
dataset = MagicMock(data_source_type="other_type")
with (
app.test_request_context("/?credential_id=c1&dataset_id=ds1"),
patch(
"controllers.console.datasets.data_source.DatasourceProviderService.get_datasource_credentials",
return_value={"token": "t"},
),
patch(
"controllers.console.datasets.data_source.DatasetService.get_dataset",
return_value=dataset,
),
):
with pytest.raises(ValueError):
method(api, MagicMock(), "tenant-1", current_user)
class TestDataSourceNotionPreviewApi:
@pytest.fixture
def app(self, flask_app_with_containers: Flask) -> Flask:
return flask_app_with_containers
def test_get_preview_success(self, app: Flask) -> None:
api = DataSourceNotionPreviewApi()
method = inspect.unwrap(api.get)
extractor = MagicMock(extract=lambda: [MagicMock(page_content="hello")])
with (
app.test_request_context("/?credential_id=c1"),
patch(
"controllers.console.datasets.data_source.DatasourceProviderService.get_datasource_credentials",
return_value={"integration_secret": "t"},
),
patch(
"controllers.console.datasets.data_source.NotionExtractor",
return_value=extractor,
),
):
response, status = method(api, "tenant-1", "p1", "page")
assert status == 200
class TestDataSourceNotionIndexingEstimateApi:
@pytest.fixture
def app(self, flask_app_with_containers: Flask) -> Flask:
return flask_app_with_containers
def test_post_indexing_estimate_success(self, app: Flask) -> None:
api = DataSourceNotionIndexingEstimateApi()
method = inspect.unwrap(api.post)
empty_rules: dict[str, object] = {}
payload: dict[str, object] = {
"notion_info_list": [
{
"workspace_id": "w1",
"credential_id": "c1",
"pages": [{"page_id": "p1", "type": "page"}],
}
],
"process_rule": {"rules": empty_rules},
"doc_form": IndexStructureType.PARAGRAPH_INDEX,
"doc_language": "English",
}
with (
app.test_request_context("/", method="POST", json=payload, headers={"Content-Type": "application/json"}),
patch(
"controllers.console.datasets.data_source.DocumentService.estimate_args_validate",
),
patch(
"controllers.console.datasets.data_source.IndexingRunner.indexing_estimate",
return_value=MagicMock(model_dump=lambda: {"total_pages": 1}),
),
):
response, status = method(api, MagicMock(), "tenant-1")
assert status == 200
class TestDataSourceNotionDatasetSyncApi:
@pytest.fixture
def app(self, flask_app_with_containers: Flask) -> Flask:
return flask_app_with_containers
def test_get_success(self, app: Flask) -> None:
api = DataSourceNotionDatasetSyncApi()
method = inspect.unwrap(api.get)
with (
app.test_request_context("/"),
patch(
"controllers.console.datasets.data_source.DatasetService.get_dataset",
return_value=MagicMock(),
),
patch(
"controllers.console.datasets.data_source.DocumentService.get_document_by_dataset_id",
return_value=[MagicMock(id="d1")],
),
patch(
"controllers.console.datasets.data_source.document_indexing_sync_task.delay",
return_value=None,
),
):
response, status = method(api, MagicMock(), "ds-1")
assert status == 200
def test_get_dataset_not_found(self, app: Flask) -> None:
api = DataSourceNotionDatasetSyncApi()
method = inspect.unwrap(api.get)
with (
app.test_request_context("/"),
patch(
"controllers.console.datasets.data_source.DatasetService.get_dataset",
return_value=None,
),
):
with pytest.raises(NotFound):
method(api, MagicMock(), "ds-1")
class TestDataSourceNotionDocumentSyncApi:
@pytest.fixture
def app(self, flask_app_with_containers: Flask) -> Flask:
return flask_app_with_containers
def test_get_success(self, app: Flask) -> None:
api = DataSourceNotionDocumentSyncApi()
method = inspect.unwrap(api.get)
with (
app.test_request_context("/"),
patch(
"controllers.console.datasets.data_source.DatasetService.get_dataset",
return_value=MagicMock(),
),
patch(
"controllers.console.datasets.data_source.DocumentService.get_document",
return_value=MagicMock(),
),
patch(
"controllers.console.datasets.data_source.document_indexing_sync_task.delay",
return_value=None,
),
):
response, status = method(api, MagicMock(), "ds-1", "doc-1")
assert status == 200
def test_get_document_not_found(self, app: Flask) -> None:
api = DataSourceNotionDocumentSyncApi()
method = inspect.unwrap(api.get)
with (
app.test_request_context("/"),
patch(
"controllers.console.datasets.data_source.DatasetService.get_dataset",
return_value=MagicMock(),
),
patch(
"controllers.console.datasets.data_source.DocumentService.get_document",
return_value=None,
),
):
with pytest.raises(NotFound):
method(api, MagicMock(), "ds-1", "doc-1")
assert status == 200
assert response["notion_info"][0]["pages"][0]["is_bound"] is True
@@ -4,7 +4,7 @@ import datetime
import json
import uuid
from decimal import Decimal
from unittest.mock import MagicMock, patch
from unittest.mock import patch
import pytest
from faker import Faker
@@ -1172,65 +1172,8 @@ class TestMessagesCleanServiceIntegration:
# Verify all messages were deleted
assert db_session_with_containers.query(Message).where(Message.id.in_(msg_ids)).count() == 0
def test_from_time_range_validation(self):
"""Test that from_time_range raises ValueError for invalid inputs."""
policy = MagicMock(spec=BillingDisabledPolicy)
now = datetime.datetime.now()
with pytest.raises(ValueError, match="start_from .* must be less than end_before"):
MessagesCleanService.from_time_range(policy, now, now)
with pytest.raises(ValueError, match="batch_size .* must be greater than 0"):
MessagesCleanService.from_time_range(policy, now - datetime.timedelta(days=1), now, batch_size=0)
def test_from_time_range_success(self):
"""Test that from_time_range creates a service with correct parameters."""
policy = MagicMock(spec=BillingDisabledPolicy)
start = datetime.datetime(2024, 1, 1)
end = datetime.datetime(2024, 2, 1)
service = MessagesCleanService.from_time_range(policy, start, end)
assert service._start_from == start
assert service._end_before == end
def test_from_days_validation(self):
"""Test that from_days raises ValueError for invalid inputs."""
policy = MagicMock(spec=BillingDisabledPolicy)
with pytest.raises(ValueError, match="days .* must be greater than or equal to 0"):
MessagesCleanService.from_days(policy, days=-1)
with pytest.raises(ValueError, match="batch_size .* must be greater than 0"):
MessagesCleanService.from_days(policy, days=30, batch_size=0)
def test_from_days_success(self):
"""Test that from_days creates a service with correct parameters."""
policy = MagicMock(spec=BillingDisabledPolicy)
with patch("services.retention.conversation.messages_clean_service.naive_utc_now") as mock_now:
fixed_now = datetime.datetime(2024, 6, 1)
mock_now.return_value = fixed_now
service = MessagesCleanService.from_days(policy, days=10)
assert service._start_from is None
assert service._end_before == fixed_now - datetime.timedelta(days=10)
def test_batch_delete_message_relations_empty(self, db_session_with_containers: Session):
"""Test that batch_delete_message_relations with empty list does nothing."""
# Get execute call count before
MessagesCleanService._batch_delete_message_relations(db_session_with_containers, [])
# No exception means success — empty list is a no-op
def test_run_calls_clean_messages(self):
"""Test that run() delegates to _clean_messages_by_time_range."""
policy = MagicMock(spec=BillingDisabledPolicy)
service = MessagesCleanService(
policy=policy,
end_before=datetime.datetime.now(),
batch_size=10,
)
with patch.object(service, "_clean_messages_by_time_range") as mock_clean:
mock_clean.return_value = {"total_deleted": 5}
result = service.run()
assert result == {"total_deleted": 5}
mock_clean.assert_called_once()
@@ -4,7 +4,7 @@ from unittest.mock import MagicMock
import pytest
from sqlalchemy.orm import Session
from commands import system as system_commands
from commands import app_maintenance as app_maintenance_commands
def test_fix_app_site_missing_passes_loaded_session_to_signal(monkeypatch: pytest.MonkeyPatch) -> None:
@@ -30,15 +30,15 @@ def test_fix_app_site_missing_passes_loaded_session_to_signal(monkeypatch: pytes
engine = MagicMock()
engine.begin.return_value.__enter__.return_value = connection
monkeypatch.setattr(system_commands, "db", SimpleNamespace(engine=engine, session=scoped_session))
monkeypatch.setattr(app_maintenance_commands, "db", SimpleNamespace(engine=engine, session=scoped_session))
send = MagicMock(side_effect=lambda *_args, **_kwargs: phase_events.append("signal"))
monkeypatch.setattr(system_commands.app_was_created, "send", send)
monkeypatch.setattr(app_maintenance_commands.app_was_created, "send", send)
system_commands.fix_app_site_missing.callback()
app_maintenance_commands.fix_app_site_missing.callback()
scoped_session.assert_called_once_with()
scalar.assert_called_once()
get.assert_called_once_with(system_commands.Tenant, app.tenant_id)
get.assert_called_once_with(app_maintenance_commands.Tenant, app.tenant_id)
tenant.get_accounts.assert_called_once_with(session=session)
send.assert_called_once_with(app, account=account, session=session)
commit.assert_called_once_with()
@@ -62,15 +62,19 @@ def test_fix_app_site_missing_rolls_back_when_signal_fails(monkeypatch: pytest.M
engine = MagicMock()
engine.begin.return_value.__enter__.return_value = connection
monkeypatch.setattr(system_commands, "db", SimpleNamespace(engine=engine, session=MagicMock(return_value=session)))
monkeypatch.setattr(
app_maintenance_commands,
"db",
SimpleNamespace(engine=engine, session=MagicMock(return_value=session)),
)
def fail_signal(*_args, **_kwargs) -> None:
phase_events.append("signal")
raise RuntimeError("failed")
monkeypatch.setattr(system_commands.app_was_created, "send", MagicMock(side_effect=fail_signal))
monkeypatch.setattr(app_maintenance_commands.app_was_created, "send", MagicMock(side_effect=fail_signal))
system_commands.fix_app_site_missing.callback()
app_maintenance_commands.fix_app_site_missing.callback()
session.rollback.assert_called_once_with()
session.commit.assert_not_called()
@@ -6,6 +6,7 @@ import json
import os
import threading
import time
from collections.abc import Iterator
from datetime import datetime, timedelta
from pathlib import Path
from types import SimpleNamespace
@@ -15,9 +16,12 @@ import pytest
import sqlalchemy as sa
from click.testing import CliRunner
from sqlalchemy.exc import OperationalError
from sqlalchemy.orm import Session, SessionTransaction, sessionmaker
from graphon.model_runtime.entities.model_entities import ModelType
from models import Dataset, DatasetPermission, DatasetPermissionEnum
from models.account import Tenant
from models.base import TypeBase
from models.enums import CredentialSourceType
from models.provider import ProviderModel
from tests.helpers.legacy_model_type_migration import (
@@ -59,6 +63,40 @@ def command_module():
)
@pytest.fixture
def rbac_session(sqlite_engine: sa.Engine, monkeypatch: pytest.MonkeyPatch) -> Iterator[Session]:
"""Bind RBAC command reads to persisted SQLite dataset rows."""
TypeBase.metadata.create_all(
sqlite_engine,
tables=[Dataset.__table__, DatasetPermission.__table__],
)
factory = sessionmaker(bind=sqlite_engine, expire_on_commit=False)
monkeypatch.setattr("commands.rbac.session_factory.create_session", factory)
with factory() as session:
yield session
def _persist_dataset(
session: Session,
*,
dataset_id: str = "dataset-1",
tenant_id: str = "tenant-1",
permission: DatasetPermissionEnum = DatasetPermissionEnum.ONLY_ME,
created_by: str = "creator-account-1",
) -> Dataset:
dataset = Dataset(
id=dataset_id,
tenant_id=tenant_id,
name=f"Dataset {dataset_id}",
permission=permission,
created_by=created_by,
)
session.add(dataset)
session.commit()
return dataset
def _parse_json_lines(output: io.StringIO) -> list[dict[str, object]]:
return [json.loads(line) for line in output.getvalue().splitlines() if line.strip()]
@@ -363,56 +401,35 @@ def test_dataset_permission_rbac_migration_maps_legacy_permissions_to_enum_scope
def test_dataset_permission_rbac_migration_uses_dataset_creator_as_operator(
command_module,
rbac_session: Session,
monkeypatch: pytest.MonkeyPatch,
) -> None:
rbac_module = importlib.import_module("commands.rbac")
dataset_row = SimpleNamespace(
id="dataset-1",
tenant_id="tenant-1",
permission="only_me",
created_by="creator-account-1",
)
execute_results = [[dataset_row], [], []]
_persist_dataset(rbac_session)
calls: list[dict[str, object]] = []
session_closed = False
class FakeExecuteResult:
def __init__(self, rows: list[object]) -> None:
self._rows = rows
def all(self) -> list[object]:
return self._rows
class FakeSession:
def __enter__(self):
return self
def __exit__(self, exc_type, exc, traceback) -> None:
nonlocal session_closed
session_closed = True
pass
def execute(self, stmt):
return FakeExecuteResult(execute_results.pop(0))
class FakeSessionFactory:
@staticmethod
def create_session() -> FakeSession:
return FakeSession()
read_transaction_ended = False
def fake_replace_whitelist(**kwargs):
assert session_closed is True
assert read_transaction_ended is True
calls.append(kwargs)
monkeypatch.setattr(rbac_module, "session_factory", FakeSessionFactory)
monkeypatch.setattr(rbac_module.RBACService.DatasetAccess, "replace_whitelist", fake_replace_whitelist)
def _record_transaction_end(session: Session, transaction: object) -> None:
nonlocal read_transaction_ended
del transaction
if session.get_bind() is rbac_session.get_bind():
read_transaction_ended = True
command_module.migrate_dataset_permissions_to_rbac.callback(
tenant_id=None,
dataset_id=None,
batch_size=500,
dry_run=False,
)
sa.event.listen(Session, "after_transaction_end", _record_transaction_end)
monkeypatch.setattr(rbac_module.RBACService.DatasetAccess, "replace_whitelist", fake_replace_whitelist)
try:
command_module.migrate_dataset_permissions_to_rbac.callback(
tenant_id=None,
dataset_id=None,
batch_size=500,
dry_run=False,
)
finally:
sa.event.remove(Session, "after_transaction_end", _record_transaction_end)
assert calls[0]["tenant_id"] == "tenant-1"
assert calls[0]["account_id"] == "creator-account-1"
@@ -422,41 +439,19 @@ def test_dataset_permission_rbac_migration_uses_dataset_creator_as_operator(
def test_dataset_permission_rbac_migration_dry_run_outputs_structured_proposed_changes(
command_module,
rbac_session: Session,
monkeypatch: pytest.MonkeyPatch,
) -> None:
rbac_module = importlib.import_module("commands.rbac")
dataset_row = SimpleNamespace(
id="dataset-1",
tenant_id="tenant-1",
permission="partial_members",
created_by="creator-account-1",
dataset = _persist_dataset(rbac_session, permission=DatasetPermissionEnum.PARTIAL_TEAM)
rbac_session.add(
DatasetPermission(
dataset_id=dataset.id,
account_id="member-account-1",
tenant_id=dataset.tenant_id,
)
)
permission_row = SimpleNamespace(dataset_id="dataset-1", account_id="member-account-1")
execute_results = [[dataset_row], [permission_row], []]
class FakeExecuteResult:
def __init__(self, rows: list[object]) -> None:
self._rows = rows
def all(self) -> list[object]:
return self._rows
class FakeSession:
def __enter__(self):
return self
def __exit__(self, exc_type, exc, traceback) -> None:
pass
def execute(self, stmt):
return FakeExecuteResult(execute_results.pop(0))
class FakeSessionFactory:
@staticmethod
def create_session() -> FakeSession:
return FakeSession()
monkeypatch.setattr(rbac_module, "session_factory", FakeSessionFactory)
rbac_session.commit()
monkeypatch.setattr(
rbac_module.RBACService.DatasetAccess,
"replace_whitelist",
@@ -1306,50 +1301,36 @@ def test_provider_models_processing_uses_same_plan_locking_and_transaction_entry
begin_calls: list[str] = []
configure_calls: list[str] = []
class _FakeBeginContext:
def __init__(self, phase: str) -> None:
self._phase = phase
def _record_begin(session: Session, transaction: SessionTransaction) -> None:
if session.get_bind() is sqlite_engine and transaction.parent is None:
begin_calls.append(current_phase["name"])
def __enter__(self) -> None:
begin_calls.append(self._phase)
def __exit__(self, exc_type, exc, tb) -> bool:
return False
class _FakeSession:
def __init__(self, phase: str) -> None:
self._phase = phase
def __enter__(self):
return self
def __exit__(self, exc_type, exc, tb) -> bool:
return False
def begin(self) -> _FakeBeginContext:
return _FakeBeginContext(self._phase)
def _fake_session_factory(engine: sa.Engine) -> _FakeSession:
return _FakeSession(current_phase["name"])
def _fake_build_plan(self, session, candidate, *, lock_rows: bool):
def _fake_build_plan(self, session: Session, candidate, *, lock_rows: bool):
assert session.get_bind() is sqlite_engine
lock_rows_seen.append((current_phase["name"], lock_rows))
return SimpleNamespace(group_row_ids=[str(candidate.row.id)], winner=None, loser_rows=[])
return migration_module._ProviderModelGroupPlan(
group_row_ids=[str(candidate.row.id)],
winner=None,
loser_rows=[],
)
def _fake_emit_plan(self, plan, *, session, tx_id: str, business_key: dict[str, object]) -> None:
return None
def _fake_configure(self, session) -> None:
def _fake_configure(self, session: Session) -> None:
assert session.get_bind() is sqlite_engine
configure_calls.append(current_phase["name"])
monkeypatch.setattr(migration_module, "_session_factory", _fake_session_factory)
monkeypatch.setattr(migration_module.Migration, "_build_provider_model_group_plan", _fake_build_plan)
monkeypatch.setattr(migration_module.Migration, "_emit_provider_model_group_plan", _fake_emit_plan)
monkeypatch.setattr(migration_module.Migration, "_configure_lock_timeout", _fake_configure)
dry_migration._process_provider_model_group(candidate, business_key)
current_phase["name"] = "apply"
apply_migration._process_provider_model_group(candidate, business_key)
sa.event.listen(Session, "after_transaction_create", _record_begin)
try:
dry_migration._process_provider_model_group(candidate, business_key)
current_phase["name"] = "apply"
apply_migration._process_provider_model_group(candidate, business_key)
finally:
sa.event.remove(Session, "after_transaction_create", _record_begin)
assert [phase for phase, _ in lock_rows_seen] == ["dry", "apply"]
assert lock_rows_seen[0][1] == lock_rows_seen[1][1]
@@ -1392,6 +1373,22 @@ def test_process_load_balancing_model_config_row_logs_stacktrace_for_lock_timeou
sqlite_engine: sa.Engine,
monkeypatch: pytest.MonkeyPatch,
) -> None:
create_minimal_legacy_model_type_schema(sqlite_engine)
created_at = datetime(2025, 1, 1, 12, 0, 0)
_insert_load_balancing_model_config(
sqlite_engine,
row_id="40000000-0000-0000-0000-000000000001",
tenant_id="tenant-1",
provider_name="openai",
model_name="gpt-4o-mini",
model_type="text-generation",
name="credential",
encrypted_config="{}",
credential_id="50000000-0000-0000-0000-000000000001",
enabled=True,
created_at=created_at,
updated_at=created_at,
)
output = io.StringIO()
migration = migration_module.Migration(
tenant_id="tenant-1",
@@ -1401,37 +1398,18 @@ def test_process_load_balancing_model_config_row_logs_stacktrace_for_lock_timeou
model_types=(ModelType.LLM,),
orm_models=(migration_module.LoadBalancingModelConfig,),
)
candidate = migration_module._RowWithRawModelType(
row=SimpleNamespace(id="lb-row-1"),
raw_model_type="text-generation",
canonical_model_type=ModelType.LLM,
)
candidate = migration._load_load_balancing_model_config_candidates(None)[0]
lock_timeout_exc = OperationalError("SELECT 1", {}, SimpleNamespace(pgcode="55P03"))
transaction_begins = 0
class _FakeBeginContext:
def __enter__(self) -> None:
return None
def __exit__(self, exc_type, exc, tb) -> bool:
return False
class _FakeSession:
def __enter__(self):
return self
def __exit__(self, exc_type, exc, tb) -> bool:
return False
def begin(self) -> _FakeBeginContext:
return _FakeBeginContext()
def _fake_session_factory(engine: sa.Engine) -> _FakeSession:
return _FakeSession()
def _record_begin(session: Session, transaction: SessionTransaction) -> None:
nonlocal transaction_begins
if session.get_bind() is sqlite_engine and transaction.parent is None:
transaction_begins += 1
def _fake_reload(self, session, original_candidate, *, lock_rows: bool):
raise lock_timeout_exc
monkeypatch.setattr(migration_module, "_session_factory", _fake_session_factory)
monkeypatch.setattr(migration_module.Migration, "_configure_lock_timeout", lambda self, session: None)
monkeypatch.setattr(
migration_module.Migration,
@@ -1439,17 +1417,22 @@ def test_process_load_balancing_model_config_row_logs_stacktrace_for_lock_timeou
_fake_reload,
)
migration._process_load_balancing_model_config_row(candidate)
sa.event.listen(Session, "after_transaction_create", _record_begin)
try:
migration._process_load_balancing_model_config_row(candidate)
finally:
sa.event.remove(Session, "after_transaction_create", _record_begin)
lines = _parse_json_lines(output)
assert len(lines) == 1
assert lines[0]["event"] == "lock_timeout_skipped"
attrs = cast(dict[str, object], lines[0]["attrs"])
assert attrs["table_name"] == "load_balancing_model_configs"
assert attrs["id"] == "lb-row-1"
assert attrs["id"] == str(candidate.row.id)
assert attrs["error"] == str(lock_timeout_exc)
assert isinstance(attrs["stacktrace"], str)
assert "OperationalError" in attrs["stacktrace"]
assert transaction_begins == 1
def test_process_load_balancing_model_config_row_logs_update_after_sql_execution(
@@ -1457,6 +1440,23 @@ def test_process_load_balancing_model_config_row_logs_update_after_sql_execution
sqlite_engine: sa.Engine,
monkeypatch: pytest.MonkeyPatch,
) -> None:
create_minimal_legacy_model_type_schema(sqlite_engine)
created_at = datetime(2025, 1, 1, 12, 0, 0)
row_id = "40000000-0000-0000-0000-000000000002"
_insert_load_balancing_model_config(
sqlite_engine,
row_id=row_id,
tenant_id="tenant-1",
provider_name="openai",
model_name="gpt-4o-mini",
model_type="text-generation",
name="credential",
encrypted_config="{}",
credential_id="50000000-0000-0000-0000-000000000002",
enabled=True,
created_at=created_at,
updated_at=created_at,
)
migration = migration_module.Migration(
tenant_id="tenant-1",
engine=sqlite_engine,
@@ -1465,42 +1465,33 @@ def test_process_load_balancing_model_config_row_logs_update_after_sql_execution
model_types=(ModelType.LLM,),
orm_models=(migration_module.LoadBalancingModelConfig,),
)
candidate = migration_module._RowWithRawModelType(
row=SimpleNamespace(id="lb-row-1"),
raw_model_type="text-generation",
canonical_model_type=ModelType.LLM,
)
candidate = migration._load_load_balancing_model_config_candidates(None)[0]
action_log: list[str] = []
class _FakeBeginContext:
def __enter__(self) -> None:
def _record_begin(session: Session, transaction: SessionTransaction) -> None:
if session.get_bind() is sqlite_engine and transaction.parent is None:
action_log.append("begin")
def __exit__(self, exc_type, exc, tb) -> bool:
return False
class _FakeSession:
def __enter__(self):
return self
def __exit__(self, exc_type, exc, tb) -> bool:
return False
def begin(self) -> _FakeBeginContext:
return _FakeBeginContext()
def execute(self, stmt) -> None:
def _record_sql(
connection: sa.Connection,
cursor: object,
statement: str,
parameters: object,
context: object,
executemany: bool,
) -> None:
del connection, cursor, parameters, context, executemany
if statement.lstrip().upper().startswith("UPDATE"):
action_log.append("sql_execute")
def _fake_session_factory(engine: sa.Engine) -> _FakeSession:
return _FakeSession()
def _fake_configure(self, session) -> None:
action_log.append("configure_lock_timeout")
def _fake_reload(self, session, original_candidate, *, lock_rows: bool):
original_reload = migration_module.Migration._reload_load_balancing_model_config_candidate
def _record_reload(self, session: Session, original_candidate, *, lock_rows: bool):
action_log.append(f"reload_candidate:{lock_rows}")
return candidate
return original_reload(self, session, original_candidate, lock_rows=lock_rows)
def _fake_log_row_updated(self, *args, **kwargs) -> None:
action_log.append("log_row_updated")
@@ -1508,12 +1499,11 @@ def test_process_load_balancing_model_config_row_logs_update_after_sql_execution
def _fake_cache_cleanup(self, *, row_id: str, tx_id: str) -> None:
action_log.append("cache_cleanup")
monkeypatch.setattr(migration_module, "_session_factory", _fake_session_factory)
monkeypatch.setattr(migration_module.Migration, "_configure_lock_timeout", _fake_configure)
monkeypatch.setattr(
migration_module.Migration,
"_reload_load_balancing_model_config_candidate",
_fake_reload,
_record_reload,
)
monkeypatch.setattr(migration_module.Migration, "_log_row_updated", _fake_log_row_updated)
monkeypatch.setattr(
@@ -1522,7 +1512,13 @@ def test_process_load_balancing_model_config_row_logs_update_after_sql_execution
_fake_cache_cleanup,
)
migration._process_load_balancing_model_config_row(candidate)
sa.event.listen(Session, "after_transaction_create", _record_begin)
sa.event.listen(sqlite_engine, "before_cursor_execute", _record_sql)
try:
migration._process_load_balancing_model_config_row(candidate)
finally:
sa.event.remove(sqlite_engine, "before_cursor_execute", _record_sql)
sa.event.remove(Session, "after_transaction_create", _record_begin)
assert action_log == [
"begin",
@@ -1532,6 +1528,10 @@ def test_process_load_balancing_model_config_row_logs_update_after_sql_execution
"log_row_updated",
"cache_cleanup",
]
with Session(sqlite_engine) as session:
persisted = session.get(migration_module.LoadBalancingModelConfig, row_id)
assert persisted is not None
assert persisted.model_type == ModelType.LLM
def test_load_balancing_model_config_cache_delete_failure_logs_stacktrace(
@@ -0,0 +1,236 @@
from types import SimpleNamespace
from unittest.mock import MagicMock
import click
import pytest
from commands import migrate_legacy_sys_files_workflows
from commands import workflow_migration as workflow_migration_commands
def test_migrate_legacy_sys_files_workflows_command_passes_batch_options(mocker, capsys):
runner = mocker.patch.object(
workflow_migration_commands,
"run_legacy_sys_files_workflow_migration",
return_value=workflow_migration_commands.LegacySysFilesWorkflowMigrationStats(
scanned=10,
migrated=2,
failed=0,
batches=1,
last_id="workflow-10",
),
)
migrate_legacy_sys_files_workflows.callback(
batch_size=200,
limit=500,
start_after_id="workflow-1",
tenant_id="tenant-1",
app_id="app-1",
dry_run=True,
)
runner.assert_called_once_with(
batch_size=200,
limit=500,
start_after_id="workflow-1",
tenant_id="tenant-1",
app_id="app-1",
dry_run=True,
)
captured = capsys.readouterr()
assert "scanned=10" in captured.out
assert "migrated=2" in captured.out
assert "last_id=workflow-10" in captured.out
def test_migrate_legacy_sys_files_workflows_rejects_non_positive_batch_size():
with pytest.raises(click.UsageError, match="batch-size"):
migrate_legacy_sys_files_workflows.callback(
batch_size=0,
limit=None,
start_after_id=None,
tenant_id=None,
app_id=None,
dry_run=False,
)
def test_migrate_legacy_sys_files_workflows_rejects_non_positive_limit():
with pytest.raises(click.UsageError, match="limit"):
migrate_legacy_sys_files_workflows.callback(
batch_size=100,
limit=0,
start_after_id=None,
tenant_id=None,
app_id=None,
dry_run=False,
)
def test_build_legacy_sys_files_workflow_query_uses_keyset_pagination():
stmt = workflow_migration_commands._build_legacy_sys_files_workflow_query(
start_after_id="workflow-1",
batch_size=200,
tenant_id="tenant-1",
app_id="app-1",
)
compiled = str(stmt.compile(compile_kwargs={"literal_binds": True}))
assert "workflows.id > 'workflow-1'" in compiled
assert "workflows.tenant_id = 'tenant-1'" in compiled
assert "workflows.app_id = 'app-1'" in compiled
assert "ORDER BY workflows.id" in compiled
assert "LIMIT 200" in compiled
assert "workflows.environment_variables" not in compiled
def test_migrate_legacy_sys_files_workflow_batch_dry_run_rolls_back():
migrated_workflow = MagicMock()
migrated_workflow.id = "workflow-1"
migrated_workflow.migrate_legacy_sys_files_graph_in_place.return_value = True
untouched_workflow = MagicMock()
untouched_workflow.id = "workflow-2"
untouched_workflow.migrate_legacy_sys_files_graph_in_place.return_value = False
session = MagicMock()
session.scalars.return_value.all.return_value = [migrated_workflow, untouched_workflow]
stats = workflow_migration_commands._migrate_legacy_sys_files_workflow_batch(
session=session,
start_after_id=None,
batch_size=200,
tenant_id=None,
app_id=None,
dry_run=True,
)
assert stats.scanned == 2
assert stats.migrated == 1
assert stats.failed == 0
assert stats.last_id == "workflow-2"
session.rollback.assert_called_once()
session.commit.assert_not_called()
def test_migrate_legacy_sys_files_workflow_batch_commits_and_counts_failures(caplog):
migrated_workflow = MagicMock()
migrated_workflow.id = "workflow-1"
migrated_workflow.migrate_legacy_sys_files_graph_in_place.return_value = True
failing_workflow = MagicMock()
failing_workflow.id = "workflow-2"
failing_workflow.migrate_legacy_sys_files_graph_in_place.side_effect = RuntimeError("boom")
session = MagicMock()
session.scalars.return_value.all.return_value = [migrated_workflow, failing_workflow]
stats = workflow_migration_commands._migrate_legacy_sys_files_workflow_batch(
session=session,
start_after_id=None,
batch_size=200,
tenant_id=None,
app_id=None,
dry_run=False,
)
assert stats.scanned == 2
assert stats.migrated == 1
assert stats.failed == 1
assert stats.last_id == "workflow-2"
assert "Failed to migrate legacy" in caplog.text
session.commit.assert_called_once()
session.rollback.assert_not_called()
def test_run_legacy_sys_files_workflow_migration_uses_keyset_batches(mocker):
session_maker = MagicMock()
sessions = [MagicMock(), MagicMock()]
session_maker.side_effect = sessions
mocker.patch.object(workflow_migration_commands, "sessionmaker", return_value=session_maker)
mocker.patch.object(workflow_migration_commands, "db", SimpleNamespace(engine=object()))
migrate_batch = mocker.patch.object(
workflow_migration_commands,
"_migrate_legacy_sys_files_workflow_batch",
side_effect=[
workflow_migration_commands.LegacySysFilesWorkflowMigrationStats(
scanned=2,
migrated=1,
failed=0,
last_id="workflow-2",
),
workflow_migration_commands.LegacySysFilesWorkflowMigrationStats(
scanned=1,
migrated=1,
failed=0,
last_id="workflow-3",
),
],
)
stats = workflow_migration_commands.run_legacy_sys_files_workflow_migration(
batch_size=2,
limit=3,
start_after_id="workflow-0",
tenant_id="tenant-1",
app_id="app-1",
dry_run=True,
)
assert stats.scanned == 3
assert stats.migrated == 2
assert stats.batches == 2
assert stats.last_id == "workflow-3"
assert migrate_batch.call_args_list[0].kwargs["start_after_id"] == "workflow-0"
assert migrate_batch.call_args_list[0].kwargs["batch_size"] == 2
assert migrate_batch.call_args_list[1].kwargs["start_after_id"] == "workflow-2"
assert migrate_batch.call_args_list[1].kwargs["batch_size"] == 1
def test_run_legacy_sys_files_workflow_migration_stops_on_empty_batch(mocker):
session_maker = MagicMock(return_value=MagicMock())
mocker.patch.object(workflow_migration_commands, "sessionmaker", return_value=session_maker)
mocker.patch.object(workflow_migration_commands, "db", SimpleNamespace(engine=object()))
mocker.patch.object(
workflow_migration_commands,
"_migrate_legacy_sys_files_workflow_batch",
return_value=workflow_migration_commands.LegacySysFilesWorkflowMigrationStats(scanned=0),
)
stats = workflow_migration_commands.run_legacy_sys_files_workflow_migration(
batch_size=2,
limit=None,
start_after_id=None,
tenant_id=None,
app_id=None,
dry_run=False,
)
assert stats.scanned == 0
assert stats.batches == 0
def test_run_legacy_sys_files_workflow_migration_stops_on_short_batch(mocker):
session_maker = MagicMock(return_value=MagicMock())
mocker.patch.object(workflow_migration_commands, "sessionmaker", return_value=session_maker)
mocker.patch.object(workflow_migration_commands, "db", SimpleNamespace(engine=object()))
migrate_batch = mocker.patch.object(
workflow_migration_commands,
"_migrate_legacy_sys_files_workflow_batch",
return_value=workflow_migration_commands.LegacySysFilesWorkflowMigrationStats(
scanned=1,
migrated=1,
failed=0,
last_id="workflow-1",
),
)
stats = workflow_migration_commands.run_legacy_sys_files_workflow_migration(
batch_size=2,
limit=None,
start_after_id=None,
tenant_id=None,
app_id=None,
dry_run=False,
)
assert stats.scanned == 1
assert stats.batches == 1
migrate_batch.assert_called_once()
@@ -14,7 +14,7 @@ from sqlalchemy import select
from sqlalchemy.orm import Session
import commands
from commands import system as system_commands
from commands import workspace as workspace_commands
from core.tools.entities.tool_entities import ApiProviderSchemaType
from graphon.model_runtime.entities.model_entities import ModelType
from models import Tenant
@@ -83,11 +83,11 @@ def _encrypted_rows(tenant_id: str, *, suffix: str = "1") -> tuple[object, ...]:
def _bind_command_to_sqlite(monkeypatch: pytest.MonkeyPatch, session: Session) -> None:
monkeypatch.setattr(system_commands, "db", SimpleNamespace(engine=session.get_bind()))
monkeypatch.setattr(workspace_commands, "db", SimpleNamespace(engine=session.get_bind()))
def test_reset_aborts_when_not_self_hosted(monkeypatch, capsys):
monkeypatch.setattr(system_commands.dify_config, "EDITION", "CLOUD")
monkeypatch.setattr(workspace_commands.dify_config, "EDITION", "CLOUD")
exit_code = _invoke_reset()
captured = capsys.readouterr()
@@ -106,8 +106,8 @@ def test_reset_purges_provider_and_tool_tables_for_each_tenant(
) -> None:
"""The command must purge LLM provider rows AND every tool provider table
that stores ciphertext encrypted under the tenant key (#35396)."""
monkeypatch.setattr(system_commands.dify_config, "EDITION", "SELF_HOSTED")
monkeypatch.setattr(system_commands, "generate_key_pair", lambda tenant_id: f"new-key-{tenant_id}")
monkeypatch.setattr(workspace_commands.dify_config, "EDITION", "SELF_HOSTED")
monkeypatch.setattr(workspace_commands, "generate_key_pair", lambda tenant_id: f"new-key-{tenant_id}")
_bind_command_to_sqlite(monkeypatch, sqlite_session)
tenant = _tenant(TENANT_ID)
@@ -146,8 +146,8 @@ def test_reset_purges_provider_and_tool_tables_for_each_tenant(
)
def test_reset_iterates_all_tenants(monkeypatch: pytest.MonkeyPatch, sqlite_session: Session) -> None:
"""Multi-tenant deployments must purge every tenant, not just the first."""
monkeypatch.setattr(system_commands.dify_config, "EDITION", "SELF_HOSTED")
monkeypatch.setattr(system_commands, "generate_key_pair", lambda tenant_id: f"new-key-{tenant_id}")
monkeypatch.setattr(workspace_commands.dify_config, "EDITION", "SELF_HOSTED")
monkeypatch.setattr(workspace_commands, "generate_key_pair", lambda tenant_id: f"new-key-{tenant_id}")
_bind_command_to_sqlite(monkeypatch, sqlite_session)
tenant_ids = [f"11111111-1111-1111-1111-{index:012d}" for index in range(3)]
@@ -4,7 +4,7 @@ import types
from unittest.mock import MagicMock
import commands
from commands import system as system_commands
from commands import database as database_commands
from libs.db_migration_lock import LockNotOwnedError, RedisError
HEARTBEAT_WAIT_TIMEOUT_SECONDS = 5.0
@@ -25,11 +25,11 @@ def _invoke_upgrade_db() -> int:
def test_upgrade_db_skips_when_lock_not_acquired(monkeypatch, capsys):
monkeypatch.setattr(system_commands, "DB_UPGRADE_LOCK_TTL_SECONDS", 1234)
monkeypatch.setattr(database_commands, "DB_UPGRADE_LOCK_TTL_SECONDS", 1234)
lock = MagicMock()
lock.acquire.return_value = False
system_commands.redis_client.lock.return_value = lock
database_commands.redis_client.lock.return_value = lock
exit_code = _invoke_upgrade_db()
captured = capsys.readouterr()
@@ -37,18 +37,20 @@ def test_upgrade_db_skips_when_lock_not_acquired(monkeypatch, capsys):
assert exit_code == 0
assert "Database migration skipped" in captured.out
system_commands.redis_client.lock.assert_called_once_with(name="db_upgrade_lock", timeout=1234, thread_local=False)
database_commands.redis_client.lock.assert_called_once_with(
name="db_upgrade_lock", timeout=1234, thread_local=False
)
lock.acquire.assert_called_once_with(blocking=False)
lock.release.assert_not_called()
def test_upgrade_db_failure_not_masked_by_lock_release(monkeypatch, capsys):
monkeypatch.setattr(system_commands, "DB_UPGRADE_LOCK_TTL_SECONDS", 321)
monkeypatch.setattr(database_commands, "DB_UPGRADE_LOCK_TTL_SECONDS", 321)
lock = MagicMock()
lock.acquire.return_value = True
lock.release.side_effect = LockNotOwnedError("simulated")
system_commands.redis_client.lock.return_value = lock
database_commands.redis_client.lock.return_value = lock
def _upgrade():
raise RuntimeError("boom")
@@ -61,18 +63,18 @@ def test_upgrade_db_failure_not_masked_by_lock_release(monkeypatch, capsys):
assert exit_code == 1
assert "Database migration failed: boom" in captured.out
system_commands.redis_client.lock.assert_called_once_with(name="db_upgrade_lock", timeout=321, thread_local=False)
database_commands.redis_client.lock.assert_called_once_with(name="db_upgrade_lock", timeout=321, thread_local=False)
lock.acquire.assert_called_once_with(blocking=False)
lock.release.assert_called_once()
def test_upgrade_db_success_ignores_lock_not_owned_on_release(monkeypatch, capsys):
monkeypatch.setattr(system_commands, "DB_UPGRADE_LOCK_TTL_SECONDS", 999)
monkeypatch.setattr(database_commands, "DB_UPGRADE_LOCK_TTL_SECONDS", 999)
lock = MagicMock()
lock.acquire.return_value = True
lock.release.side_effect = LockNotOwnedError("simulated")
system_commands.redis_client.lock.return_value = lock
database_commands.redis_client.lock.return_value = lock
_install_fake_flask_migrate(monkeypatch, lambda: None)
@@ -82,7 +84,7 @@ def test_upgrade_db_success_ignores_lock_not_owned_on_release(monkeypatch, capsy
assert exit_code == 0
assert "Database migration successful!" in captured.out
system_commands.redis_client.lock.assert_called_once_with(name="db_upgrade_lock", timeout=999, thread_local=False)
database_commands.redis_client.lock.assert_called_once_with(name="db_upgrade_lock", timeout=999, thread_local=False)
lock.acquire.assert_called_once_with(blocking=False)
lock.release.assert_called_once()
@@ -93,11 +95,11 @@ def test_upgrade_db_renews_lock_during_migration(monkeypatch, capsys):
"""
# Use a small TTL so the heartbeat interval triggers quickly.
monkeypatch.setattr(system_commands, "DB_UPGRADE_LOCK_TTL_SECONDS", 0.3)
monkeypatch.setattr(database_commands, "DB_UPGRADE_LOCK_TTL_SECONDS", 0.3)
lock = MagicMock()
lock.acquire.return_value = True
system_commands.redis_client.lock.return_value = lock
database_commands.redis_client.lock.return_value = lock
renewed = threading.Event()
@@ -121,11 +123,11 @@ def test_upgrade_db_renews_lock_during_migration(monkeypatch, capsys):
def test_upgrade_db_ignores_reacquire_errors(monkeypatch, capsys):
# Use a small TTL so heartbeat runs during the upgrade call.
monkeypatch.setattr(system_commands, "DB_UPGRADE_LOCK_TTL_SECONDS", 0.3)
monkeypatch.setattr(database_commands, "DB_UPGRADE_LOCK_TTL_SECONDS", 0.3)
lock = MagicMock()
lock.acquire.return_value = True
system_commands.redis_client.lock.return_value = lock
database_commands.redis_client.lock.return_value = lock
attempted = threading.Event()
@@ -0,0 +1,49 @@
from unittest.mock import MagicMock
from commands import reset_encrypt_key_pair
from commands import workspace as workspace_commands
def test_reset_encrypt_key_pair_skips_non_self_hosted(monkeypatch, capsys):
monkeypatch.setattr(workspace_commands.dify_config, "EDITION", "CLOUD")
reset_encrypt_key_pair.callback()
captured = capsys.readouterr()
assert "only for SELF_HOSTED" in captured.out
def test_reset_encrypt_key_pair_rotates_keys_and_removes_custom_provider_data(monkeypatch, capsys):
monkeypatch.setattr(workspace_commands.dify_config, "EDITION", "SELF_HOSTED")
monkeypatch.setattr(workspace_commands, "generate_key_pair", lambda tenant_id: f"public-key-{tenant_id}")
tenant = MagicMock()
tenant.id = "tenant-1"
session = MagicMock()
session.scalars.return_value.all.return_value = [tenant]
session_manager = MagicMock()
session_manager.begin.return_value.__enter__.return_value = session
monkeypatch.setattr(workspace_commands, "sessionmaker", lambda *args, **kwargs: session_manager)
monkeypatch.setattr(workspace_commands, "db", MagicMock(engine=object()))
reset_encrypt_key_pair.callback()
assert tenant.encrypt_public_key == "public-key-tenant-1"
assert session.execute.call_count == 5
captured = capsys.readouterr()
assert "tenant-1 has been reset" in captured.out
def test_reset_encrypt_key_pair_stops_when_workspace_record_is_missing(monkeypatch, capsys):
monkeypatch.setattr(workspace_commands.dify_config, "EDITION", "SELF_HOSTED")
session = MagicMock()
session.scalars.return_value.all.return_value = [None]
session_manager = MagicMock()
session_manager.begin.return_value.__enter__.return_value = session
monkeypatch.setattr(workspace_commands, "sessionmaker", lambda *args, **kwargs: session_manager)
monkeypatch.setattr(workspace_commands, "db", MagicMock(engine=object()))
reset_encrypt_key_pair.callback()
session.execute.assert_not_called()
captured = capsys.readouterr()
assert "No workspaces found" in captured.out
@@ -4,6 +4,7 @@ from dotenv import dotenv_values
BASE_API_AND_DOCKER_CONFIG_SET_DIFF: frozenset[str] = frozenset(
(
"AGENT_BACKEND_API_TOKEN",
"APP_MAX_EXECUTION_TIME",
"BATCH_UPLOAD_LIMIT",
"CELERY_BEAT_SCHEDULER_TIME",
@@ -43,6 +44,7 @@ BASE_API_AND_DOCKER_CONFIG_SET_DIFF: frozenset[str] = frozenset(
BASE_API_AND_DOCKER_COMPOSE_CONFIG_SET_DIFF: frozenset[str] = frozenset(
(
"AGENT_BACKEND_API_TOKEN",
"BATCH_UPLOAD_LIMIT",
"CELERY_BEAT_SCHEDULER_TIME",
"HTTP_REQUEST_MAX_CONNECT_TIMEOUT",
+56 -26
View File
@@ -1,11 +1,13 @@
import os
import shutil
from collections.abc import Iterator
from pathlib import Path
from unittest.mock import MagicMock, patch
import pytest
from flask import Flask
from sqlalchemy import create_engine
from sqlalchemy.engine import Engine
from sqlalchemy.engine import URL, Engine
from sqlalchemy.orm import Session, sessionmaker
# Getting the absolute path of the current file's directory
@@ -35,7 +37,7 @@ os.environ.setdefault("OPENDAL_SCHEME", "fs")
os.environ.setdefault("OPENDAL_FS_ROOT", "/tmp/dify-storage")
os.environ.setdefault("STORAGE_TYPE", "opendal")
from core.db.session_factory import configure_session_factory, session_factory
import core.db.session_factory as session_factory_module
from extensions import ext_redis
from models.account import Account, Tenant, TenantAccountJoin, TenantAccountRole
from models.base import TypeBase
@@ -111,42 +113,70 @@ def reset_secret_key():
dify_config.SECRET_KEY = original
@pytest.fixture(scope="session")
def _unit_test_engine():
engine = create_engine("sqlite:///:memory:")
yield engine
engine.dispose()
@pytest.fixture
def sqlite_engine() -> Iterator[Engine]:
"""Create an isolated in-memory SQLite engine for tests that need a disposable database."""
def _sqlite_engine(_sqlite_database_template: Path, tmp_path: Path) -> Iterator[Engine]:
"""Create an engine over a pristine per-test copy of the SQLite schema."""
database_path = tmp_path / "unit-tests.sqlite3"
shutil.copyfile(_sqlite_database_template, database_path)
engine = create_engine(URL.create("sqlite", database=str(database_path)))
engine = create_engine("sqlite:///:memory:")
try:
yield engine
finally:
engine.dispose()
database_path.unlink(missing_ok=True)
@pytest.fixture
def sqlite_session(request: pytest.FixtureRequest, sqlite_engine: Engine) -> Iterator[Session]:
"""Yield a SQLite session after creating the model tables passed through ``request.param``."""
@pytest.fixture(scope="session")
def _sqlite_database_template(tmp_path_factory: pytest.TempPathFactory) -> Path:
"""Create one empty full-schema SQLite database per pytest worker."""
models: tuple[type[TypeBase], ...] = request.param
tables = [model.metadata.tables[model.__tablename__] for model in models]
TypeBase.metadata.create_all(sqlite_engine, tables=tables)
session_factory = sessionmaker(bind=sqlite_engine, expire_on_commit=False)
with session_factory() as session:
yield session
database_path = tmp_path_factory.mktemp("sqlite-template") / "unit-tests.sqlite3"
engine = create_engine(URL.create("sqlite", database=str(database_path)))
try:
TypeBase.metadata.create_all(engine)
finally:
engine.dispose()
return database_path
@pytest.fixture(autouse=True)
def _configure_session_factory(_unit_test_engine):
try:
session_factory.get_session_maker()
except RuntimeError:
configure_session_factory(_unit_test_engine, expire_on_commit=False)
def _sqlite_session_factory(
_sqlite_engine: Engine,
monkeypatch: pytest.MonkeyPatch,
) -> sessionmaker[Session]:
"""Bind all unit-test Sessions to the pristine full-schema SQLite database."""
factory = sessionmaker(bind=_sqlite_engine, expire_on_commit=False)
monkeypatch.setattr(session_factory_module, "_session_maker", factory)
return factory
@pytest.fixture
def sqlite_engine(_sqlite_engine: Engine) -> Engine:
"""Expose the pristine full-schema SQLite engine to tests."""
return _sqlite_engine
@pytest.fixture
def sqlite_session_factory(_sqlite_session_factory: sessionmaker[Session]) -> sessionmaker[Session]:
"""Expose the shared SQLite session factory to tests."""
return _sqlite_session_factory
@pytest.fixture
def sqlite_session(_sqlite_session_factory: sessionmaker[Session]) -> Iterator[Session]:
"""Yield a session over the pristine full-schema SQLite database.
Legacy indirect model parameters remain accepted by pytest but are ignored.
Remove those decorators as their test files receive individual review.
"""
with _sqlite_session_factory() as session:
yield session
def persist_service_api_tenant_owner(session: Session, tenant: Tenant, owner: Account) -> TenantAccountJoin:
@@ -708,6 +708,121 @@ def test_app_list_api_attaches_permission_keys(app, app_module):
assert resp["data"][0]["permission_keys"] == ["app.acl.view_layout", "app.acl.edit"]
def test_recent_app_list_api_returns_only_home_card_fields(app, app_module):
method = app_module.RecentAppListApi.get
while hasattr(method, "__wrapped__"):
method = method.__wrapped__
recent_app = SimpleNamespace(
id="app-1",
name="Recent App",
icon_type="emoji",
icon="🚀",
icon_background="#FFFFFF",
mode="chat",
author_name="Recent Author",
updated_at=_ts(15),
maintainer="acct-1",
)
get_recent_apps = MagicMock(return_value=[recent_app])
with app.test_request_context("/apps/recent?limit=8"):
with pytest.MonkeyPatch.context() as monkeypatch:
monkeypatch.setattr(dify_config, "RBAC_ENABLED", False)
monkeypatch.setattr(app_module.AppService, "get_recent_apps", get_recent_apps)
monkeypatch.setattr(
app_module.enterprise_rbac_service.RBACService.MyPermissions,
"get",
lambda tenant_id, account_id, session: app_module.enterprise_rbac_service.MyPermissionsResponse(
app=app_module.enterprise_rbac_service.ResourcePermissionSnapshot(
overrides=[
app_module.enterprise_rbac_service.ResourcePermissionKeys(
resource_id="app-1",
permission_keys=["app.acl.monitor"],
)
]
)
),
)
resp, status = method(app_module.RecentAppListApi(), "tenant-1", "acct-1", MagicMock())
assert status == 200
assert resp == {
"data": [
{
"id": "app-1",
"name": "Recent App",
"icon_type": "emoji",
"icon": "🚀",
"icon_background": "#FFFFFF",
"mode": "chat",
"author_name": "Recent Author",
"updated_at": int(_ts(15).timestamp()),
"permission_keys": ["app.acl.monitor"],
"maintainer": "acct-1",
"icon_url": None,
}
]
}
params = get_recent_apps.call_args.args[2]
assert params.limit == 8
assert "total" not in resp
assert "description" not in resp["data"][0]
assert "tags" not in resp["data"][0]
assert "workflow" not in resp["data"][0]
@pytest.mark.parametrize("mode", ["channel", "rag-pipeline", "agent"])
def test_recent_app_response_rejects_non_home_app_modes(app_module, mode: str) -> None:
with pytest.raises(ValidationError):
app_module.RecentAppResponse.model_validate(
{
"id": "app-1",
"name": "Recent App",
"mode": mode,
"updated_at": _ts(),
}
)
def test_recent_app_list_api_applies_rbac_visibility_filter(app, app_module):
method = app_module.RecentAppListApi.get
while hasattr(method, "__wrapped__"):
method = method.__wrapped__
get_recent_apps = MagicMock(return_value=[])
with app.test_request_context("/apps/recent"):
with pytest.MonkeyPatch.context() as monkeypatch:
monkeypatch.setattr(dify_config, "RBAC_ENABLED", True)
monkeypatch.setattr(app_module.AppService, "get_recent_apps", get_recent_apps)
monkeypatch.setattr(
app_module.enterprise_rbac_service.RBACService.MyPermissions,
"get",
lambda tenant_id, account_id, session: app_module.enterprise_rbac_service.MyPermissionsResponse(
workspace=app_module.enterprise_rbac_service.WorkspacePermissionSnapshot(
permission_keys=["app.create_and_management"]
)
),
)
monkeypatch.setattr(
app_module.enterprise_rbac_service.RBACService.AppAccess,
"whitelist_resources",
lambda tenant_id, account_id: SimpleNamespace(
unrestricted=False,
resource_ids=["app-shared"],
),
)
resp, status = method(app_module.RecentAppListApi(), "tenant-1", "acct-1", MagicMock())
assert status == 200
assert resp == {"data": []}
params = get_recent_apps.call_args.args[2]
assert params.accessible_app_ids == ["app-shared"]
assert params.include_own_apps is True
def test_app_list_api_limits_to_apps_created_by_current_user_without_view_permission(app, app_module):
method = app_module.AppListApi.get
while hasattr(method, "__wrapped__"):
@@ -1,18 +1,22 @@
from __future__ import annotations
import inspect
from collections.abc import Callable
from collections.abc import Callable, Iterator
from datetime import UTC, datetime
from typing import cast
from typing import Literal, cast
from unittest.mock import MagicMock, PropertyMock, patch
from uuid import uuid4
from uuid import UUID
import pytest
from flask import Flask
from sqlalchemy import select
from sqlalchemy.orm import Session
from werkzeug.exceptions import NotFound
from controllers.console.datasets import data_source as module
from controllers.console.datasets.data_source import DataSourceApi, DataSourceNotionListApi
from models import Account, DataSourceOauthBinding
from models.engine import db
ControllerMethod = Callable[..., tuple[dict[str, object], int]]
@@ -22,10 +26,15 @@ def unwrap(func: object) -> ControllerMethod:
@pytest.fixture
def flask_app() -> Flask:
def flask_app() -> Iterator[Flask]:
app = Flask(__name__)
app.config["TESTING"] = True
return app
app.config["SQLALCHEMY_DATABASE_URI"] = "sqlite:///:memory:"
db.init_app(app)
with app.app_context():
DataSourceOauthBinding.__table__.create(db.engine)
yield app
@pytest.fixture
@@ -35,9 +44,13 @@ def current_user() -> Account:
return account
def test_get_data_source_integrates_serializes_orm_binding(flask_app: Flask) -> None:
TENANT_ID = "11111111-1111-1111-1111-111111111111"
BINDING_ID = "22222222-2222-2222-2222-222222222222"
def _add_binding(session: Session, *, disabled: bool) -> DataSourceOauthBinding:
binding = DataSourceOauthBinding(
tenant_id="tenant-1",
tenant_id=TENANT_ID,
access_token="token",
provider="notion",
source_info={
@@ -55,24 +68,31 @@ def test_get_data_source_integrates_serializes_orm_binding(flask_app: Flask) ->
}
],
},
disabled=disabled,
)
binding.id = "binding-1"
binding.id = BINDING_ID
binding.created_at = datetime(2026, 5, 25, 1, 2, 3, tzinfo=UTC)
binding.disabled = False
session.add(binding)
session.commit()
return binding
with (
flask_app.test_request_context("/"),
patch.object(module.db.session, "scalars", return_value=MagicMock(all=lambda: [binding])),
):
response, status = unwrap(DataSourceApi().get)(DataSourceApi(), "tenant-1")
def test_get_data_source_integrates_serializes_orm_binding(
flask_app: Flask,
) -> None:
binding = _add_binding(db.session, disabled=False)
expected_created_at = int(binding.created_at.timestamp())
with flask_app.test_request_context("/"):
response, status = unwrap(DataSourceApi().get)(DataSourceApi(), TENANT_ID)
assert status == 200
assert response == {
"data": [
{
"id": "binding-1",
"id": BINDING_ID,
"provider": "notion",
"created_at": 1779670923,
"created_at": expected_created_at,
"is_bound": True,
"disabled": False,
"source_info": {
@@ -96,34 +116,75 @@ def test_get_data_source_integrates_serializes_orm_binding(flask_app: Flask) ->
}
def test_get_data_source_integrates_preserves_empty_list_when_no_binding(flask_app: Flask) -> None:
with (
flask_app.test_request_context("/"),
patch.object(module.db.session, "scalars", return_value=MagicMock(all=lambda: [])),
):
response, status = unwrap(DataSourceApi().get)(DataSourceApi(), "tenant-1")
def test_get_data_source_integrates_preserves_empty_list_when_no_binding(
flask_app: Flask,
) -> None:
with flask_app.test_request_context("/"):
response, status = unwrap(DataSourceApi().get)(DataSourceApi(), TENANT_ID)
assert status == 200
assert response == {"data": []}
def test_patch_data_source_binding_uses_injected_session(flask_app: Flask) -> None:
binding = MagicMock(disabled=True)
session = MagicMock()
session.scalar.return_value = binding
@pytest.mark.parametrize(
("disabled", "action", "expected_disabled"),
[(True, "enable", False), (False, "disable", True)],
)
@pytest.mark.parametrize("sqlite_session", [(DataSourceOauthBinding,)], indirect=True)
def test_patch_data_source_binding_updates_state(
flask_app: Flask,
sqlite_session: Session,
disabled: bool,
action: Literal["enable", "disable"],
expected_disabled: bool,
) -> None:
_add_binding(sqlite_session, disabled=disabled)
sqlite_session.expunge_all()
with flask_app.test_request_context("/"):
response, status = unwrap(DataSourceApi().patch)(DataSourceApi(), session, "tenant-1", uuid4(), "enable")
response, status = unwrap(DataSourceApi().patch)(
DataSourceApi(), sqlite_session, TENANT_ID, UUID(BINDING_ID), action
)
sqlite_session.flush()
sqlite_session.expire_all()
binding = sqlite_session.scalar(select(DataSourceOauthBinding).where(DataSourceOauthBinding.id == BINDING_ID))
assert status == 200
assert response == {"result": "success"}
assert binding.disabled is False
session.scalar.assert_called_once()
session.add.assert_not_called()
session.commit.assert_not_called()
assert binding is not None
assert binding.disabled is expected_disabled
def test_notion_pre_import_pages_serializes_frontend_list_shape(flask_app: Flask, current_user: Account) -> None:
@pytest.mark.parametrize("sqlite_session", [(DataSourceOauthBinding,)], indirect=True)
def test_patch_data_source_binding_rejects_unknown_binding(
flask_app: Flask,
sqlite_session: Session,
) -> None:
with flask_app.test_request_context("/"), pytest.raises(NotFound, match="Data source binding not found"):
unwrap(DataSourceApi().patch)(DataSourceApi(), sqlite_session, TENANT_ID, UUID(BINDING_ID), "enable")
@pytest.mark.parametrize(("disabled", "action"), [(False, "enable"), (True, "disable")])
@pytest.mark.parametrize("sqlite_session", [(DataSourceOauthBinding,)], indirect=True)
def test_patch_data_source_binding_rejects_current_state(
flask_app: Flask,
sqlite_session: Session,
disabled: bool,
action: Literal["enable", "disable"],
) -> None:
_add_binding(sqlite_session, disabled=disabled)
sqlite_session.expunge_all()
with flask_app.test_request_context("/"), pytest.raises(ValueError):
unwrap(DataSourceApi().patch)(DataSourceApi(), sqlite_session, TENANT_ID, UUID(BINDING_ID), action)
@pytest.mark.parametrize("sqlite_session", [()], indirect=True)
def test_notion_pre_import_pages_serializes_frontend_list_shape(
flask_app: Flask,
current_user: Account,
sqlite_session: Session,
) -> None:
page = MagicMock(
page_id="page-1",
page_name="Page",
@@ -145,8 +206,6 @@ def test_notion_pre_import_pages_serializes_frontend_list_shape(flask_app: Flask
get_online_document_pages=MagicMock(return_value=iter([online_document_message])),
datasource_provider_type=MagicMock(return_value="online_document"),
)
session = MagicMock()
with (
flask_app.test_request_context("/?credential_id=credential-1"),
patch.object(
@@ -158,7 +217,7 @@ def test_notion_pre_import_pages_serializes_frontend_list_shape(flask_app: Flask
patch("core.datasource.datasource_manager.DatasourceManager.get_datasource_runtime", return_value=runtime),
):
response, status = unwrap(DataSourceNotionListApi().get)(
DataSourceNotionListApi(), session, "tenant-1", current_user
DataSourceNotionListApi(), sqlite_session, "tenant-1", current_user
)
assert status == 200
@@ -183,3 +242,38 @@ def test_notion_pre_import_pages_serializes_frontend_list_shape(flask_app: Flask
}
runtime.get_online_document_pages.assert_called_once()
assert runtime.get_online_document_pages.call_args.kwargs["datasource_parameters"] == {}
@pytest.mark.parametrize("sqlite_session", [()], indirect=True)
def test_notion_pre_import_pages_rejects_missing_credential(
flask_app: Flask,
current_user: Account,
sqlite_session: Session,
) -> None:
with (
flask_app.test_request_context("/?credential_id=credential-1"),
patch.object(module.DatasourceProviderService, "get_datasource_credentials", return_value=None),
pytest.raises(NotFound, match="Credential not found"),
):
unwrap(DataSourceNotionListApi().get)(DataSourceNotionListApi(), sqlite_session, TENANT_ID, current_user)
@pytest.mark.parametrize("sqlite_session", [()], indirect=True)
def test_notion_pre_import_pages_rejects_non_notion_dataset(
flask_app: Flask,
current_user: Account,
sqlite_session: Session,
) -> None:
dataset = MagicMock(data_source_type="other_type")
with (
flask_app.test_request_context("/?credential_id=credential-1&dataset_id=dataset-1"),
patch.object(
module.DatasourceProviderService,
"get_datasource_credentials",
return_value={"token": "token"},
),
patch.object(module.DatasetService, "get_dataset", return_value=dataset),
pytest.raises(ValueError, match="Dataset is not notion type"),
):
unwrap(DataSourceNotionListApi().get)(DataSourceNotionListApi(), sqlite_session, TENANT_ID, current_user)
@@ -0,0 +1,171 @@
"""Unit tests for controllers.console.datasets.data_source Notion endpoints."""
from __future__ import annotations
import inspect
from unittest.mock import MagicMock, patch
import pytest
from flask import Flask
from sqlalchemy.orm import Session
from werkzeug.exceptions import NotFound
from controllers.console.datasets.data_source import (
DataSourceNotionDatasetSyncApi,
DataSourceNotionDocumentSyncApi,
DataSourceNotionIndexingEstimateApi,
DataSourceNotionPreviewApi,
)
from core.rag.index_processor.constant.index_type import IndexStructureType
from models import Account
@pytest.fixture
def current_user() -> Account:
account = Account(name="Test User", email="u1@example.com")
account.id = "u1"
return account
class TestDataSourceNotionPreviewApi:
def test_get_preview_success(self, app: Flask) -> None:
api = DataSourceNotionPreviewApi()
method = inspect.unwrap(api.get)
extractor = MagicMock(extract=lambda: [MagicMock(page_content="hello")])
with (
app.test_request_context("/?credential_id=c1"),
patch(
"controllers.console.datasets.data_source.DatasourceProviderService.get_datasource_credentials",
return_value={"integration_secret": "t"},
),
patch(
"controllers.console.datasets.data_source.NotionExtractor",
return_value=extractor,
),
):
response, status = method(api, "tenant-1", "p1", "page")
assert status == 200
class TestDataSourceNotionIndexingEstimateApi:
@pytest.mark.parametrize("sqlite_session", [()], indirect=True)
def test_post_indexing_estimate_success(self, app: Flask, sqlite_session: Session) -> None:
api = DataSourceNotionIndexingEstimateApi()
method = inspect.unwrap(api.post)
empty_rules: dict[str, object] = {}
payload: dict[str, object] = {
"notion_info_list": [
{
"workspace_id": "w1",
"credential_id": "c1",
"pages": [{"page_id": "p1", "type": "page"}],
}
],
"process_rule": {"rules": empty_rules},
"doc_form": IndexStructureType.PARAGRAPH_INDEX,
"doc_language": "English",
}
with (
app.test_request_context("/", method="POST", json=payload, headers={"Content-Type": "application/json"}),
patch(
"controllers.console.datasets.data_source.DocumentService.estimate_args_validate",
),
patch(
"controllers.console.datasets.data_source.IndexingRunner.indexing_estimate",
return_value=MagicMock(model_dump=lambda: {"total_pages": 1}),
),
):
response, status = method(api, sqlite_session, "tenant-1")
assert status == 200
class TestDataSourceNotionDatasetSyncApi:
@pytest.mark.parametrize("sqlite_session", [()], indirect=True)
def test_get_success(self, app: Flask, sqlite_session: Session) -> None:
api = DataSourceNotionDatasetSyncApi()
method = inspect.unwrap(api.get)
with (
app.test_request_context("/"),
patch(
"controllers.console.datasets.data_source.DatasetService.get_dataset",
return_value=MagicMock(),
),
patch(
"controllers.console.datasets.data_source.DocumentService.get_document_by_dataset_id",
return_value=[MagicMock(id="d1")],
),
patch(
"controllers.console.datasets.data_source.document_indexing_sync_task.delay",
return_value=None,
),
):
response, status = method(api, sqlite_session, "ds-1")
assert status == 200
@pytest.mark.parametrize("sqlite_session", [()], indirect=True)
def test_get_dataset_not_found(self, app: Flask, sqlite_session: Session) -> None:
api = DataSourceNotionDatasetSyncApi()
method = inspect.unwrap(api.get)
with (
app.test_request_context("/"),
patch(
"controllers.console.datasets.data_source.DatasetService.get_dataset",
return_value=None,
),
):
with pytest.raises(NotFound):
method(api, sqlite_session, "ds-1")
class TestDataSourceNotionDocumentSyncApi:
@pytest.mark.parametrize("sqlite_session", [()], indirect=True)
def test_get_success(self, app: Flask, sqlite_session: Session) -> None:
api = DataSourceNotionDocumentSyncApi()
method = inspect.unwrap(api.get)
with (
app.test_request_context("/"),
patch(
"controllers.console.datasets.data_source.DatasetService.get_dataset",
return_value=MagicMock(),
),
patch(
"controllers.console.datasets.data_source.DocumentService.get_document",
return_value=MagicMock(),
),
patch(
"controllers.console.datasets.data_source.document_indexing_sync_task.delay",
return_value=None,
),
):
response, status = method(api, sqlite_session, "ds-1", "doc-1")
assert status == 200
@pytest.mark.parametrize("sqlite_session", [()], indirect=True)
def test_get_document_not_found(self, app: Flask, sqlite_session: Session) -> None:
api = DataSourceNotionDocumentSyncApi()
method = inspect.unwrap(api.get)
with (
app.test_request_context("/"),
patch(
"controllers.console.datasets.data_source.DatasetService.get_dataset",
return_value=MagicMock(),
),
patch(
"controllers.console.datasets.data_source.DocumentService.get_document",
return_value=None,
),
):
with pytest.raises(NotFound):
method(api, sqlite_session, "ds-1", "doc-1")
@@ -1,64 +0,0 @@
import inspect
from unittest.mock import call, patch
import pytest
from flask import Flask
from pydantic import ValidationError
from controllers.inner_api.workspace.plugin_model_providers import (
EnterprisePluginModelProvidersCacheInvalidate,
InvalidatePluginModelProvidersCachePayload,
)
class TestInvalidatePluginModelProvidersCachePayload:
def test_valid_payload(self):
payload = InvalidatePluginModelProvidersCachePayload.model_validate(
{"tenant_ids": ["tenant-alpha", "tenant-beta"]}
)
assert payload.tenant_ids == ["tenant-alpha", "tenant-beta"]
def test_missing_tenant_ids_defaults_to_empty(self):
payload = InvalidatePluginModelProvidersCachePayload.model_validate({})
assert payload.tenant_ids == []
def test_unknown_field_rejected(self):
with pytest.raises(ValidationError):
InvalidatePluginModelProvidersCachePayload.model_validate({"tenant_ids": ["tenant-alpha"], "generation": 7})
class TestEnterprisePluginModelProvidersCacheInvalidate:
@pytest.fixture
def api_instance(self):
return EnterprisePluginModelProvidersCacheInvalidate()
def _post(self, api_instance, app: Flask, payload):
unwrapped_post = inspect.unwrap(api_instance.post)
with app.test_request_context():
with patch("controllers.inner_api.workspace.plugin_model_providers.inner_api_ns") as mock_ns:
mock_ns.payload = payload
return unwrapped_post(api_instance)
@patch("controllers.inner_api.workspace.plugin_model_providers.PluginService")
def test_post_invalidates_once_per_tenant(self, mock_plugin_service, api_instance, app: Flask):
result = self._post(api_instance, app, {"tenant_ids": ["tenant-alpha", "tenant-beta"]})
assert result == ({"result": "success"}, 200)
assert mock_plugin_service.invalidate_plugin_model_providers_cache.call_args_list == [
call("tenant-alpha"),
call("tenant-beta"),
]
@patch("controllers.inner_api.workspace.plugin_model_providers.PluginService")
def test_post_with_empty_list_is_a_no_op(self, mock_plugin_service, api_instance, app: Flask):
result = self._post(api_instance, app, {"tenant_ids": []})
assert result == ({"result": "success"}, 200)
mock_plugin_service.invalidate_plugin_model_providers_cache.assert_not_called()
@patch("controllers.inner_api.workspace.plugin_model_providers.PluginService")
def test_post_with_missing_payload_is_a_no_op(self, mock_plugin_service, api_instance, app: Flask):
result = self._post(api_instance, app, None)
assert result == ({"result": "success"}, 200)
mock_plugin_service.invalidate_plugin_model_providers_cache.assert_not_called()
@@ -1,17 +1,31 @@
"""
Unit tests for Service API File Preview endpoint
"""Unit tests for the Service API file-preview endpoint.
Ownership checks run against persisted message, file, app, and upload rows so the
tests exercise the same SQLAlchemy statements and tenant boundary as production.
Storage remains mocked because it is the external I/O boundary of the endpoint.
"""
import logging
import uuid
from collections.abc import Iterator
from dataclasses import dataclass
from datetime import datetime
from decimal import Decimal
from typing import Protocol, cast
from unittest.mock import Mock, patch
from uuid import uuid4
import pytest
from sqlalchemy import event
from sqlalchemy.engine import Engine
from sqlalchemy.orm import Session
from controllers.service_api.app.error import FileAccessDeniedError, FileNotFoundError
from controllers.service_api.app.file_preview import FilePreviewApi
from models.model import App, EndUser, Message, MessageFile, UploadFile
from extensions.storage.storage_type import StorageType
from graphon.file import FileTransferMethod, FileType
from models.base import TypeBase
from models.enums import ConversationFromSource, CreatorUserRole
from models.model import App, AppMode, Message, MessageFile, UploadFile
class _FilePreviewLogRecord(Protocol):
@@ -20,367 +34,252 @@ class _FilePreviewLogRecord(Protocol):
error: str
@dataclass(frozen=True)
class _Database:
"""Expose the real test session through the interface used by the controller."""
session: Session
@dataclass(frozen=True)
class _PreviewRecords:
app: App
message: Message
message_file: MessageFile
upload_file: UploadFile
@pytest.fixture
def database(sqlite_engine: Engine) -> Iterator[_Database]:
"""Create only the tables required by file ownership validation."""
models = (App, Message, MessageFile, UploadFile)
tables = [TypeBase.metadata.tables[model.__tablename__] for model in models]
TypeBase.metadata.create_all(sqlite_engine, tables=tables)
with Session(sqlite_engine, expire_on_commit=False) as session:
yield _Database(session)
@pytest.fixture
def file_preview_api() -> FilePreviewApi:
"""Create the resource instance under test."""
return FilePreviewApi()
def _upload_file(*, tenant_id: str, file_id: str | None = None) -> UploadFile:
upload_file = UploadFile(
tenant_id=tenant_id,
storage_type=StorageType.LOCAL,
key="storage/key/test_file.jpg",
name="test_file.jpg",
size=1024,
extension="jpg",
mime_type="image/jpeg",
created_by_role=CreatorUserRole.ACCOUNT,
created_by=str(uuid4()),
created_at=datetime(2026, 1, 1),
used=True,
)
if file_id is not None:
upload_file.id = file_id
return upload_file
def _persist_preview_records(
session: Session,
*,
app_id: str | None = None,
app_tenant_id: str | None = None,
upload_tenant_id: str | None = None,
) -> _PreviewRecords:
app_id = app_id or str(uuid4())
app_tenant_id = app_tenant_id or str(uuid4())
upload_file = _upload_file(tenant_id=upload_tenant_id or app_tenant_id)
app = App(
id=app_id,
tenant_id=app_tenant_id,
name="Preview app",
description="",
mode=AppMode.CHAT,
icon_type=None,
icon="",
icon_background=None,
enable_site=True,
enable_api=True,
)
message = Message(
id=str(uuid4()),
app_id=app_id,
conversation_id=str(uuid4()),
_inputs={},
query="preview",
message={},
message_unit_price=Decimal(0),
answer="answer",
answer_unit_price=Decimal(0),
currency="USD",
from_source=ConversationFromSource.API,
)
message_file = MessageFile(
message_id=message.id,
type=FileType.IMAGE,
transfer_method=FileTransferMethod.LOCAL_FILE,
created_by_role=CreatorUserRole.ACCOUNT,
created_by=str(uuid4()),
upload_file_id=upload_file.id,
)
session.add_all([app, message, message_file, upload_file])
session.commit()
return _PreviewRecords(app=app, message=message, message_file=message_file, upload_file=upload_file)
class TestFilePreviewApi:
"""Test suite for FilePreviewApi"""
"""Exercise ownership validation and response construction."""
@pytest.fixture
def file_preview_api(self):
"""Create FilePreviewApi instance for testing"""
return FilePreviewApi()
def test_validate_file_ownership_success(self, file_preview_api: FilePreviewApi, database: _Database):
records = _persist_preview_records(database.session)
@pytest.fixture
def mock_app(self):
"""Mock App model"""
app = Mock(spec=App)
app.id = str(uuid.uuid4())
app.tenant_id = str(uuid.uuid4())
return app
with patch("controllers.service_api.app.file_preview.db", database):
message_file, upload_file = file_preview_api._validate_file_ownership(
records.upload_file.id, records.app.id
)
@pytest.fixture
def mock_end_user(self):
"""Mock EndUser model"""
end_user = Mock(spec=EndUser)
end_user.id = str(uuid.uuid4())
return end_user
assert message_file.id == records.message_file.id
assert upload_file.id == records.upload_file.id
assert upload_file.tenant_id == records.app.tenant_id
@pytest.fixture
def mock_upload_file(self):
"""Mock UploadFile model"""
upload_file = Mock(spec=UploadFile)
upload_file.id = str(uuid.uuid4())
upload_file.name = "test_file.jpg"
upload_file.extension = "jpg"
upload_file.mime_type = "image/jpeg"
upload_file.size = 1024
upload_file.key = "storage/key/test_file.jpg"
upload_file.tenant_id = str(uuid.uuid4())
return upload_file
def test_validate_file_ownership_file_not_found(self, file_preview_api: FilePreviewApi, database: _Database):
with patch("controllers.service_api.app.file_preview.db", database):
with pytest.raises(FileNotFoundError, match="File not found in message context"):
file_preview_api._validate_file_ownership(str(uuid4()), str(uuid4()))
@pytest.fixture
def mock_message_file(self):
"""Mock MessageFile model"""
message_file = Mock(spec=MessageFile)
message_file.id = str(uuid.uuid4())
message_file.upload_file_id = str(uuid.uuid4())
message_file.message_id = str(uuid.uuid4())
return message_file
def test_validate_file_ownership_access_denied(self, file_preview_api: FilePreviewApi, database: _Database):
records = _persist_preview_records(database.session)
@pytest.fixture
def mock_message(self):
"""Mock Message model"""
message = Mock(spec=Message)
message.id = str(uuid.uuid4())
message.app_id = str(uuid.uuid4())
return message
with patch("controllers.service_api.app.file_preview.db", database):
with pytest.raises(FileAccessDeniedError, match="not owned by requesting app"):
file_preview_api._validate_file_ownership(records.upload_file.id, str(uuid4()))
def test_validate_file_ownership_success(
self, file_preview_api: FilePreviewApi, mock_app, mock_upload_file, mock_message_file, mock_message
):
"""Test successful file ownership validation"""
file_id = str(uuid.uuid4())
app_id = mock_app.id
def test_validate_file_ownership_upload_file_not_found(self, file_preview_api: FilePreviewApi, database: _Database):
records = _persist_preview_records(database.session)
database.session.delete(records.upload_file)
database.session.commit()
# Set up the mocks
mock_upload_file.tenant_id = mock_app.tenant_id
mock_message.app_id = app_id
mock_message_file.upload_file_id = file_id
mock_message_file.message_id = mock_message.id
with patch("controllers.service_api.app.file_preview.db", database):
with pytest.raises(FileNotFoundError, match="Upload file record not found"):
file_preview_api._validate_file_ownership(records.upload_file.id, records.app.id)
with patch("controllers.service_api.app.file_preview.db") as mock_db:
# Mock scalar() for MessageFile and Message queries
mock_db.session.scalar.side_effect = [
mock_message_file, # MessageFile query
mock_message, # Message query
]
# Mock get() for UploadFile and App PK lookups
mock_db.session.get.side_effect = [
mock_upload_file, # UploadFile query
mock_app, # App query for tenant validation
]
def test_validate_file_ownership_tenant_mismatch(self, file_preview_api: FilePreviewApi, database: _Database):
records = _persist_preview_records(database.session, upload_tenant_id=str(uuid4()))
# Execute the method
result_message_file, result_upload_file = file_preview_api._validate_file_ownership(file_id, app_id)
# Assertions
assert result_message_file == mock_message_file
assert result_upload_file == mock_upload_file
def test_validate_file_ownership_file_not_found(self, file_preview_api: FilePreviewApi):
"""Test file ownership validation when MessageFile not found"""
file_id = str(uuid.uuid4())
app_id = str(uuid.uuid4())
with patch("controllers.service_api.app.file_preview.db") as mock_db:
# Mock MessageFile not found via scalar()
mock_db.session.scalar.return_value = None
# Execute and assert exception
with pytest.raises(FileNotFoundError) as exc_info:
file_preview_api._validate_file_ownership(file_id, app_id)
assert "File not found in message context" in str(exc_info.value)
def test_validate_file_ownership_access_denied(self, file_preview_api: FilePreviewApi, mock_message_file):
"""Test file ownership validation when Message not owned by app"""
file_id = str(uuid.uuid4())
app_id = str(uuid.uuid4())
with patch("controllers.service_api.app.file_preview.db") as mock_db:
# Mock MessageFile found but Message not owned by app via scalar()
mock_db.session.scalar.side_effect = [
mock_message_file, # MessageFile query - found
None, # Message query - not found (access denied)
]
# Execute and assert exception
with pytest.raises(FileAccessDeniedError) as exc_info:
file_preview_api._validate_file_ownership(file_id, app_id)
assert "not owned by requesting app" in str(exc_info.value)
def test_validate_file_ownership_upload_file_not_found(
self, file_preview_api: FilePreviewApi, mock_message_file, mock_message
):
"""Test file ownership validation when UploadFile not found"""
file_id = str(uuid.uuid4())
app_id = str(uuid.uuid4())
with patch("controllers.service_api.app.file_preview.db") as mock_db:
# Mock scalar() for MessageFile and Message
mock_db.session.scalar.side_effect = [
mock_message_file, # MessageFile query - found
mock_message, # Message query - found
]
# Mock get() for UploadFile - not found
mock_db.session.get.return_value = None
# Execute and assert exception
with pytest.raises(FileNotFoundError) as exc_info:
file_preview_api._validate_file_ownership(file_id, app_id)
assert "Upload file record not found" in str(exc_info.value)
def test_validate_file_ownership_tenant_mismatch(
self, file_preview_api: FilePreviewApi, mock_app, mock_upload_file, mock_message_file, mock_message
):
"""Test file ownership validation with tenant mismatch"""
file_id = str(uuid.uuid4())
app_id = mock_app.id
# Set up tenant mismatch
mock_upload_file.tenant_id = "different_tenant_id"
mock_app.tenant_id = "app_tenant_id"
mock_message.app_id = app_id
mock_message_file.upload_file_id = file_id
mock_message_file.message_id = mock_message.id
with patch("controllers.service_api.app.file_preview.db") as mock_db:
# Mock scalar() for MessageFile and Message queries
mock_db.session.scalar.side_effect = [
mock_message_file, # MessageFile query
mock_message, # Message query
]
# Mock get() for UploadFile and App PK lookups
mock_db.session.get.side_effect = [
mock_upload_file, # UploadFile query
mock_app, # App query for tenant validation
]
# Execute and assert exception
with pytest.raises(FileAccessDeniedError) as exc_info:
file_preview_api._validate_file_ownership(file_id, app_id)
assert "tenant mismatch" in str(exc_info.value)
with patch("controllers.service_api.app.file_preview.db", database):
with pytest.raises(FileAccessDeniedError, match="tenant mismatch"):
file_preview_api._validate_file_ownership(records.upload_file.id, records.app.id)
def test_validate_file_ownership_invalid_input(self, file_preview_api: FilePreviewApi):
"""Test file ownership validation with invalid input"""
# Test with empty file_id
with pytest.raises(FileAccessDeniedError) as exc_info:
with pytest.raises(FileAccessDeniedError, match="Invalid file or app identifier"):
file_preview_api._validate_file_ownership("", "app_id")
assert "Invalid file or app identifier" in str(exc_info.value)
# Test with empty app_id
with pytest.raises(FileAccessDeniedError) as exc_info:
with pytest.raises(FileAccessDeniedError, match="Invalid file or app identifier"):
file_preview_api._validate_file_ownership("file_id", "")
assert "Invalid file or app identifier" in str(exc_info.value)
def test_build_file_response_basic(self, file_preview_api: FilePreviewApi, mock_upload_file):
"""Test basic file response building"""
mock_generator = Mock()
@pytest.mark.parametrize(
("as_attachment", "mime_type", "name", "extension", "size"),
[
(False, "image/jpeg", "test_file.jpg", "jpg", 1024),
(True, "image/jpeg", "test_file.jpg", "jpg", 1024),
(False, "text/html", "unsafe.html", "html", 1024),
(False, "video/mp4", "test_file.mp4", "mp4", 1024),
(False, "image/jpeg", "test_file.jpg", "jpg", 0),
],
)
def test_build_file_response(
self,
file_preview_api: FilePreviewApi,
as_attachment: bool,
mime_type: str,
name: str,
extension: str,
size: int,
):
upload_file = _upload_file(tenant_id=str(uuid4()))
upload_file.mime_type = mime_type
upload_file.name = name
upload_file.extension = extension
upload_file.size = size
response = file_preview_api._build_file_response(mock_generator, mock_upload_file, False)
response = file_preview_api._build_file_response(Mock(), upload_file, as_attachment)
# Check response properties
assert response.mimetype == mock_upload_file.mime_type
assert response.direct_passthrough is True
assert response.headers["Content-Length"] == str(mock_upload_file.size)
assert "Cache-Control" in response.headers
def test_build_file_response_as_attachment(self, file_preview_api: FilePreviewApi, mock_upload_file):
"""Test file response building with attachment flag"""
mock_generator = Mock()
response = file_preview_api._build_file_response(mock_generator, mock_upload_file, True)
# Check attachment-specific headers
assert "attachment" in response.headers["Content-Disposition"]
assert mock_upload_file.name in response.headers["Content-Disposition"]
assert response.headers["Content-Type"] == "application/octet-stream"
def test_build_file_response_html_forces_attachment(self, file_preview_api: FilePreviewApi, mock_upload_file):
"""Test HTML files are forced to download"""
mock_generator = Mock()
mock_upload_file.mime_type = "text/html"
mock_upload_file.name = "unsafe.html"
mock_upload_file.extension = "html"
response = file_preview_api._build_file_response(mock_generator, mock_upload_file, False)
assert "attachment" in response.headers["Content-Disposition"]
assert response.headers["Content-Type"] == "application/octet-stream"
assert response.headers["X-Content-Type-Options"] == "nosniff"
def test_build_file_response_audio_video(self, file_preview_api: FilePreviewApi, mock_upload_file):
"""Test file response building for audio/video files"""
mock_generator = Mock()
mock_upload_file.mime_type = "video/mp4"
response = file_preview_api._build_file_response(mock_generator, mock_upload_file, False)
# Check Range support for media files
assert response.headers["Accept-Ranges"] == "bytes"
def test_build_file_response_no_size(self, file_preview_api: FilePreviewApi, mock_upload_file):
"""Test file response building when size is unknown"""
mock_generator = Mock()
mock_upload_file.size = 0 # Unknown size
response = file_preview_api._build_file_response(mock_generator, mock_upload_file, False)
# Content-Length should not be set when size is unknown
assert "Content-Length" not in response.headers
assert ("Content-Length" in response.headers) is bool(size)
if as_attachment or mime_type == "text/html":
assert "attachment" in response.headers["Content-Disposition"]
assert response.headers["Content-Type"] == "application/octet-stream"
else:
assert response.mimetype == mime_type
if mime_type == "text/html":
assert response.headers["X-Content-Type-Options"] == "nosniff"
if mime_type.startswith("video/"):
assert response.headers["Accept-Ranges"] == "bytes"
@patch("controllers.service_api.app.file_preview.storage")
def test_get_method_integration(
self,
mock_storage,
file_preview_api: FilePreviewApi,
mock_app,
mock_end_user,
mock_upload_file,
mock_message_file,
mock_message,
def test_components_use_validated_file(
self, mock_storage: Mock, file_preview_api: FilePreviewApi, database: _Database
):
"""Test the full GET method integration (without decorator)"""
file_id = str(uuid.uuid4())
app_id = mock_app.id
records = _persist_preview_records(database.session)
generator = Mock()
# Set up mocks
mock_upload_file.tenant_id = mock_app.tenant_id
mock_message.app_id = app_id
mock_message_file.upload_file_id = file_id
mock_message_file.message_id = mock_message.id
with patch("controllers.service_api.app.file_preview.db", database):
message_file, upload_file = file_preview_api._validate_file_ownership(
records.upload_file.id, records.app.id
)
response = file_preview_api._build_file_response(generator, upload_file, False)
mock_generator = Mock()
mock_storage.load.return_value = mock_generator
with patch("controllers.service_api.app.file_preview.db") as mock_db:
# Mock scalar() for MessageFile and Message queries
mock_db.session.scalar.side_effect = [
mock_message_file, # MessageFile query
mock_message, # Message query
]
# Mock get() for UploadFile and App PK lookups
mock_db.session.get.side_effect = [
mock_upload_file, # UploadFile query
mock_app, # App query for tenant validation
]
# Test the core logic directly without Flask decorators
# Validate file ownership
result_message_file, result_upload_file = file_preview_api._validate_file_ownership(file_id, app_id)
assert result_message_file == mock_message_file
assert result_upload_file == mock_upload_file
# Test file response building
response = file_preview_api._build_file_response(mock_generator, mock_upload_file, False)
assert response is not None
# Verify storage was called correctly
mock_storage.load.assert_not_called() # Since we're testing components separately
assert message_file.id == records.message_file.id
assert response.mimetype == "image/jpeg"
mock_storage.load.assert_not_called()
@patch("controllers.service_api.app.file_preview.storage")
def test_storage_error_handling(
self,
mock_storage,
file_preview_api: FilePreviewApi,
mock_app,
mock_upload_file,
mock_message_file,
mock_message,
def test_storage_error_remains_external(
self, mock_storage: Mock, file_preview_api: FilePreviewApi, database: _Database
):
"""Test storage error handling in the core logic"""
file_id = str(uuid.uuid4())
app_id = mock_app.id
records = _persist_preview_records(database.session)
mock_storage.load.side_effect = OSError("Storage error")
# Set up mocks
mock_upload_file.tenant_id = mock_app.tenant_id
mock_message.app_id = app_id
mock_message_file.upload_file_id = file_id
mock_message_file.message_id = mock_message.id
with patch("controllers.service_api.app.file_preview.db", database):
_, upload_file = file_preview_api._validate_file_ownership(records.upload_file.id, records.app.id)
# Mock storage error
mock_storage.load.side_effect = Exception("Storage error")
with patch("controllers.service_api.app.file_preview.db") as mock_db:
# Mock scalar() for MessageFile and Message queries
mock_db.session.scalar.side_effect = [
mock_message_file, # MessageFile query
mock_message, # Message query
]
# Mock get() for UploadFile and App PK lookups
mock_db.session.get.side_effect = [
mock_upload_file, # UploadFile query
mock_app, # App query for tenant validation
]
# First validate file ownership works
result_message_file, result_upload_file = file_preview_api._validate_file_ownership(file_id, app_id)
assert result_message_file == mock_message_file
assert result_upload_file == mock_upload_file
# Test storage error handling
with pytest.raises(Exception) as exc_info:
mock_storage.load(mock_upload_file.key, stream=True)
assert "Storage error" in str(exc_info.value)
with pytest.raises(OSError, match="Storage error"):
mock_storage.load(upload_file.key, stream=True)
def test_validate_file_ownership_unexpected_error_logging(
self, file_preview_api: FilePreviewApi, caplog: pytest.LogCaptureFixture
self,
file_preview_api: FilePreviewApi,
database: _Database,
sqlite_engine: Engine,
caplog: pytest.LogCaptureFixture,
):
"""Test that unexpected errors are logged properly"""
file_id = str(uuid.uuid4())
app_id = str(uuid.uuid4())
file_id = str(uuid4())
app_id = str(uuid4())
with patch("controllers.service_api.app.file_preview.db") as mock_db:
# Mock database scalar to raise unexpected exception
mock_db.session.scalar.side_effect = Exception("Unexpected database error")
def fail_statement(*_args: object) -> None:
raise RuntimeError("Unexpected database error")
# Execute and assert exception
with caplog.at_level(logging.ERROR, logger="controllers.service_api.app.file_preview"):
with pytest.raises(FileAccessDeniedError) as exc_info:
file_preview_api._validate_file_ownership(file_id, app_id)
event.listen(sqlite_engine, "before_cursor_execute", fail_statement)
try:
with patch("controllers.service_api.app.file_preview.db", database):
with caplog.at_level(logging.ERROR, logger="controllers.service_api.app.file_preview"):
with pytest.raises(FileAccessDeniedError, match="File access validation failed"):
file_preview_api._validate_file_ownership(file_id, app_id)
finally:
event.remove(sqlite_engine, "before_cursor_execute", fail_statement)
# Verify error message
assert "File access validation failed" in str(exc_info.value)
# Verify logging was called with the structured context fields. The ``extra`` keys
# are attached to the LogRecord as attributes, so they are not in ``caplog.text``.
assert len(caplog.records) == 1
log_record = caplog.records[0]
assert log_record.getMessage() == "Unexpected error during file ownership validation"
record = cast(_FilePreviewLogRecord, log_record)
assert record.file_id == file_id
assert record.app_id == app_id
assert record.error == "Unexpected database error"
assert len(caplog.records) == 1
log_record = caplog.records[0]
assert log_record.getMessage() == "Unexpected error during file ownership validation"
record = cast(_FilePreviewLogRecord, log_record)
assert record.file_id == file_id
assert record.app_id == app_id
assert record.error == "Unexpected database error"
@@ -0,0 +1,89 @@
from unittest.mock import MagicMock
from controllers.service_api.app.legacy_system_files import (
attach_legacy_system_file_warning_for_service_api,
normalize_legacy_system_file_args_for_service_api,
)
from core.app.entities.app_invoke_entities import InvokeFrom
from services.app_generate_service import AppGenerateService
_LEGACY_FILE_TEMPLATE = "{{#" + ".".join(("sys", "files")) + "#}}"
_USER_INPUT_FILE_INPUT_KEY = ".".join(("userinput", "files"))
def _legacy_file_graph() -> dict:
return {
"nodes": [
{"id": "start", "data": {"type": "start", "variables": []}},
{"id": "answer", "data": {"type": "answer", "answer": _LEGACY_FILE_TEMPLATE}},
],
"edges": [],
}
def test_hidden_service_api_file_payload_maps_to_userinput_files(mocker):
workflow = MagicMock()
workflow.graph_dict = _legacy_file_graph()
get_workflow = mocker.patch.object(AppGenerateService, "get_workflow", return_value=workflow)
app_model = MagicMock()
session = MagicMock()
files = [{"transfer_method": "remote_url", "url": "https://example.com/a.png"}]
args, compat_variable = normalize_legacy_system_file_args_for_service_api(
session=session,
app_model=app_model,
args={"inputs": {}, "files": None},
raw_payload={"system": {"files": files}},
)
get_workflow.assert_called_once_with(app_model, InvokeFrom.SERVICE_API, None, session=session)
assert compat_variable is not None
assert args["files"] == files
assert args["inputs"][_USER_INPUT_FILE_INPUT_KEY] == files
def test_service_api_file_payload_is_ignored_when_absent(mocker):
get_workflow = mocker.patch.object(AppGenerateService, "get_workflow")
app_model = MagicMock()
original_args = {"inputs": {}}
args, compat_variable = normalize_legacy_system_file_args_for_service_api(
session=MagicMock(),
app_model=app_model,
args=original_args,
raw_payload={},
)
assert args is original_args
assert compat_variable is None
get_workflow.assert_not_called()
def test_top_level_service_api_file_payload_still_checks_workflow_graph(mocker):
workflow = MagicMock()
workflow.graph_dict = {"nodes": []}
get_workflow = mocker.patch.object(AppGenerateService, "get_workflow", return_value=workflow)
app_model = MagicMock()
session = MagicMock()
files = [{"id": "file-1"}]
args, compat_variable = normalize_legacy_system_file_args_for_service_api(
session=session,
app_model=app_model,
args={"inputs": {}, "files": files},
raw_payload={},
)
get_workflow.assert_called_once_with(app_model, InvokeFrom.SERVICE_API, None, session=session)
assert args["files"] == files
assert compat_variable is None
def test_service_api_warning_is_attached_only_when_compatibility_was_used():
compat_variable = MagicMock(node_id="userinput", variable_name="files")
response = attach_legacy_system_file_warning_for_service_api({"answer": "ok"}, compat_variable)
response_without_warning = attach_legacy_system_file_warning_for_service_api({"answer": "ok"}, None)
assert response["warnings"]
assert response_without_warning == {"answer": "ok"}
@@ -15,12 +15,15 @@ Focus on:
"""
import uuid
from collections.abc import Iterator
from inspect import unwrap
from types import SimpleNamespace
from unittest.mock import Mock, patch
import pytest
from flask import Flask
from sqlalchemy import Engine
from sqlalchemy.orm import Session
from werkzeug.exceptions import BadRequest, InternalServerError, NotFound
from controllers.service_api.app.error import NotChatAppError
@@ -44,6 +47,14 @@ from services.errors.message import (
from services.message_service import MessageService
@pytest.fixture
def orm_session(sqlite_engine: Engine) -> Iterator[Session]:
"""Provide a real caller-owned session for MessageService interface tests."""
with Session(sqlite_engine, expire_on_commit=False) as session:
yield session
class TestMessageListQuery:
"""Test suite for MessageListQuery Pydantic model."""
@@ -253,7 +264,7 @@ class TestMessageService:
assert callable(MessageService.get_suggested_questions_after_answer)
@patch.object(MessageService, "pagination_by_first_id")
def test_pagination_by_first_id_returns_pagination_result(self, mock_pagination):
def test_pagination_by_first_id_returns_pagination_result(self, mock_pagination, orm_session: Session):
"""Test pagination_by_first_id returns expected format."""
mock_result = Mock()
mock_result.data = []
@@ -267,7 +278,7 @@ class TestMessageService:
conversation_id=str(uuid.uuid4()),
first_id=None,
limit=20,
session=Mock(),
session=orm_session,
)
assert hasattr(result, "data")
@@ -275,7 +286,7 @@ class TestMessageService:
assert hasattr(result, "has_more")
@patch.object(MessageService, "pagination_by_first_id")
def test_pagination_raises_conversation_not_exists_error(self, mock_pagination):
def test_pagination_raises_conversation_not_exists_error(self, mock_pagination, orm_session: Session):
"""Test pagination raises ConversationNotExistsError."""
import services.errors.conversation
@@ -288,11 +299,11 @@ class TestMessageService:
conversation_id="invalid_id",
first_id=None,
limit=20,
session=Mock(),
session=orm_session,
)
@patch.object(MessageService, "pagination_by_first_id")
def test_pagination_raises_first_message_not_exists_error(self, mock_pagination):
def test_pagination_raises_first_message_not_exists_error(self, mock_pagination, orm_session: Session):
"""Test pagination raises FirstMessageNotExistsError."""
mock_pagination.side_effect = FirstMessageNotExistsError()
@@ -303,11 +314,11 @@ class TestMessageService:
conversation_id=str(uuid.uuid4()),
first_id="invalid_first_id",
limit=20,
session=Mock(),
session=orm_session,
)
@patch.object(MessageService, "create_feedback")
def test_create_feedback_with_rating_and_content(self, mock_create_feedback):
def test_create_feedback_with_rating_and_content(self, mock_create_feedback, orm_session: Session):
"""Test create_feedback with rating and content."""
mock_create_feedback.return_value = None
@@ -317,13 +328,13 @@ class TestMessageService:
user=Mock(spec=EndUser),
rating=FeedbackRating.LIKE,
content="Great response!",
session=Mock(),
session=orm_session,
)
mock_create_feedback.assert_called_once()
@patch.object(MessageService, "create_feedback")
def test_create_feedback_raises_message_not_exists_error(self, mock_create_feedback):
def test_create_feedback_raises_message_not_exists_error(self, mock_create_feedback, orm_session: Session):
"""Test create_feedback raises MessageNotExistsError."""
mock_create_feedback.side_effect = MessageNotExistsError()
@@ -334,11 +345,11 @@ class TestMessageService:
user=Mock(spec=EndUser),
rating=FeedbackRating.LIKE,
content=None,
session=Mock(),
session=orm_session,
)
@patch.object(MessageService, "get_all_messages_feedbacks")
def test_get_all_messages_feedbacks_returns_list(self, mock_get_feedbacks):
def test_get_all_messages_feedbacks_returns_list(self, mock_get_feedbacks, orm_session: Session):
"""Test get_all_messages_feedbacks returns list of feedbacks."""
mock_feedbacks = [
{"message_id": str(uuid.uuid4()), "rating": "like"},
@@ -346,13 +357,15 @@ class TestMessageService:
]
mock_get_feedbacks.return_value = mock_feedbacks
result = MessageService.get_all_messages_feedbacks(app_model=Mock(spec=App), page=1, limit=20, session=Mock())
result = MessageService.get_all_messages_feedbacks(
app_model=Mock(spec=App), page=1, limit=20, session=orm_session
)
assert len(result) == 2
assert result[0]["rating"] == "like"
@patch.object(MessageService, "get_suggested_questions_after_answer")
def test_get_suggested_questions_returns_questions_list(self, mock_get_questions):
def test_get_suggested_questions_returns_questions_list(self, mock_get_questions, orm_session: Session):
"""Test get_suggested_questions_after_answer returns list of questions."""
mock_questions = ["What about this aspect?", "Can you elaborate on that?", "How does this relate to...?"]
mock_get_questions.return_value = mock_questions
@@ -362,14 +375,14 @@ class TestMessageService:
user=Mock(spec=EndUser),
message_id=str(uuid.uuid4()),
invoke_from=Mock(),
session=Mock(),
session=orm_session,
)
assert len(result) == 3
assert isinstance(result[0], str)
@patch.object(MessageService, "get_suggested_questions_after_answer")
def test_get_suggested_questions_raises_disabled_error(self, mock_get_questions):
def test_get_suggested_questions_raises_disabled_error(self, mock_get_questions, orm_session: Session):
"""Test get_suggested_questions_after_answer raises SuggestedQuestionsAfterAnswerDisabledError."""
mock_get_questions.side_effect = SuggestedQuestionsAfterAnswerDisabledError()
@@ -379,11 +392,11 @@ class TestMessageService:
user=Mock(spec=EndUser),
message_id=str(uuid.uuid4()),
invoke_from=Mock(),
session=Mock(),
session=orm_session,
)
@patch.object(MessageService, "get_suggested_questions_after_answer")
def test_get_suggested_questions_raises_message_not_exists_error(self, mock_get_questions):
def test_get_suggested_questions_raises_message_not_exists_error(self, mock_get_questions, orm_session: Session):
"""Test get_suggested_questions_after_answer raises MessageNotExistsError."""
mock_get_questions.side_effect = MessageNotExistsError()
@@ -393,7 +406,7 @@ class TestMessageService:
user=Mock(spec=EndUser),
message_id="invalid_message_id",
invoke_from=Mock(),
session=Mock(),
session=orm_session,
)
@@ -24,6 +24,7 @@ from unittest.mock import Mock, patch
import pytest
from flask import Flask
from sqlalchemy.orm import Session
from werkzeug.datastructures import FileStorage
from werkzeug.exceptions import Forbidden, NotFound
@@ -38,6 +39,7 @@ from controllers.service_api.dataset.rag_pipeline.rag_pipeline_workflow import (
)
from core.app.entities.app_invoke_entities import InvokeFrom
from models.account import Account
from models.dataset import Dataset
from services.errors.file import FileTooLargeError, UnsupportedFileTypeError
from services.rag_pipeline.entity.pipeline_service_api_entities import (
DatasourceNodeRunApiEntity,
@@ -46,6 +48,20 @@ from services.rag_pipeline.entity.pipeline_service_api_entities import (
from services.rag_pipeline.rag_pipeline import RagPipelineService
def _persist_dataset(session: Session, *, tenant_id: str, dataset_id: str) -> Dataset:
dataset = Dataset(
id=dataset_id,
tenant_id=tenant_id,
name="Pipeline dataset",
created_by="account-1",
data_source_type=None,
indexing_technique=None,
)
session.add(dataset)
session.commit()
return dataset
class TestDatasourceNodeRunPayload:
"""Test suite for DatasourceNodeRunPayload Pydantic model."""
@@ -550,13 +566,15 @@ class TestPipelineRunApiPost:
)
@patch("controllers.service_api.dataset.rag_pipeline.rag_pipeline_workflow.RagPipelineService")
@patch("controllers.service_api.dataset.rag_pipeline.rag_pipeline_workflow.service_api_ns")
def test_post_success_streaming(self, mock_ns, mock_svc_cls, mock_current_user, mock_gen_svc, mock_helper, app):
@pytest.mark.parametrize("sqlite_session", [(Dataset,)], indirect=True)
def test_post_success_streaming(
self, mock_ns, mock_svc_cls, mock_current_user, mock_gen_svc, mock_helper, app, sqlite_session: Session
):
"""Test successful pipeline run with streaming response."""
tenant_id = str(uuid.uuid4())
dataset_id = str(uuid.uuid4())
session = Mock()
session.scalar.return_value = Mock()
_persist_dataset(sqlite_session, tenant_id=tenant_id, dataset_id=dataset_id)
mock_ns.payload = {
"inputs": {"key": "val"},
@@ -577,33 +595,33 @@ class TestPipelineRunApiPost:
with app.test_request_context("/datasets/test/pipeline/run", method="POST"):
api = PipelineRunApi()
response = api.post.__wrapped__(api, session, tenant_id=tenant_id, dataset_id=dataset_id)
response = api.post.__wrapped__(api, sqlite_session, tenant_id=tenant_id, dataset_id=dataset_id)
assert response == {"result": "ok"}
mock_svc_cls.assert_called_once_with(session)
mock_svc_cls.assert_called_once_with(sqlite_session)
mock_gen_svc.generate.assert_called_once()
def test_post_not_found(self, app: Flask):
@pytest.mark.parametrize("sqlite_session", [(Dataset,)], indirect=True)
def test_post_not_found(self, app: Flask, sqlite_session: Session):
"""Test NotFound when dataset check fails."""
session = Mock()
session.scalar.return_value = None
with app.test_request_context("/datasets/test/pipeline/run", method="POST"):
api = PipelineRunApi()
with pytest.raises(NotFound):
api.post.__wrapped__(
api,
session,
sqlite_session,
tenant_id=str(uuid.uuid4()),
dataset_id=str(uuid.uuid4()),
)
@patch("controllers.service_api.dataset.rag_pipeline.rag_pipeline_workflow.current_user", new="not_account")
@patch("controllers.service_api.dataset.rag_pipeline.rag_pipeline_workflow.service_api_ns")
def test_post_forbidden_non_account_user(self, mock_ns, app: Flask):
@pytest.mark.parametrize("sqlite_session", [(Dataset,)], indirect=True)
def test_post_forbidden_non_account_user(self, mock_ns, app: Flask, sqlite_session: Session):
"""Test Forbidden when current_user is not an Account."""
session = Mock()
session.scalar.return_value = Mock()
tenant_id = str(uuid.uuid4())
dataset_id = str(uuid.uuid4())
_persist_dataset(sqlite_session, tenant_id=tenant_id, dataset_id=dataset_id)
mock_ns.payload = {
"inputs": {},
"datasource_type": "online_document",
@@ -618,9 +636,9 @@ class TestPipelineRunApiPost:
with pytest.raises(Forbidden):
api.post.__wrapped__(
api,
session,
tenant_id=str(uuid.uuid4()),
dataset_id=str(uuid.uuid4()),
sqlite_session,
tenant_id=tenant_id,
dataset_id=dataset_id,
)
@@ -3,10 +3,13 @@ Unit tests for Service API wraps (authentication decorators)
"""
import uuid
from types import SimpleNamespace
from unittest.mock import MagicMock, Mock, patch
import pytest
from flask import Flask
from sqlalchemy import select
from sqlalchemy.orm import Session
from werkzeug.exceptions import Forbidden, NotFound, Unauthorized
from controllers.service_api.wraps import (
@@ -21,12 +24,11 @@ from controllers.service_api.wraps import (
validate_dataset_token,
)
from enums.cloud_plan import CloudPlan
from models.account import TenantStatus
from models.model import ApiToken
from tests.unit_tests.conftest import (
setup_mock_dataset_owner_execute_result,
setup_mock_tenant_owner_execute_result,
)
from models import Account, Tenant, TenantAccountJoin
from models.account import TenantAccountRole
from models.dataset import Dataset, RateLimitLog
from models.enums import ApiTokenType
from models.model import ApiToken, App, AppMode, IconType
def _configure_current_app_mock(mock_current_app):
@@ -34,6 +36,51 @@ def _configure_current_app_mock(mock_current_app):
mock_current_app._get_current_object = Mock(return_value=Mock())
def _session_proxy(session: Session) -> MagicMock:
"""Emulate Flask-SQLAlchemy's callable scoped-session proxy around a test session."""
proxy = MagicMock(wraps=session)
proxy.return_value = session
return proxy
def _api_token(*, tenant_id: str, app_id: str | None = None, token_type: ApiTokenType) -> ApiToken:
return ApiToken(
id=str(uuid.uuid4()),
tenant_id=tenant_id,
app_id=app_id,
type=token_type,
token="test_token",
)
def _persist_workspace(session: Session) -> tuple[Tenant, Account, TenantAccountJoin]:
tenant = Tenant(name="Workspace")
account = Account(name="Owner", email=f"owner-{uuid.uuid4()}@example.com")
membership = TenantAccountJoin(
tenant_id=tenant.id,
account_id=account.id,
current=True,
role=TenantAccountRole.OWNER,
)
session.add_all([tenant, account, membership])
session.commit()
return tenant, account, membership
def _app_model(*, tenant_id: str, enable_api: bool = True) -> App:
return App(
id=str(uuid.uuid4()),
tenant_id=tenant_id,
name="Service API App",
mode=AppMode.CHAT,
icon_type=IconType.EMOJI,
icon="chat",
icon_background="#FFFFFF",
enable_site=False,
enable_api=enable_api,
)
class TestValidateAndGetApiToken:
"""Test suite for validate_and_get_api_token function"""
@@ -70,21 +117,24 @@ class TestValidateAndGetApiToken:
def test_valid_token_returns_api_token(self, mock_fetch_token, mock_cache_cls, mock_record_usage, app: Flask):
"""Test that valid token returns the ApiToken object."""
# Arrange
mock_api_token = Mock(spec=ApiToken)
mock_api_token.token = "valid_token_123"
mock_api_token.type = "app"
api_token = _api_token(
tenant_id=str(uuid.uuid4()),
app_id=str(uuid.uuid4()),
token_type=ApiTokenType.APP,
)
api_token.token = "valid_token_123"
mock_cache_instance = Mock()
mock_cache_instance.get.return_value = None # Cache miss
mock_cache_cls.get = mock_cache_instance.get
mock_fetch_token.return_value = mock_api_token
mock_fetch_token.return_value = api_token
# Act
with app.test_request_context("/", method="GET", headers={"Authorization": "Bearer valid_token_123"}):
result = validate_and_get_api_token("app")
# Assert
assert result == mock_api_token
assert result == api_token
@patch("controllers.service_api.wraps.record_token_usage")
@patch("controllers.service_api.wraps.ApiTokenCache")
@@ -117,116 +167,124 @@ class TestValidateAppToken:
return app
@patch("controllers.service_api.wraps.user_logged_in")
@patch("controllers.service_api.wraps.db")
@patch("controllers.service_api.wraps.validate_and_get_api_token")
@patch("controllers.service_api.wraps.current_app")
@pytest.mark.parametrize(
"sqlite_session",
[(App, ApiToken, Tenant, Account, TenantAccountJoin)],
indirect=True,
)
def test_valid_app_token_allows_access(
self, mock_current_app, mock_validate_token, mock_db, mock_user_logged_in, app
self,
mock_current_app,
mock_validate_token,
mock_user_logged_in,
app: Flask,
sqlite_session: Session,
):
"""Test that valid app token allows access to decorated view."""
# Arrange
_configure_current_app_mock(mock_current_app)
mock_api_token = Mock()
mock_api_token.app_id = str(uuid.uuid4())
mock_api_token.tenant_id = str(uuid.uuid4())
mock_validate_token.return_value = mock_api_token
mock_app = Mock()
mock_app.id = mock_api_token.app_id
mock_app.status = "normal"
mock_app.enable_api = True
mock_app.tenant_id = mock_api_token.tenant_id
mock_tenant = Mock()
mock_tenant.status = TenantStatus.NORMAL
mock_tenant.id = mock_api_token.tenant_id
mock_account = Mock()
mock_account.id = str(uuid.uuid4())
# Use side_effect to return app first, then tenant via session.get()
mock_db.session.get.side_effect = [mock_app, mock_tenant]
# Mock the tenant owner execute result (execute(select(...)).one_or_none())
setup_mock_tenant_owner_execute_result(mock_db, mock_tenant, mock_account)
tenant, account, _ = _persist_workspace(sqlite_session)
app_model = _app_model(tenant_id=tenant.id)
api_token = _api_token(tenant_id=tenant.id, app_id=app_model.id, token_type=ApiTokenType.APP)
sqlite_session.add_all([app_model, api_token])
sqlite_session.commit()
mock_validate_token.return_value = api_token
@validate_app_token
def protected_view(app_model):
return {"success": True, "app_id": app_model.id}
# Act
with app.test_request_context("/", method="GET", headers={"Authorization": "Bearer test_token"}):
with (
app.test_request_context("/", method="GET", headers={"Authorization": "Bearer test_token"}),
patch("controllers.service_api.wraps.db.session", _session_proxy(sqlite_session)),
):
result = protected_view()
# Assert
assert result["success"] is True
assert result["app_id"] == mock_app.id
assert result["app_id"] == app_model.id
assert account.current_tenant_id == tenant.id
@patch("controllers.service_api.wraps.db")
@patch("controllers.service_api.wraps.validate_and_get_api_token")
def test_app_not_found_raises_forbidden(self, mock_validate_token, mock_db, app: Flask):
@pytest.mark.parametrize("sqlite_session", [(App,)], indirect=True)
def test_app_not_found_raises_forbidden(self, mock_validate_token, app: Flask, sqlite_session: Session):
"""Test that Forbidden is raised when app no longer exists."""
# Arrange
mock_api_token = Mock()
mock_api_token.app_id = str(uuid.uuid4())
mock_validate_token.return_value = mock_api_token
mock_db.session.get.return_value = None
api_token = _api_token(
tenant_id=str(uuid.uuid4()),
app_id=str(uuid.uuid4()),
token_type=ApiTokenType.APP,
)
mock_validate_token.return_value = api_token
@validate_app_token
def protected_view(**kwargs):
return {"success": True}
# Act & Assert
with app.test_request_context("/", method="GET"):
with (
app.test_request_context("/", method="GET"),
patch("controllers.service_api.wraps.db.session", sqlite_session),
):
with pytest.raises(Forbidden) as exc_info:
protected_view()
assert "no longer exists" in str(exc_info.value)
@patch("controllers.service_api.wraps.db")
@patch("controllers.service_api.wraps.validate_and_get_api_token")
def test_app_status_abnormal_raises_forbidden(self, mock_validate_token, mock_db, app: Flask):
@pytest.mark.parametrize("sqlite_session", [(App,)], indirect=True)
def test_app_status_abnormal_raises_forbidden(self, mock_validate_token, app: Flask, sqlite_session: Session):
"""Test that Forbidden is raised when app status is abnormal."""
# Arrange
mock_api_token = Mock()
mock_api_token.app_id = str(uuid.uuid4())
mock_validate_token.return_value = mock_api_token
mock_app = Mock()
mock_app.status = "abnormal"
mock_db.session.get.return_value = mock_app
app_model = _app_model(tenant_id=str(uuid.uuid4()))
sqlite_session.add(app_model)
sqlite_session.commit()
app_model.status = "abnormal"
mock_validate_token.return_value = _api_token(
tenant_id=app_model.tenant_id,
app_id=app_model.id,
token_type=ApiTokenType.APP,
)
@validate_app_token
def protected_view(**kwargs):
return {"success": True}
# Act & Assert
with app.test_request_context("/", method="GET"):
with (
app.test_request_context("/", method="GET"),
patch("controllers.service_api.wraps.db.session", sqlite_session),
):
with pytest.raises(Forbidden) as exc_info:
protected_view()
assert "status is abnormal" in str(exc_info.value)
@patch("controllers.service_api.wraps.db")
@patch("controllers.service_api.wraps.validate_and_get_api_token")
def test_app_api_disabled_raises_forbidden(self, mock_validate_token, mock_db, app: Flask):
@pytest.mark.parametrize("sqlite_session", [(App,)], indirect=True)
def test_app_api_disabled_raises_forbidden(self, mock_validate_token, app: Flask, sqlite_session: Session):
"""Test that Forbidden is raised when app API is disabled."""
# Arrange
mock_api_token = Mock()
mock_api_token.app_id = str(uuid.uuid4())
mock_validate_token.return_value = mock_api_token
mock_app = Mock()
mock_app.status = "normal"
mock_app.enable_api = False
mock_db.session.get.return_value = mock_app
app_model = _app_model(tenant_id=str(uuid.uuid4()), enable_api=False)
sqlite_session.add(app_model)
sqlite_session.commit()
mock_validate_token.return_value = _api_token(
tenant_id=app_model.tenant_id,
app_id=app_model.id,
token_type=ApiTokenType.APP,
)
@validate_app_token
def protected_view(**kwargs):
return {"success": True}
# Act & Assert
with app.test_request_context("/", method="GET"):
with (
app.test_request_context("/", method="GET"),
patch("controllers.service_api.wraps.db.session", sqlite_session),
):
with pytest.raises(Forbidden) as exc_info:
protected_view()
assert "API service has been disabled" in str(exc_info.value)
@@ -468,26 +526,35 @@ class TestCloudEditionBillingRateLimitCheck:
@patch("controllers.service_api.wraps.validate_and_get_api_token")
@patch("controllers.service_api.wraps.FeatureService.get_knowledge_rate_limit")
@patch("controllers.service_api.wraps.db")
@patch("controllers.service_api.wraps.sessionmaker")
@pytest.mark.parametrize("sqlite_session", [(RateLimitLog,)], indirect=True)
def test_rejects_over_rate_limit(
self, mock_sessionmaker, mock_db, mock_get_rate_limit, mock_validate_token, app: Flask
self,
mock_get_rate_limit,
mock_validate_token,
app: Flask,
sqlite_session: Session,
):
"""Test that Forbidden is raised when over rate limit."""
# Arrange
mock_validate_token.return_value = Mock(tenant_id="tenant123")
tenant_id = str(uuid.uuid4())
mock_validate_token.return_value = _api_token(
tenant_id=tenant_id,
token_type=ApiTokenType.DATASET,
)
mock_rate_limit = Mock()
mock_rate_limit.enabled = True
mock_rate_limit.limit = 10
mock_rate_limit.subscription_plan = "pro"
mock_get_rate_limit.return_value = mock_rate_limit
rate_limit_log_session = MagicMock()
session_factory = MagicMock()
session_factory.begin.return_value.__enter__.return_value = rate_limit_log_session
mock_sessionmaker.return_value = session_factory
with patch("controllers.service_api.wraps.redis_client") as mock_redis:
with (
patch("controllers.service_api.wraps.redis_client") as mock_redis,
patch(
"controllers.service_api.wraps.db",
SimpleNamespace(engine=sqlite_session.get_bind()),
),
):
mock_redis.zcard.return_value = 15 # Over limit
@cloud_edition_billing_rate_limit_check("knowledge", "dataset")
@@ -499,9 +566,12 @@ class TestCloudEditionBillingRateLimitCheck:
with pytest.raises(Forbidden) as exc_info:
knowledge_request()
assert "rate limit" in str(exc_info.value)
mock_sessionmaker.assert_called_once_with(bind=mock_db.engine, expire_on_commit=False)
rate_limit_log_session.add.assert_called_once()
mock_db.session.commit.assert_not_called()
persisted_logs = sqlite_session.scalars(select(RateLimitLog)).all()
assert len(persisted_logs) == 1
assert persisted_logs[0].tenant_id == tenant_id
assert persisted_logs[0].subscription_plan == "pro"
assert persisted_logs[0].operation == "knowledge"
class TestValidateDatasetToken:
@@ -515,65 +585,62 @@ class TestValidateDatasetToken:
return app
@patch("controllers.service_api.wraps.user_logged_in")
@patch("controllers.service_api.wraps.db")
@patch("controllers.service_api.wraps.validate_and_get_api_token")
@patch("controllers.service_api.wraps.current_app")
def test_valid_dataset_token(self, mock_current_app, mock_validate_token, mock_db, mock_user_logged_in, app: Flask):
@pytest.mark.parametrize(
"sqlite_session",
[(Tenant, Account, TenantAccountJoin)],
indirect=True,
)
def test_valid_dataset_token(
self,
mock_current_app,
mock_validate_token,
mock_user_logged_in,
app: Flask,
sqlite_session: Session,
):
"""Test that valid dataset token allows access."""
# Arrange
_configure_current_app_mock(mock_current_app)
tenant_id = str(uuid.uuid4())
mock_api_token = Mock()
mock_api_token.tenant_id = tenant_id
mock_validate_token.return_value = mock_api_token
mock_tenant = Mock()
mock_tenant.id = tenant_id
mock_tenant.status = TenantStatus.NORMAL
mock_ta = Mock()
mock_ta.account_id = str(uuid.uuid4())
mock_account = Mock()
mock_account.id = mock_ta.account_id
mock_account.current_tenant = mock_tenant
# Mock the tenant account join query (execute(select(...)).one_or_none())
setup_mock_dataset_owner_execute_result(mock_db, mock_tenant, mock_ta)
# Mock the account lookup via session.get()
mock_db.session.get.return_value = mock_account
tenant, account, _ = _persist_workspace(sqlite_session)
api_token = _api_token(tenant_id=tenant.id, token_type=ApiTokenType.DATASET)
mock_validate_token.return_value = api_token
@validate_dataset_token
def protected_view(tenant_id):
return {"success": True, "tenant_id": tenant_id}
# Act
with app.test_request_context("/", method="GET", headers={"Authorization": "Bearer test_token"}):
with (
app.test_request_context("/", method="GET", headers={"Authorization": "Bearer test_token"}),
patch("controllers.service_api.wraps.db.session", _session_proxy(sqlite_session)),
):
result = protected_view()
# Assert
assert result["success"] is True
assert result["tenant_id"] == tenant_id
assert result["tenant_id"] == tenant.id
assert account.current_tenant_id == tenant.id
@patch("controllers.service_api.wraps.db")
@patch("controllers.service_api.wraps.validate_and_get_api_token")
def test_dataset_not_found_raises_not_found(self, mock_validate_token, mock_db, app: Flask):
@pytest.mark.parametrize("sqlite_session", [(Dataset,)], indirect=True)
def test_dataset_not_found_raises_not_found(self, mock_validate_token, app: Flask, sqlite_session: Session):
"""Test that NotFound is raised when dataset doesn't exist."""
# Arrange
mock_api_token = Mock()
mock_api_token.tenant_id = str(uuid.uuid4())
mock_validate_token.return_value = mock_api_token
mock_db.session.scalar.return_value = None
api_token = _api_token(tenant_id=str(uuid.uuid4()), token_type=ApiTokenType.DATASET)
mock_validate_token.return_value = api_token
@validate_dataset_token
def protected_view(dataset_id=None, **kwargs):
return {"success": True}
# Act & Assert
with app.test_request_context("/", method="GET"):
with (
app.test_request_context("/", method="GET"),
patch("controllers.service_api.wraps.db.session", sqlite_session),
):
with pytest.raises(NotFound) as exc_info:
protected_view(dataset_id=str(uuid.uuid4()))
assert "Dataset not found" in str(exc_info.value)
@@ -46,7 +46,7 @@ class TestAdvancedChatAppGeneratorValidation:
with pytest.raises(ValueError, match="query must be a string"):
generator.generate(
app_model=SimpleNamespace(),
workflow=SimpleNamespace(),
workflow=SimpleNamespace(graph_dict={"nodes": []}),
user=SimpleNamespace(),
args={"inputs": {}, "query": 123},
invoke_from=InvokeFrom.WEB_APP,
@@ -186,7 +186,7 @@ class TestAdvancedChatAppGeneratorInternals:
result = generator.generate(
app_model=SimpleNamespace(id="app", tenant_id="tenant"),
workflow=SimpleNamespace(features_dict={}),
workflow=SimpleNamespace(features_dict={}, graph_dict={"nodes": []}),
user=user,
args={
"query": "hello",
@@ -442,6 +442,13 @@ class TestAdvancedChatAppGeneratorInternals:
def start(self):
thread_data["started"] = True
def join(self, timeout):
thread_data["joined"] = True
thread_data["join_timeout"] = timeout
def is_alive(self):
return False
monkeypatch.setattr("core.app.apps.advanced_chat.app_generator.threading.Thread", _Thread)
monkeypatch.setattr(
"core.app.apps.advanced_chat.app_generator.db", SimpleNamespace(engine=object(), session=db_session)
@@ -475,6 +482,8 @@ class TestAdvancedChatAppGeneratorInternals:
assert response["response"] == {"raw": True}
assert thread_data["started"] is True
assert thread_data["joined"] is True
assert thread_data["join_timeout"] == 300
assert "pause-layer" in thread_data["kwargs"]["graph_engine_layers"]
assert generator._dialogue_count == 3
assert init_records.call_args.kwargs["session"] is db_session
@@ -542,6 +551,13 @@ class TestAdvancedChatAppGeneratorInternals:
def start(self):
thread_data["started"] = True
def join(self, timeout):
thread_data["joined"] = True
thread_data["join_timeout"] = timeout
def is_alive(self):
return False
monkeypatch.setattr("core.app.apps.advanced_chat.app_generator.threading.Thread", _Thread)
monkeypatch.setattr(
"core.app.apps.advanced_chat.app_generator.db", SimpleNamespace(engine=object(), session=db_session)
@@ -574,6 +590,8 @@ class TestAdvancedChatAppGeneratorInternals:
init_records.assert_not_called()
get_thread_messages_length.assert_called_once_with(conversation.id, session=db_session)
assert thread_data["started"] is True
assert thread_data["joined"] is True
assert thread_data["join_timeout"] == 300
db_session.commit.assert_not_called()
db_session.refresh.assert_not_called()
db_session.close.assert_called_once()
@@ -1191,7 +1209,7 @@ class TestAdvancedChatAppGeneratorInternals:
monkeypatch.setattr(generator, "_generate", _fake_generate)
app_model = SimpleNamespace(id="app", tenant_id="tenant")
workflow = SimpleNamespace(features_dict={})
workflow = SimpleNamespace(features_dict={}, graph_dict={"nodes": []})
from models import Account
user = Account(name="Tester", email="tester@example.com")
@@ -1271,7 +1289,7 @@ class TestAdvancedChatAppGeneratorInternals:
monkeypatch.setattr(generator, "_generate", _fake_generate)
app_model = SimpleNamespace(id="app", tenant_id="tenant")
workflow = SimpleNamespace(features_dict={})
workflow = SimpleNamespace(features_dict={}, graph_dict={"nodes": []})
from models.model import EndUser
user = EndUser(tenant_id="tenant", type="session", name="tester", session_id="session")
@@ -22,7 +22,7 @@ def _build_converter() -> WorkflowResponseConverter:
app_config=SimpleNamespace(app_id="app-1", tenant_id="tenant-1"),
invoke_from=InvokeFrom.EXPLORE,
files=[],
inputs={},
inputs={"userinput.files": []},
workflow_execution_id="run-1",
call_depth=0,
)
@@ -54,3 +54,17 @@ def test_workflow_start_stream_response_carries_initial_reason():
reason=WorkflowStartReason.INITIAL,
)
assert resp.data.reason is WorkflowStartReason.INITIAL
def test_workflow_start_stream_response_exposes_only_canonical_file_input():
converter = _build_converter()
resp = converter.workflow_start_to_stream_response(
task_id="task-1",
workflow_run_id="run-1",
workflow_id="wf-1",
reason=WorkflowStartReason.INITIAL,
)
assert resp.data.inputs["userinput.files"] == []
assert "sys.files" not in resp.data.inputs
@@ -435,6 +435,7 @@ def test_generate_success_returns_converted(generator, mocker: MockerFixture):
mocker.patch.object(module, "PipelineQueueManager", return_value=queue_manager)
worker_thread = MagicMock()
worker_thread.is_alive.return_value = False
mocker.patch.object(module.threading, "Thread", return_value=worker_thread)
mocker.patch.object(generator, "_get_draft_var_saver_factory", return_value=MagicMock())
@@ -461,6 +462,7 @@ def test_generate_success_returns_converted(generator, mocker: MockerFixture):
)
assert result == "converted"
worker_thread.join.assert_called_once_with(timeout=300)
def test_single_iteration_generate_validates_inputs(generator, mocker: MockerFixture):
@@ -1,3 +1,6 @@
import logging
from unittest.mock import Mock
import pytest
from core.app.apps.base_app_generator import BaseAppGenerator
@@ -369,6 +372,58 @@ def test_validate_inputs_optional_file_with_empty_string_ignores_default():
class TestBaseAppGeneratorExtras:
def test_wrap_stream_joins_worker_after_stream_exhaustion(self):
base_app_generator = BaseAppGenerator()
worker_thread = Mock()
worker_thread.is_alive.return_value = False
def response_stream():
yield {"event": "workflow_finished"}
managed_stream = base_app_generator._wrap_stream_with_worker_thread_join(
response_stream(),
worker_thread,
)
assert next(managed_stream) == {"event": "workflow_finished"}
worker_thread.join.assert_not_called()
with pytest.raises(StopIteration):
next(managed_stream)
worker_thread.join.assert_called_once_with(timeout=300)
def test_wrap_stream_joins_worker_when_stream_closes(self):
base_app_generator = BaseAppGenerator()
worker_thread = Mock()
worker_thread.is_alive.return_value = False
def response_stream():
yield {"event": "workflow_started"}
yield {"event": "workflow_finished"}
managed_stream = base_app_generator._wrap_stream_with_worker_thread_join(
response_stream(),
worker_thread,
)
assert next(managed_stream) == {"event": "workflow_started"}
managed_stream.close()
worker_thread.join.assert_called_once_with(timeout=300)
def test_join_worker_thread_warns_when_thread_remains_alive(self, caplog: pytest.LogCaptureFixture):
worker_thread = Mock()
worker_thread.name = "leaked-app-worker"
worker_thread.is_alive.return_value = True
with caplog.at_level(logging.WARNING, logger="core.app.apps.base_app_generator"):
BaseAppGenerator._join_worker_thread(worker_thread)
worker_thread.join.assert_called_once_with(timeout=300)
assert "Possible app worker thread leak" in caplog.text
assert "leaked-app-worker" in caplog.text
def test_prepare_user_inputs_converts_files_and_lists(self, monkeypatch: pytest.MonkeyPatch):
base_app_generator = BaseAppGenerator()
@@ -111,7 +111,7 @@ def test_generate_includes_parent_trace_context_in_extras(monkeypatch):
result = generator.generate(
app_model=SimpleNamespace(tenant_id="tenant-1", id="app-1"),
workflow=SimpleNamespace(features_dict={}),
workflow=SimpleNamespace(features_dict={}, graph_dict={"nodes": []}),
user=SimpleNamespace(id="user-1", session_id="session-1"),
args={
"inputs": {"query": "hello"},
@@ -211,6 +211,13 @@ def test_generate_appends_pause_layer_and_forwards_state(mocker: MockerFixture):
def start(self):
return None
def join(self, timeout):
worker_kwargs["joined"] = True
worker_kwargs["join_timeout"] = timeout
def is_alive(self):
return False
mocker.patch("core.app.apps.workflow.app_generator.threading.Thread", DummyThread)
app_model = SimpleNamespace(mode="workflow", tenant_id="tenant")
@@ -244,6 +251,8 @@ def test_generate_appends_pause_layer_and_forwards_state(mocker: MockerFixture):
assert result == "converted"
assert worker_kwargs["kwargs"]["graph_engine_layers"] == ("base-layer", pause_layer)
assert worker_kwargs["kwargs"]["graph_runtime_state"] is graph_runtime_state
assert worker_kwargs["joined"] is True
assert worker_kwargs["join_timeout"] == 300
assert draft_saver_factory.call_args.kwargs["tenant_id"] == app_model.tenant_id
@@ -286,6 +295,8 @@ def test_resume_path_runs_worker_with_runtime_state(mocker: MockerFixture):
mocker.patch("core.app.apps.workflow.app_generator.WorkflowAppRunner", side_effect=runner_ctor)
worker_lifecycle: dict[str, bool] = {}
class ImmediateThread:
def __init__(self, target, kwargs):
target(**kwargs)
@@ -293,6 +304,13 @@ def test_resume_path_runs_worker_with_runtime_state(mocker: MockerFixture):
def start(self):
return None
def join(self, timeout):
worker_lifecycle["joined"] = True
worker_lifecycle["join_timeout"] = timeout
def is_alive(self):
return False
mocker.patch("core.app.apps.workflow.app_generator.threading.Thread", ImmediateThread)
mocker.patch(
@@ -331,5 +349,7 @@ def test_resume_path_runs_worker_with_runtime_state(mocker: MockerFixture):
)
assert result == "raw-response"
assert worker_lifecycle["joined"] is True
assert worker_lifecycle["join_timeout"] == 300
runner_instance.run.assert_called_once()
queue_manager.graph_runtime_state = runtime_state
@@ -1,5 +1,9 @@
import threading
from collections.abc import Generator
import pytest
from core.app.apps.base_app_generator import BaseAppGenerator
from core.app.apps.workflow.active_workflow_tasks import (
active_workflow_task,
get_active_workflow_task_count,
@@ -28,3 +32,51 @@ def test_active_workflow_task_rejects_duplicate_task_id() -> None:
with pytest.raises(ValueError, match="already active"):
with active_workflow_task("task-a"):
pass
def test_managed_stream_waits_for_active_worker_cleanup() -> None:
worker_started = threading.Event()
release_worker = threading.Event()
stream_exhausted = threading.Event()
consumer_finished = threading.Event()
consumer_errors: list[BaseException] = []
def run_worker() -> None:
with active_workflow_task("task-a"):
worker_started.set()
release_worker.wait()
def response_stream() -> Generator[dict[str, str], None, None]:
yield {"event": "workflow_finished"}
stream_exhausted.set()
worker_thread = threading.Thread(target=run_worker)
worker_thread.start()
assert worker_started.wait(timeout=2)
managed_stream = BaseAppGenerator._wrap_stream_with_worker_thread_join(response_stream(), worker_thread)
assert next(managed_stream) == {"event": "workflow_finished"}
def finish_stream() -> None:
try:
list(managed_stream)
except BaseException as exc:
consumer_errors.append(exc)
finally:
consumer_finished.set()
consumer_thread = threading.Thread(target=finish_stream)
consumer_thread.start()
try:
assert stream_exhausted.wait(timeout=2)
assert not consumer_finished.is_set()
assert get_active_workflow_task_count() == 1
finally:
release_worker.set()
consumer_thread.join(timeout=2)
worker_thread.join(timeout=2)
assert not consumer_thread.is_alive()
assert not worker_thread.is_alive()
assert consumer_errors == []
assert get_active_workflow_task_count() == 0
@@ -15,6 +15,70 @@ from models.model import AppMode
class TestWorkflowAppGeneratorValidation:
def test_generate_stream_joins_worker_after_response_exhaustion(self, monkeypatch: pytest.MonkeyPatch):
generator = WorkflowAppGenerator()
worker_thread = Mock()
worker_thread.is_alive.return_value = False
app_config = WorkflowUIBasedAppConfig(
tenant_id="tenant",
app_id="app",
app_mode=AppMode.WORKFLOW,
additional_features=AppAdditionalFeatures(),
variables=[],
workflow_id="workflow-id",
)
application_generate_entity = WorkflowAppGenerateEntity.model_construct(
task_id="task",
app_config=app_config,
inputs={},
files=[],
user_id="user",
stream=True,
invoke_from=InvokeFrom.WEB_APP,
extras={},
)
def response_stream():
yield {"event": "workflow_finished"}
monkeypatch.setattr(generator, "_bind_file_access_scope", lambda **kwargs: contextlib.nullcontext())
monkeypatch.setattr(
"core.app.apps.workflow.app_generator.WorkflowAppQueueManager",
lambda **kwargs: SimpleNamespace(**kwargs),
)
monkeypatch.setattr(
"core.app.apps.workflow.app_generator.current_app",
SimpleNamespace(_get_current_object=lambda: SimpleNamespace(name="flask")),
)
monkeypatch.setattr("core.app.apps.workflow.app_generator.contextvars.copy_context", lambda: "ctx")
monkeypatch.setattr("core.app.apps.workflow.app_generator.threading.Thread", lambda **kwargs: worker_thread)
monkeypatch.setattr(
"core.app.apps.workflow.app_generator.db",
SimpleNamespace(session=SimpleNamespace(close=Mock())),
)
monkeypatch.setattr(generator, "_get_draft_var_saver_factory", lambda *args, **kwargs: "draft-factory")
monkeypatch.setattr(generator, "_handle_response", lambda **kwargs: response_stream())
monkeypatch.setattr(
"core.app.apps.workflow.app_generator.WorkflowAppGenerateResponseConverter.convert",
lambda response, invoke_from: response,
)
managed_stream = generator._generate(
app_model=SimpleNamespace(mode=AppMode.WORKFLOW, tenant_id="tenant"),
workflow=SimpleNamespace(id="workflow-id"),
user=SimpleNamespace(id="user"),
application_generate_entity=application_generate_entity,
invoke_from=InvokeFrom.WEB_APP,
workflow_execution_repository=SimpleNamespace(),
workflow_node_execution_repository=SimpleNamespace(),
streaming=True,
)
worker_thread.start.assert_called_once_with()
worker_thread.join.assert_not_called()
assert list(managed_stream) == [{"event": "workflow_finished"}]
worker_thread.join.assert_called_once_with(timeout=300)
def test_ensure_snippet_start_node_returns_original_for_non_snippet_workflow(self):
workflow = SimpleNamespace(kind_or_standard="workflow")
session = SimpleNamespace(scalar=Mock())
@@ -313,7 +377,7 @@ class TestWorkflowAppGeneratorGenerate:
result = generator.generate(
app_model=SimpleNamespace(id="app", tenant_id="tenant"),
workflow=SimpleNamespace(features_dict={}),
workflow=SimpleNamespace(features_dict={}, graph_dict={"nodes": []}),
user=SimpleNamespace(id="user", session_id="session"),
args={"inputs": {}, SKIP_PREPARE_USER_INPUTS_KEY: True},
invoke_from=InvokeFrom.WEB_APP,
@@ -1,24 +1,95 @@
"""Unit tests for the message cycle manager optimization."""
import logging
from collections.abc import Iterator
from dataclasses import dataclass
from types import SimpleNamespace
from unittest.mock import Mock, patch
import pytest
from flask import Flask, current_app
from sqlalchemy import Engine, event, select
from sqlalchemy.orm import Session, sessionmaker
from core.app.entities.queue_entities import QueueAnnotationReplyEvent, QueueRetrieverResourcesEvent
from core.app.entities.task_entities import MessageStreamResponse, StreamEvent, TaskStateMetadata
from core.app.task_pipeline import message_cycle_manager as message_cycle_manager_module
from core.app.task_pipeline.message_cycle_manager import MessageCycleManager
from core.rag.entities import RetrievalSourceMetadata
from models.model import App, AppMode
from graphon.file import FileTransferMethod, FileType
from models import model as model_module
from models.base import TypeBase
from models.enums import ConversationFromSource, CreatorUserRole, MessageFileBelongsTo
from models.model import App, AppMode, Conversation, MessageFile
def _patch_create_session(mock_session):
session_cm = Mock()
session_cm.__enter__ = Mock(return_value=mock_session)
session_cm.__exit__ = Mock(return_value=False)
return patch("core.app.task_pipeline.message_cycle_manager.session_factory.create_session", return_value=session_cm)
@dataclass(frozen=True)
class _SQLiteDb:
engine: Engine
session: Session
@pytest.fixture
def cycle_db(sqlite_engine: Engine, monkeypatch: pytest.MonkeyPatch) -> Iterator[Session]:
"""Bind request-owned and cycle-manager-owned sessions to isolated SQLite."""
TypeBase.metadata.create_all(
sqlite_engine,
tables=[App.__table__, Conversation.__table__, MessageFile.__table__],
)
owned_session_factory = sessionmaker(bind=sqlite_engine, expire_on_commit=False)
with owned_session_factory() as request_session:
sqlite_db = _SQLiteDb(engine=sqlite_engine, session=request_session)
monkeypatch.setattr(message_cycle_manager_module, "db", sqlite_db)
monkeypatch.setattr(model_module, "db", sqlite_db)
monkeypatch.setattr(message_cycle_manager_module.session_factory, "create_session", owned_session_factory)
yield request_session
def _app(*, app_id: str = "app-id", tenant_id: str = "tenant-1") -> App:
return App(
id=app_id,
tenant_id=tenant_id,
name="Test App",
description="",
mode=AppMode.CHAT,
enable_site=True,
enable_api=True,
max_active_requests=0,
)
def _conversation(*, conversation_id: str = "conv-1", app_id: str = "app-id") -> Conversation:
conversation = Conversation(
app_id=app_id,
mode=AppMode.CHAT,
name="",
status="normal",
from_source=ConversationFromSource.API,
inputs={},
)
conversation.id = conversation_id
return conversation
def _message_file(
*,
file_id: str = "file-1",
message_id: str = "test-message-id",
belongs_to: MessageFileBelongsTo | None = MessageFileBelongsTo.ASSISTANT,
url: str | None = "http://example.com/image.png",
file_type: FileType = FileType.IMAGE,
) -> MessageFile:
message_file = MessageFile(
message_id=message_id,
type=file_type,
transfer_method=FileTransferMethod.TOOL_FILE,
created_by_role=CreatorUserRole.ACCOUNT,
created_by="account-id",
belongs_to=belongs_to,
url=url,
)
message_file.id = file_id
return message_file
class TestMessageCycleManagerOptimization:
@@ -37,30 +108,22 @@ class TestMessageCycleManagerOptimization:
task_state = Mock()
return MessageCycleManager(application_generate_entity=mock_application_generate_entity, task_state=task_state)
def test_get_message_event_type_with_assistant_file(self, message_cycle_manager):
def test_get_message_event_type_with_assistant_file(self, message_cycle_manager, cycle_db: Session):
"""Test get_message_event_type returns MESSAGE_FILE when message has assistant-generated files.
This ensures that AI-generated images (belongs_to='assistant') trigger the MESSAGE_FILE event,
allowing the frontend to properly display generated image files with url field.
"""
with patch("core.app.task_pipeline.message_cycle_manager.session_factory") as mock_session_factory:
# Setup mock session and message file
mock_session = Mock()
mock_session_factory.create_session.return_value.__enter__.return_value = mock_session
cycle_db.add(_message_file())
cycle_db.commit()
mock_message_file = Mock()
mock_message_file.belongs_to = "assistant"
mock_session.scalar.return_value = mock_message_file
with current_app.app_context():
result = message_cycle_manager.get_message_event_type("test-message-id")
# Execute
with current_app.app_context():
result = message_cycle_manager.get_message_event_type("test-message-id")
assert result == StreamEvent.MESSAGE_FILE
assert "test-message-id" in message_cycle_manager._message_has_file
# Assert
assert result == StreamEvent.MESSAGE_FILE
mock_session.scalar.assert_called_once()
def test_get_message_event_type_with_user_file(self, message_cycle_manager):
def test_get_message_event_type_with_user_file(self, message_cycle_manager, cycle_db: Session):
"""Test get_message_event_type returns MESSAGE when message only has user-uploaded files.
This is a regression test for the issue where user-uploaded images (belongs_to='user')
@@ -68,90 +131,81 @@ class TestMessageCycleManagerOptimization:
resulting in broken images in the chat UI. The query filters for belongs_to='assistant',
so when only user files exist, the database query returns None, resulting in MESSAGE event type.
"""
with patch("core.app.task_pipeline.message_cycle_manager.session_factory") as mock_session_factory:
# Setup mock session and message file
mock_session = Mock()
mock_session_factory.create_session.return_value.__enter__.return_value = mock_session
cycle_db.add(_message_file(belongs_to=MessageFileBelongsTo.USER))
cycle_db.commit()
# When querying for assistant files with only user files present, return None
# (simulates database query with belongs_to='assistant' filter returning no results)
mock_session.scalar.return_value = None
with current_app.app_context():
result = message_cycle_manager.get_message_event_type("test-message-id")
# Execute
with current_app.app_context():
result = message_cycle_manager.get_message_event_type("test-message-id")
assert result == StreamEvent.MESSAGE
assert "test-message-id" not in message_cycle_manager._message_has_file
# Assert
assert result == StreamEvent.MESSAGE
mock_session.scalar.assert_called_once()
def test_get_message_event_type_without_message_file(self, message_cycle_manager):
def test_get_message_event_type_without_message_file(self, message_cycle_manager, cycle_db: Session):
"""Test get_message_event_type returns MESSAGE when message has no files."""
with patch("core.app.task_pipeline.message_cycle_manager.session_factory") as mock_session_factory:
# Setup mock session and no message file
mock_session = Mock()
mock_session_factory.create_session.return_value.__enter__.return_value = mock_session
# Current implementation uses session.scalar(select(...))
mock_session.scalar.return_value = None
assert list(cycle_db.scalars(select(MessageFile)).all()) == []
# Execute
with current_app.app_context():
result = message_cycle_manager.get_message_event_type("test-message-id")
with current_app.app_context():
result = message_cycle_manager.get_message_event_type("test-message-id")
# Assert
assert result == StreamEvent.MESSAGE
mock_session.scalar.assert_called_once()
assert result == StreamEvent.MESSAGE
def test_get_message_event_type_uses_cache_without_query(self, message_cycle_manager):
def test_get_message_event_type_uses_cache_without_query(
self, message_cycle_manager, cycle_db: Session, sqlite_engine: Engine
):
"""Return MESSAGE_FILE directly from in-memory cache without opening a DB session."""
message_cycle_manager._message_has_file.add("cached-message")
statements: list[str] = []
with patch("core.app.task_pipeline.message_cycle_manager.session_factory") as mock_session_factory:
def record_statement(_conn, _cursor, statement, _parameters, _context, _executemany) -> None:
statements.append(statement)
event.listen(sqlite_engine, "before_cursor_execute", record_statement)
try:
result = message_cycle_manager.get_message_event_type("cached-message")
finally:
event.remove(sqlite_engine, "before_cursor_execute", record_statement)
assert result == StreamEvent.MESSAGE_FILE
mock_session_factory.create_session.assert_not_called()
assert statements == []
def test_message_to_stream_response_with_precomputed_event_type(self, message_cycle_manager):
def test_message_to_stream_response_with_precomputed_event_type(self, message_cycle_manager, cycle_db: Session):
"""MessageCycleManager.message_to_stream_response expects a valid event_type; callers should precompute it."""
with patch("core.app.task_pipeline.message_cycle_manager.session_factory") as mock_session_factory:
# Setup mock session and message file
mock_session = Mock()
mock_session_factory.create_session.return_value.__enter__.return_value = mock_session
cycle_db.add(_message_file())
cycle_db.commit()
mock_message_file = Mock()
mock_message_file.belongs_to = "assistant"
mock_session.scalar.return_value = mock_message_file
with current_app.app_context():
event_type = message_cycle_manager.get_message_event_type("test-message-id")
result = message_cycle_manager.message_to_stream_response(
answer="Hello world", message_id="test-message-id", event_type=event_type
)
# Execute: compute event type once, then pass to message_to_stream_response
with current_app.app_context():
event_type = message_cycle_manager.get_message_event_type("test-message-id")
result = message_cycle_manager.message_to_stream_response(
answer="Hello world", message_id="test-message-id", event_type=event_type
)
assert isinstance(result, MessageStreamResponse)
assert result.answer == "Hello world"
assert result.id == "test-message-id"
assert result.event == StreamEvent.MESSAGE_FILE
# Assert
assert isinstance(result, MessageStreamResponse)
assert result.answer == "Hello world"
assert result.id == "test-message-id"
assert result.event == StreamEvent.MESSAGE_FILE
mock_session.scalar.assert_called_once()
def test_message_to_stream_response_with_event_type_skips_query(self, message_cycle_manager):
def test_message_to_stream_response_with_event_type_skips_query(
self, message_cycle_manager, cycle_db: Session, sqlite_engine: Engine
):
"""Test that message_to_stream_response skips database query when event_type is provided."""
with patch("core.app.task_pipeline.message_cycle_manager.session_factory") as mock_session_factory:
# Execute with event_type provided
statements: list[str] = []
def record_statement(_conn, _cursor, statement, _parameters, _context, _executemany) -> None:
statements.append(statement)
event.listen(sqlite_engine, "before_cursor_execute", record_statement)
try:
result = message_cycle_manager.message_to_stream_response(
answer="Hello world", message_id="test-message-id", event_type=StreamEvent.MESSAGE
)
finally:
event.remove(sqlite_engine, "before_cursor_execute", record_statement)
# Assert
assert isinstance(result, MessageStreamResponse)
assert result.answer == "Hello world"
assert result.id == "test-message-id"
assert result.event == StreamEvent.MESSAGE
# Should not open a session when event_type is provided
mock_session_factory.create_session.assert_not_called()
assert isinstance(result, MessageStreamResponse)
assert result.answer == "Hello world"
assert result.id == "test-message-id"
assert result.event == StreamEvent.MESSAGE
assert statements == []
def test_message_to_stream_response_with_from_variable_selector(self, message_cycle_manager):
"""Test message_to_stream_response with from_variable_selector parameter."""
@@ -168,40 +222,32 @@ class TestMessageCycleManagerOptimization:
assert result.from_variable_selector == ["var1", "var2"]
assert result.event == StreamEvent.MESSAGE
def test_optimization_usage_example(self, message_cycle_manager):
def test_optimization_usage_example(self, message_cycle_manager, cycle_db: Session, sqlite_engine: Engine):
"""Test the optimization pattern that should be used by callers."""
# Step 1: Get event type once (this queries database)
with patch("core.app.task_pipeline.message_cycle_manager.session_factory") as mock_session_factory:
mock_session = Mock()
mock_session_factory.create_session.return_value.__enter__.return_value = mock_session
# Current implementation uses session.scalar(select(...))
mock_session.scalar.return_value = None # No files
statements: list[str] = []
def record_statement(_conn, _cursor, statement, _parameters, _context, _executemany) -> None:
statements.append(statement)
event.listen(sqlite_engine, "before_cursor_execute", record_statement)
try:
with current_app.app_context():
event_type = message_cycle_manager.get_message_event_type("test-message-id")
# Should open session once
mock_session_factory.create_session.assert_called_once()
assert event_type == StreamEvent.MESSAGE
# Step 2: Use event_type for multiple calls (no additional queries)
with patch("core.app.task_pipeline.message_cycle_manager.session_factory") as mock_session_factory:
mock_session_factory.create_session.return_value.__enter__.return_value = Mock()
chunk1_response = message_cycle_manager.message_to_stream_response(
answer="Chunk 1", message_id="test-message-id", event_type=event_type
)
chunk2_response = message_cycle_manager.message_to_stream_response(
answer="Chunk 2", message_id="test-message-id", event_type=event_type
)
finally:
event.remove(sqlite_engine, "before_cursor_execute", record_statement)
# Should not open session again when event_type provided
mock_session_factory.create_session.assert_not_called()
assert chunk1_response.event == StreamEvent.MESSAGE
assert chunk2_response.event == StreamEvent.MESSAGE
assert chunk1_response.answer == "Chunk 1"
assert chunk2_response.answer == "Chunk 2"
assert event_type == StreamEvent.MESSAGE
assert len([statement for statement in statements if statement.lstrip().upper().startswith("SELECT")]) == 1
assert chunk1_response.event == StreamEvent.MESSAGE
assert chunk2_response.event == StreamEvent.MESSAGE
assert chunk1_response.answer == "Chunk 1"
assert chunk2_response.answer == "Chunk 2"
def test_generate_conversation_name_returns_none_for_completion(self, message_cycle_manager):
"""Return None when completion entities are used for conversation naming.
@@ -269,51 +315,38 @@ class TestMessageCycleManagerOptimization:
assert message_cycle_manager._application_generate_entity.is_new_conversation is False
mock_timer.assert_not_called()
def test_generate_conversation_name_worker_returns_when_conversation_missing(self, message_cycle_manager):
def test_generate_conversation_name_worker_returns_when_conversation_missing(
self, message_cycle_manager, cycle_db: Session
):
"""Return early when the conversation cannot be found."""
flask_app = Flask(__name__)
db_session = Mock()
db_session.scalar.return_value = None
assert list(cycle_db.scalars(select(Conversation)).all()) == []
with _patch_create_session(db_session):
message_cycle_manager._generate_conversation_name_worker(flask_app, "conv-missing", "hello")
message_cycle_manager._generate_conversation_name_worker(flask_app, "conv-missing", "hello")
db_session.commit.assert_not_called()
assert list(cycle_db.scalars(select(Conversation)).all()) == []
def test_generate_conversation_name_worker_returns_when_app_missing(self, message_cycle_manager):
def test_generate_conversation_name_worker_returns_when_app_missing(self, message_cycle_manager, cycle_db: Session):
"""Return early when non-completion conversation has no app relation."""
flask_app = Flask(__name__)
conversation = SimpleNamespace(mode=AppMode.CHAT, app=None, app_id="app-id")
db_session = Mock()
db_session.scalar.return_value = conversation
db_session.get.return_value = None
conversation = _conversation()
cycle_db.add(conversation)
cycle_db.commit()
with _patch_create_session(db_session):
message_cycle_manager._generate_conversation_name_worker(flask_app, "conv-1", "hello")
message_cycle_manager._generate_conversation_name_worker(flask_app, "conv-1", "hello")
db_session.commit.assert_not_called()
assert cycle_db.get(Conversation, "conv-1").name == ""
assert cycle_db.get(App, "app-id") is None
def test_generate_conversation_name_worker_uses_cached_name(self, message_cycle_manager):
def test_generate_conversation_name_worker_uses_cached_name(
self, message_cycle_manager, cycle_db: Session, sqlite_engine: Engine
):
"""Use cached conversation name when present and avoid LLM call."""
flask_app = Flask(__name__)
class ConversationWithPoisonedApp:
mode = AppMode.CHAT
app_id = "app-id"
name = ""
@property
def app(self):
raise AssertionError("conversation.app must not open an implicit session")
conversation = ConversationWithPoisonedApp()
app_model = SimpleNamespace(tenant_id="tenant-1")
db_session = Mock()
db_session.scalar.return_value = conversation
db_session.get.return_value = app_model
cycle_db.add_all([_app(), _conversation()])
cycle_db.commit()
with (
_patch_create_session(db_session) as create_session,
patch("core.app.task_pipeline.message_cycle_manager.redis_client") as mock_redis,
patch("core.app.task_pipeline.message_cycle_manager.LLMGenerator") as mock_llm_generator,
):
@@ -321,27 +354,23 @@ class TestMessageCycleManagerOptimization:
message_cycle_manager._generate_conversation_name_worker(flask_app, "conv-1", "hello")
assert cycle_db.in_transaction() is False
with Session(sqlite_engine) as verification_session:
conversation = verification_session.get(Conversation, "conv-1")
assert conversation is not None
assert conversation.name == "cached-title"
create_session.assert_called_once_with()
db_session.get.assert_called_once_with(App, "app-id")
db_session.commit.assert_called_once()
mock_llm_generator.generate_conversation_name.assert_not_called()
mock_redis.setex.assert_not_called()
def test_generate_conversation_name_worker_generates_and_caches_name(self, message_cycle_manager):
def test_generate_conversation_name_worker_generates_and_caches_name(
self, message_cycle_manager, cycle_db: Session, sqlite_engine: Engine
):
"""Generate conversation name and write it to redis cache on cache miss."""
flask_app = Flask(__name__)
conversation = SimpleNamespace(
mode=AppMode.CHAT,
app=SimpleNamespace(tenant_id="tenant-1"),
app_id="app-id",
name="",
)
db_session = Mock()
db_session.scalar.return_value = conversation
cycle_db.add_all([_app(), _conversation()])
cycle_db.commit()
with (
_patch_create_session(db_session),
patch("core.app.task_pipeline.message_cycle_manager.redis_client") as mock_redis,
patch("core.app.task_pipeline.message_cycle_manager.LLMGenerator") as mock_llm_generator,
):
@@ -350,27 +379,27 @@ class TestMessageCycleManagerOptimization:
message_cycle_manager._generate_conversation_name_worker(flask_app, "conv-1", "hello")
assert cycle_db.in_transaction() is False
with Session(sqlite_engine) as verification_session:
conversation = verification_session.get(Conversation, "conv-1")
assert conversation is not None
assert conversation.name == "generated-title"
db_session.commit.assert_called_once()
mock_redis.setex.assert_called_once()
def test_generate_conversation_name_worker_falls_back_when_generation_fails(
self, message_cycle_manager, caplog: pytest.LogCaptureFixture
self,
message_cycle_manager,
cycle_db: Session,
sqlite_engine: Engine,
caplog: pytest.LogCaptureFixture,
):
"""Fallback to truncated query when LLM generation fails."""
flask_app = Flask(__name__)
conversation = SimpleNamespace(
mode=AppMode.CHAT,
app=SimpleNamespace(tenant_id="tenant-1"),
app_id="app-id",
name="",
)
db_session = Mock()
db_session.scalar.return_value = conversation
cycle_db.add_all([_app(), _conversation()])
cycle_db.commit()
long_query = "q" * 60
with (
_patch_create_session(db_session),
patch("core.app.task_pipeline.message_cycle_manager.redis_client") as mock_redis,
patch("core.app.task_pipeline.message_cycle_manager.LLMGenerator") as mock_llm_generator,
patch("core.app.task_pipeline.message_cycle_manager.dify_config") as mock_dify_config,
@@ -382,8 +411,11 @@ class TestMessageCycleManagerOptimization:
with caplog.at_level(logging.ERROR, logger="core.app.task_pipeline.message_cycle_manager"):
message_cycle_manager._generate_conversation_name_worker(flask_app, "conv-1", long_query)
assert cycle_db.in_transaction() is False
with Session(sqlite_engine) as verification_session:
conversation = verification_session.get(Conversation, "conv-1")
assert conversation is not None
assert conversation.name == (long_query[:47] + "...")
db_session.commit.assert_called_once()
assert any(record.levelno == logging.ERROR for record in caplog.records)
def test_handle_annotation_reply_sets_metadata(self, message_cycle_manager):
@@ -454,33 +486,25 @@ class TestMessageCycleManagerOptimization:
assert message_cycle_manager._task_state.metadata.retriever_resources[0].position == 1
assert message_cycle_manager._task_state.metadata.retriever_resources[1].position == 2
def test_message_file_to_stream_response_builds_signed_url(self, message_cycle_manager):
def test_message_file_to_stream_response_builds_signed_url(self, message_cycle_manager, cycle_db: Session):
"""Build a stream response with a signed tool file URL.
Args: message_cycle_manager with mocked Session/db and sign_tool_file.
Args: message_cycle_manager with a persisted MessageFile and mocked sign_tool_file.
Returns: MessageStreamResponse with signed url and belongs_to normalized to user.
Side effects: Calls sign_tool_file for tool file ids.
"""
message_cycle_manager._application_generate_entity.task_id = "task-1"
message_file = SimpleNamespace(
id="file-1",
type="image",
belongs_to=None,
url="tool://file.verylongextension",
message_id="msg-1",
cycle_db.add(
_message_file(
file_id="file-1",
message_id="msg-1",
belongs_to=None,
url="tool://file.verylongextension",
)
)
cycle_db.commit()
session = Mock()
session.scalar.return_value = message_file
with (
patch("core.app.task_pipeline.message_cycle_manager.Session") as mock_session_cls,
patch("core.app.task_pipeline.message_cycle_manager.sign_tool_file") as mock_sign,
patch("core.app.task_pipeline.message_cycle_manager.db") as mock_db,
):
mock_db.engine = Mock()
mock_session_cls.return_value.__enter__.return_value = session
with patch("core.app.task_pipeline.message_cycle_manager.sign_tool_file") as mock_sign:
mock_sign.return_value = "signed-url"
response = message_cycle_manager.message_file_to_stream_response(SimpleNamespace(message_file_id="file-1"))
@@ -514,56 +538,42 @@ class TestMessageCycleManagerOptimization:
assert len(message_cycle_manager._task_state.metadata.retriever_resources) == 1
assert message_cycle_manager._task_state.metadata.retriever_resources[0].position == 1
def test_message_file_to_stream_response_uses_http_url_directly(self, message_cycle_manager):
def test_message_file_to_stream_response_uses_http_url_directly(self, message_cycle_manager, cycle_db: Session):
"""Use original URL when message file URL is already HTTP."""
message_cycle_manager._application_generate_entity.task_id = "task-http"
message_file = SimpleNamespace(
id="file-http",
type="image",
belongs_to="assistant",
url="http://example.com/pic.png",
message_id="msg-http",
)
session = Mock()
session.scalar.return_value = message_file
with (
patch("core.app.task_pipeline.message_cycle_manager.Session") as mock_session_cls,
patch("core.app.task_pipeline.message_cycle_manager.db") as mock_db,
):
mock_db.engine = Mock()
mock_session_cls.return_value.__enter__.return_value = session
response = message_cycle_manager.message_file_to_stream_response(
SimpleNamespace(message_file_id="file-http")
cycle_db.add(
_message_file(
file_id="file-http",
message_id="msg-http",
belongs_to=MessageFileBelongsTo.ASSISTANT,
url="http://example.com/pic.png",
)
)
cycle_db.commit()
response = message_cycle_manager.message_file_to_stream_response(SimpleNamespace(message_file_id="file-http"))
assert response is not None
assert response.url == "http://example.com/pic.png"
assert "msg-http" in message_cycle_manager._message_has_file
def test_message_file_to_stream_response_defaults_extension_to_bin_without_dot(self, message_cycle_manager):
def test_message_file_to_stream_response_defaults_extension_to_bin_without_dot(
self, message_cycle_manager, cycle_db: Session
):
"""Default tool file extension to .bin when URL has no extension part."""
message_cycle_manager._application_generate_entity.task_id = "task-bin"
message_file = SimpleNamespace(
id="file-bin",
type="file",
belongs_to="assistant",
url="tool-file-id",
message_id="msg-bin",
cycle_db.add(
_message_file(
file_id="file-bin",
message_id="msg-bin",
belongs_to=MessageFileBelongsTo.ASSISTANT,
url="tool-file-id",
file_type=FileType.CUSTOM,
)
)
cycle_db.commit()
session = Mock()
session.scalar.return_value = message_file
with (
patch("core.app.task_pipeline.message_cycle_manager.Session") as mock_session_cls,
patch("core.app.task_pipeline.message_cycle_manager.sign_tool_file") as mock_sign,
patch("core.app.task_pipeline.message_cycle_manager.db") as mock_db,
):
mock_db.engine = Mock()
mock_session_cls.return_value.__enter__.return_value = session
with patch("core.app.task_pipeline.message_cycle_manager.sign_tool_file") as mock_sign:
mock_sign.return_value = "signed-bin-url"
response = message_cycle_manager.message_file_to_stream_response(
@@ -574,19 +584,13 @@ class TestMessageCycleManagerOptimization:
assert response.url == "signed-bin-url"
mock_sign.assert_called_once_with(tool_file_id="tool-file-id", extension=".bin")
def test_message_file_to_stream_response_returns_none_when_file_missing(self, message_cycle_manager):
def test_message_file_to_stream_response_returns_none_when_file_missing(
self, message_cycle_manager, cycle_db: Session
):
"""Return None when message file lookup does not find a record."""
session = Mock()
session.scalar.return_value = None
assert list(cycle_db.scalars(select(MessageFile)).all()) == []
with (
patch("core.app.task_pipeline.message_cycle_manager.Session") as mock_session_cls,
patch("core.app.task_pipeline.message_cycle_manager.db") as mock_db,
):
mock_db.engine = Mock()
mock_session_cls.return_value.__enter__.return_value = session
response = message_cycle_manager.message_file_to_stream_response(SimpleNamespace(message_file_id="missing"))
response = message_cycle_manager.message_file_to_stream_response(SimpleNamespace(message_file_id="missing"))
assert response is None
@@ -1,7 +1,57 @@
import sys
from datetime import datetime, timedelta
from unittest.mock import MagicMock, patch
import pytest
from sqlalchemy import event
from sqlalchemy.exc import SQLAlchemyError
from sqlalchemy.orm import Session, sessionmaker
import core.llm_generator.llm_generator as generator_module
from core.llm_generator.llm_generator import LLMGenerator, _parse_string_list
from core.model_manager import ModelInstance, ModelManager
from core.workflow.generator import tool_catalogue as tool_catalogue_module
from core.workflow.generator.tool_catalogue import ToolCatalogueEntry
from graphon.model_runtime.entities.llm_entities import LLMResult, LLMUsage
from graphon.model_runtime.entities.message_entities import AssistantPromptMessage
from models.dataset import Dataset
from services.workflow_service import WorkflowService
@pytest.fixture
def dataset_session(sqlite_session: Session, monkeypatch: pytest.MonkeyPatch) -> Session:
"""Bind the real SQLite session to the production database extension."""
monkeypatch.setattr(generator_module.db, "session", sqlite_session)
return sqlite_session
def _llm_result(content: str) -> LLMResult:
"""Build a real non-streaming LLM response around deterministic test content."""
return LLMResult(
model="test-model",
message=AssistantPromptMessage(content=content),
usage=LLMUsage.empty_usage(),
)
def _model_manager() -> tuple[MagicMock, MagicMock]:
"""Build spec-constrained mocks for the model-manager boundary and its default model."""
model_manager = MagicMock(spec=ModelManager)
model_instance = MagicMock(spec=ModelInstance)
model_manager.get_default_model_instance.return_value = model_instance
return model_manager, model_instance
def _dataset(*, dataset_id: str, tenant_id: str, name: str, created_at: datetime) -> Dataset:
return Dataset(
id=dataset_id,
tenant_id=tenant_id,
name=name,
created_by="account-id",
created_at=created_at,
)
class TestParseStringList:
@@ -34,95 +84,115 @@ class TestParseStringList:
class TestGenerateWorkflowInstructionSuggestions:
@patch("core.llm_generator.llm_generator.ModelManager.for_tenant")
def test_no_default_model(self, mock_for_tenant):
mock_for_tenant.return_value.get_default_model_instance.side_effect = Exception("No model")
model_manager, _ = _model_manager()
model_manager.get_default_model_instance.side_effect = RuntimeError("no default model")
mock_for_tenant.return_value = model_manager
assert LLMGenerator.generate_workflow_instruction_suggestions("tenant", mode="workflow") == []
@patch("core.llm_generator.llm_generator.ModelManager.for_tenant")
@patch("core.llm_generator.llm_generator.LLMGenerator._build_suggestion_context")
def test_llm_success(self, mock_build_context, mock_for_tenant):
mock_build_context.return_value = "context"
mock_model = MagicMock()
mock_model.invoke_llm.return_value = MagicMock()
mock_model.invoke_llm.return_value.message.get_text_content.return_value = '["idea 1", "idea 2"]'
mock_for_tenant.return_value.get_default_model_instance.return_value = mock_model
model_manager, model_instance = _model_manager()
model_instance.invoke_llm.return_value = _llm_result('["idea 1", "idea 2"]')
mock_for_tenant.return_value = model_manager
result = LLMGenerator.generate_workflow_instruction_suggestions("tenant", mode="workflow")
assert result == ["idea 1", "idea 2"]
model_instance.invoke_llm.assert_called_once()
@patch("core.llm_generator.llm_generator.ModelManager.for_tenant")
@patch("core.llm_generator.llm_generator.LLMGenerator._build_suggestion_context")
def test_llm_error(self, mock_build_context, mock_for_tenant):
mock_build_context.return_value = "context"
model_manager, model_instance = _model_manager()
model_instance.invoke_llm.side_effect = RuntimeError("API error")
mock_for_tenant.return_value = model_manager
mock_model = MagicMock()
mock_model.invoke_llm.side_effect = Exception("API error")
mock_for_tenant.return_value.get_default_model_instance.return_value = mock_model
assert LLMGenerator.generate_workflow_instruction_suggestions("tenant", mode="workflow") == []
result = LLMGenerator.generate_workflow_instruction_suggestions("tenant", mode="workflow")
assert result == []
model_instance.invoke_llm.assert_called_once()
@patch("core.llm_generator.llm_generator.ModelManager.for_tenant")
@patch("core.llm_generator.llm_generator.LLMGenerator._build_suggestion_context")
def test_llm_bad_output(self, mock_build_context, mock_for_tenant):
mock_build_context.return_value = "context"
model_manager, model_instance = _model_manager()
model_instance.invoke_llm.return_value = _llm_result("Not a list")
mock_for_tenant.return_value = model_manager
mock_model = MagicMock()
mock_model.invoke_llm.return_value = MagicMock()
mock_model.invoke_llm.return_value.message.get_text_content.return_value = "Not a list"
mock_for_tenant.return_value.get_default_model_instance.return_value = mock_model
assert LLMGenerator.generate_workflow_instruction_suggestions("tenant", mode="workflow") == []
result = LLMGenerator.generate_workflow_instruction_suggestions("tenant", mode="workflow")
assert result == []
model_instance.invoke_llm.assert_called_once()
@pytest.mark.parametrize("sqlite_session", [(Dataset,)], indirect=True)
class TestBuildSuggestionContext:
@patch("core.llm_generator.llm_generator.db.session.scalars")
def test_both_success(self, mock_scalars, monkeypatch):
mock_scalars.return_value.all.return_value = ["kb1", "kb2"]
def test_both_success(self, dataset_session: Session, monkeypatch: pytest.MonkeyPatch):
now = datetime.now()
dataset_session.add_all(
(
_dataset(dataset_id="kb-1", tenant_id="tenant", name="kb1", created_at=now),
_dataset(
dataset_id="kb-2",
tenant_id="tenant",
name="kb2",
created_at=now - timedelta(seconds=1),
),
_dataset(dataset_id="other-kb", tenant_id="other", name="private", created_at=now),
)
)
dataset_session.commit()
# ``_build_suggestion_context`` imports the tool catalogue lazily, so we
# stub the module in ``sys.modules``. Use ``monkeypatch.setitem`` so the
# ORIGINAL module is RESTORED on teardown — a bare ``del`` would evict it
# from sys.modules entirely, after which a sibling test that imported
# ``build_tool_catalogue`` at collection time (e.g. test_tool_catalogue)
# diverges from a freshly re-imported module and its @patch targets stop
# applying, silently breaking it under xdist.
mock_tool_catalogue = MagicMock()
mock_tool_catalogue.build_tool_catalogue.return_value = "catalog"
mock_tool_catalogue.format_tool_catalogue.return_value = "tool1\ntool2"
monkeypatch.setitem(sys.modules, "core.workflow.generator.tool_catalogue", mock_tool_catalogue)
def build_tool_catalogue(_tenant_id: str) -> list[ToolCatalogueEntry]:
return [
ToolCatalogueEntry(
provider_name="provider",
provider_type="builtin",
plugin_id="",
tool_name="tool1",
tool_label="tool1",
description="First tool",
),
ToolCatalogueEntry(
provider_name="provider",
provider_type="builtin",
plugin_id="",
tool_name="tool2",
tool_label="tool2",
description="Second tool",
),
]
# Keep the real module and formatter; only isolate provider/plugin discovery.
monkeypatch.setattr(tool_catalogue_module, "build_tool_catalogue", build_tool_catalogue)
result = LLMGenerator._build_suggestion_context("tenant")
assert "Knowledge bases:\n- kb1\n- kb2" in result
assert "Installed tools:\ntool1\ntool2" in result
assert "Installed tools:\n- provider/tool1 — First tool\n- provider/tool2 — Second tool" in result
@patch("core.llm_generator.llm_generator.db.session.scalars")
def test_both_fail(self, mock_scalars, monkeypatch):
mock_scalars.side_effect = Exception("DB error")
def test_both_fail(self, dataset_session: Session, monkeypatch: pytest.MonkeyPatch):
def fail_query(_orm_execute_state: object) -> None:
raise SQLAlchemyError("DB error")
# See ``test_both_success``: restore the original module via monkeypatch
# rather than ``del``-ing it, so we don't evict it for sibling tests.
mock_tool_catalogue = MagicMock()
mock_tool_catalogue.build_tool_catalogue.side_effect = Exception("Tool error")
monkeypatch.setitem(sys.modules, "core.workflow.generator.tool_catalogue", mock_tool_catalogue)
def fail_tool_catalogue(_tenant_id: str) -> list[ToolCatalogueEntry]:
raise RuntimeError("Tool error")
assert LLMGenerator._build_suggestion_context("tenant") == ""
event.listen(dataset_session, "do_orm_execute", fail_query)
monkeypatch.setattr(tool_catalogue_module, "build_tool_catalogue", fail_tool_catalogue)
try:
assert LLMGenerator._build_suggestion_context("tenant") == ""
finally:
event.remove(dataset_session, "do_orm_execute", fail_query)
class TestWorkflowServiceInterface:
def test_protocol_methods(self):
# Just to cover the 'pass' statements in the Protocol definition
def test_real_workflow_service_exposes_protocol_methods(self):
from core.llm_generator.llm_generator import WorkflowServiceInterface
class MockService(WorkflowServiceInterface):
def get_draft_workflow(self, app_model, workflow_id=None, *, session):
return super().get_draft_workflow(app_model, workflow_id, session=session)
service: WorkflowServiceInterface = WorkflowService(sessionmaker())
def get_node_last_run(self, app_model, workflow, node_id):
return super().get_node_last_run(app_model, workflow, node_id)
service = MockService()
service.get_draft_workflow(None, session=None)
service.get_node_last_run(None, None, "node")
assert callable(service.get_draft_workflow)
assert callable(service.get_node_last_run)
@@ -1,11 +1,19 @@
"""Comprehensive unit tests for core/memory/token_buffer_memory.py"""
"""Comprehensive SQLite-backed tests for token-buffer memory."""
from collections.abc import Iterator
from dataclasses import dataclass
from datetime import UTC, datetime, timedelta
from decimal import Decimal
from unittest.mock import MagicMock, patch
from uuid import uuid4
import pytest
from sqlalchemy import Engine, event
from sqlalchemy.orm import Session
from core.memory import token_buffer_memory as memory_module
from core.memory.token_buffer_memory import TokenBufferMemory
from graphon.file import FileTransferMethod, FileType
from graphon.model_runtime.entities import (
AssistantPromptMessage,
ImagePromptMessageContent,
@@ -13,13 +21,44 @@ from graphon.model_runtime.entities import (
TextPromptMessageContent,
UserPromptMessage,
)
from models.model import AppMode
from models.base import TypeBase
from models.enums import ConversationFromSource, CreatorUserRole, MessageFileBelongsTo
from models.model import AppMode, Message, MessageFile
from models.workflow import Workflow, WorkflowType
# ---------------------------------------------------------------------------
# Helpers / shared fixtures
# ---------------------------------------------------------------------------
@dataclass(frozen=True)
class Database:
"""Typed SQLite binding plus executed SQL for query-count assertions."""
engine: Engine
session: Session
statements: list[tuple[str, object]]
@pytest.fixture
def database(sqlite_engine: Engine, monkeypatch: pytest.MonkeyPatch) -> Iterator[Database]:
TypeBase.metadata.create_all(
sqlite_engine,
tables=[Message.__table__, MessageFile.__table__, Workflow.__table__],
)
statements: list[tuple[str, object]] = []
def record_statement(_connection, _cursor, statement, parameters, _context, _executemany) -> None:
statements.append((statement, parameters))
event.listen(sqlite_engine, "before_cursor_execute", record_statement)
with Session(sqlite_engine, expire_on_commit=False) as session:
database = Database(engine=sqlite_engine, session=session, statements=statements)
monkeypatch.setattr(memory_module, "db", database)
yield database
event.remove(sqlite_engine, "before_cursor_execute", record_statement)
def _make_conversation(mode: AppMode = AppMode.CHAT) -> MagicMock:
"""Return a minimal Conversation mock."""
conv = MagicMock()
@@ -47,6 +86,73 @@ def _make_message(answer: str = "hello", answer_tokens: int = 5) -> MagicMock:
return msg
def _persist_message(
database: Database,
conversation_id: str,
*,
query: str = "user query",
answer: str = "hello",
answer_tokens: int = 5,
created_at: datetime | None = None,
workflow_run_id: str | None = None,
) -> Message:
message = Message(
id=str(uuid4()),
app_id="app-1",
conversation_id=conversation_id,
_inputs={},
query=query,
message={},
message_unit_price=Decimal(0),
answer=answer,
answer_tokens=answer_tokens,
answer_unit_price=Decimal(0),
currency="USD",
from_source=ConversationFromSource.API,
workflow_run_id=workflow_run_id,
created_at=created_at or datetime.now(UTC).replace(tzinfo=None),
)
database.session.add(message)
database.session.commit()
return message
def _persist_message_file(
database: Database,
message: Message,
*,
belongs_to: MessageFileBelongsTo | None,
) -> MessageFile:
message_file = MessageFile(
message_id=message.id,
type=FileType.IMAGE,
transfer_method=FileTransferMethod.REMOTE_URL,
created_by_role=CreatorUserRole.ACCOUNT,
created_by="account-1",
belongs_to=belongs_to,
url="https://example.com/image.png",
)
database.session.add(message_file)
database.session.commit()
return message_file
def _persist_workflow(database: Database, *, workflow_id: str) -> Workflow:
workflow = Workflow(
id=workflow_id,
tenant_id="tenant-1",
app_id="app-1",
type=WorkflowType.CHAT,
version="1",
graph="{}",
features="{}",
created_by="account-1",
)
database.session.add(workflow)
database.session.commit()
return workflow
# ===========================================================================
# Tests for __init__ and workflow_run_repo property
# ===========================================================================
@@ -61,25 +167,25 @@ class TestInit:
assert mem.model_instance is mi
assert mem._workflow_run_repo is None
def test_workflow_run_repo_is_created_lazily(self):
def test_workflow_run_repo_is_created_lazily(self, database: Database):
conv = _make_conversation()
mi = _make_model_instance()
mem = TokenBufferMemory(conversation=conv, model_instance=mi)
mock_repo = MagicMock()
with (
patch("core.memory.token_buffer_memory.sessionmaker") as mock_sm,
patch("core.memory.token_buffer_memory.db") as mock_db,
patch(
"core.memory.token_buffer_memory.DifyAPIRepositoryFactory.create_api_workflow_run_repository",
return_value=mock_repo,
),
):
mock_db.engine = MagicMock()
with patch(
"core.memory.token_buffer_memory.DifyAPIRepositoryFactory.create_api_workflow_run_repository",
return_value=mock_repo,
) as repository_factory:
repo = mem.workflow_run_repo
assert repo is mock_repo
assert mem._workflow_run_repo is mock_repo
session_factory = repository_factory.call_args.args[0]
with session_factory() as session:
assert isinstance(session, Session)
assert session.get_bind() is database.engine
def test_workflow_run_repo_cached_after_first_access(self):
conv = _make_conversation()
mi = _make_model_instance()
@@ -410,7 +516,7 @@ class TestBuildPromptMessageWithFiles:
)
@pytest.mark.parametrize("mode", [AppMode.ADVANCED_CHAT, AppMode.WORKFLOW])
def test_workflow_mode_workflow_not_found_raises(self, mode):
def test_workflow_mode_workflow_not_found_raises(self, mode, database: Database):
"""Raises ValueError when Workflow lookup returns None."""
conv = _make_conversation(mode)
conv.app = MagicMock()
@@ -422,22 +528,17 @@ class TestBuildPromptMessageWithFiles:
mem._workflow_run_repo = MagicMock()
mem._workflow_run_repo.get_workflow_run_by_id.return_value = mock_workflow_run
with (
patch("core.memory.token_buffer_memory.db") as mock_db,
):
mock_db.session.scalar.return_value = None # workflow not found
with pytest.raises(ValueError, match="Workflow not found"):
mem._build_prompt_message_with_files(
message_files=[],
text_content="text",
message=_make_message(),
app_record=MagicMock(),
is_user_message=True,
)
with pytest.raises(ValueError, match="Workflow not found"):
mem._build_prompt_message_with_files(
message_files=[],
text_content="text",
message=_make_message(),
app_record=MagicMock(),
is_user_message=True,
)
@pytest.mark.parametrize("mode", [AppMode.ADVANCED_CHAT, AppMode.WORKFLOW])
def test_workflow_mode_success_no_files_user(self, mode):
def test_workflow_mode_success_no_files_user(self, mode, database: Database):
"""Happy path: workflow mode, no message files → plain UserPromptMessage."""
conv = _make_conversation(mode)
conv.app = MagicMock()
@@ -445,22 +546,16 @@ class TestBuildPromptMessageWithFiles:
mock_workflow_run = MagicMock()
mock_workflow_run.workflow_id = str(uuid4())
mock_workflow = MagicMock()
mock_workflow.features_dict = {}
workflow = _persist_workflow(database, workflow_id=mock_workflow_run.workflow_id)
mem = TokenBufferMemory(conversation=conv, model_instance=_make_model_instance())
mem._workflow_run_repo = MagicMock()
mem._workflow_run_repo.get_workflow_run_by_id.return_value = mock_workflow_run
with (
patch("core.memory.token_buffer_memory.db") as mock_db,
patch(
"core.memory.token_buffer_memory.FileUploadConfigManager.convert",
return_value=None,
),
with patch(
"core.memory.token_buffer_memory.FileUploadConfigManager.convert",
return_value=None,
):
mock_db.session.scalar.return_value = mock_workflow
result = mem._build_prompt_message_with_files(
message_files=[],
text_content="wf text",
@@ -471,6 +566,7 @@ class TestBuildPromptMessageWithFiles:
assert isinstance(result, UserPromptMessage)
assert result.content == "wf text"
assert database.session.get(Workflow, workflow.id) is workflow
# ------------------------------------------------------------------
# Invalid mode
@@ -498,417 +594,140 @@ class TestBuildPromptMessageWithFiles:
class TestGetHistoryPromptMessages:
"""Tests for get_history_prompt_messages."""
"""Tests for persisted history retrieval, file batching, and pruning."""
def _make_memory(self, mode: AppMode = AppMode.CHAT) -> TokenBufferMemory:
conv = _make_conversation(mode)
conv.app = MagicMock()
return TokenBufferMemory(conversation=conv, model_instance=_make_model_instance())
def test_returns_empty_when_no_messages(self):
def test_returns_empty_when_no_messages(self, database: Database) -> None:
assert self._make_memory().get_history_prompt_messages() == []
def test_skips_newest_message_without_answer(self, database: Database) -> None:
mem = self._make_memory()
with patch("core.memory.token_buffer_memory.db") as mock_db:
mock_db.session.scalars.return_value.all.return_value = []
result = mem.get_history_prompt_messages()
assert result == []
message = _persist_message(database, mem.conversation.id, answer="", answer_tokens=0)
def test_skips_first_message_without_answer(self):
"""The newest message (index 0 after extraction) without answer and tokens==0 is skipped."""
assert mem.get_history_prompt_messages() == []
assert database.session.get(Message, message.id) is message
def test_message_with_answer_returns_user_and_assistant_prompts(self, database: Database) -> None:
mem = self._make_memory()
_persist_message(database, mem.conversation.id, query="My query", answer="My answer", answer_tokens=10)
msg_no_answer = _make_message(answer="", answer_tokens=0)
msg_no_answer.parent_message_id = None # ensures extract_thread_messages returns it
with (
patch("core.memory.token_buffer_memory.db") as mock_db,
patch(
"core.memory.token_buffer_memory.extract_thread_messages",
return_value=[msg_no_answer],
),
):
mock_db.session.scalars.return_value.all.side_effect = [
[msg_no_answer], # first call: messages query
[], # second call: user files query (never hit, but safe)
]
result = mem.get_history_prompt_messages()
assert result == []
def test_message_with_answer_not_skipped(self):
"""A message with a non-empty answer is NOT popped."""
mem = self._make_memory()
msg = _make_message(answer="some answer", answer_tokens=10)
msg.parent_message_id = None
with (
patch("core.memory.token_buffer_memory.db") as mock_db,
patch(
"core.memory.token_buffer_memory.extract_thread_messages",
return_value=[msg],
),
patch(
"core.memory.token_buffer_memory.FileUploadConfigManager.convert",
return_value=None,
),
):
# user files query → empty; assistant files query → empty
mock_db.session.scalars.return_value.all.return_value = []
result = mem.get_history_prompt_messages()
assert len(result) == 2 # one user + one assistant
def test_message_limit_default_is_500(self):
"""When message_limit is None the stmt is limited to 500."""
mem = self._make_memory()
with (
patch("core.memory.token_buffer_memory.db") as mock_db,
patch("core.memory.token_buffer_memory.select") as mock_select,
patch("core.memory.token_buffer_memory.extract_thread_messages", return_value=[]),
):
mock_stmt = MagicMock()
mock_select.return_value.where.return_value.order_by.return_value = mock_stmt
mock_stmt.limit.return_value = mock_stmt
mock_db.session.scalars.return_value.all.return_value = []
mem.get_history_prompt_messages(message_limit=None)
mock_stmt.limit.assert_called_with(500)
def test_message_limit_clipped_to_500(self):
"""A message_limit > 500 is clamped to 500."""
mem = self._make_memory()
with (
patch("core.memory.token_buffer_memory.db") as mock_db,
patch("core.memory.token_buffer_memory.select") as mock_select,
patch("core.memory.token_buffer_memory.extract_thread_messages", return_value=[]),
):
mock_stmt = MagicMock()
mock_select.return_value.where.return_value.order_by.return_value = mock_stmt
mock_stmt.limit.return_value = mock_stmt
mock_db.session.scalars.return_value.all.return_value = []
mem.get_history_prompt_messages(message_limit=9999)
mock_stmt.limit.assert_called_with(500)
def test_message_limit_positive_used(self):
"""A positive message_limit < 500 is used as-is."""
mem = self._make_memory()
with (
patch("core.memory.token_buffer_memory.db") as mock_db,
patch("core.memory.token_buffer_memory.select") as mock_select,
patch("core.memory.token_buffer_memory.extract_thread_messages", return_value=[]),
):
mock_stmt = MagicMock()
mock_select.return_value.where.return_value.order_by.return_value = mock_stmt
mock_stmt.limit.return_value = mock_stmt
mock_db.session.scalars.return_value.all.return_value = []
mem.get_history_prompt_messages(message_limit=10)
mock_stmt.limit.assert_called_with(10)
def test_message_limit_zero_uses_default(self):
"""message_limit=0 triggers the else branch → default 500."""
mem = self._make_memory()
with (
patch("core.memory.token_buffer_memory.db") as mock_db,
patch("core.memory.token_buffer_memory.select") as mock_select,
patch("core.memory.token_buffer_memory.extract_thread_messages", return_value=[]),
):
mock_stmt = MagicMock()
mock_select.return_value.where.return_value.order_by.return_value = mock_stmt
mock_stmt.limit.return_value = mock_stmt
mock_db.session.scalars.return_value.all.return_value = []
mem.get_history_prompt_messages(message_limit=0)
mock_stmt.limit.assert_called_with(500)
def test_user_files_cause_build_with_files_call(self):
"""When user_files is non-empty _build_prompt_message_with_files is invoked."""
mem = self._make_memory()
msg = _make_message()
msg.parent_message_id = None
mock_user_file = MagicMock()
mock_user_file.message_id = msg.id # must match so batched grouping keys it to this message
mock_user_prompt = UserPromptMessage(content="from build")
mock_assistant_prompt = AssistantPromptMessage(content="answer")
call_count = {"n": 0}
def scalars_side_effect(stmt):
r = MagicMock()
if call_count["n"] == 0:
# messages query
r.all.return_value = [msg]
elif call_count["n"] == 1:
# user files
r.all.return_value = [mock_user_file]
else:
# assistant files
r.all.return_value = []
call_count["n"] += 1
return r
with (
patch("core.memory.token_buffer_memory.db") as mock_db,
patch(
"core.memory.token_buffer_memory.extract_thread_messages",
return_value=[msg],
),
patch.object(
mem,
"_build_prompt_message_with_files",
side_effect=[mock_user_prompt, mock_assistant_prompt],
) as mock_build,
patch(
"core.memory.token_buffer_memory.FileUploadConfigManager.convert",
return_value=None,
),
):
mock_db.session.scalars.side_effect = scalars_side_effect
result = mem.get_history_prompt_messages()
assert mock_build.call_count >= 1
# First call should be user message
first_call_kwargs = mock_build.call_args_list[0][1]
assert first_call_kwargs["is_user_message"] is True
def test_assistant_files_cause_build_with_files_call(self):
"""When assistant_files is non-empty, build is called with is_user_message=False."""
mem = self._make_memory()
msg = _make_message()
msg.parent_message_id = None
mock_assistant_file = MagicMock()
mock_assistant_file.message_id = msg.id # must match so batched grouping keys it to this message
mock_user_prompt = UserPromptMessage(content="query")
mock_assistant_prompt = AssistantPromptMessage(content="built")
call_count = {"n": 0}
def scalars_side_effect(stmt):
r = MagicMock()
if call_count["n"] == 0:
r.all.return_value = [msg]
elif call_count["n"] == 1:
r.all.return_value = [] # no user files
else:
r.all.return_value = [mock_assistant_file]
call_count["n"] += 1
return r
with (
patch("core.memory.token_buffer_memory.db") as mock_db,
patch(
"core.memory.token_buffer_memory.extract_thread_messages",
return_value=[msg],
),
patch.object(
mem,
"_build_prompt_message_with_files",
return_value=mock_assistant_prompt,
) as mock_build,
):
mock_db.session.scalars.side_effect = scalars_side_effect
result = mem.get_history_prompt_messages()
mock_build.assert_called_once()
call_kwargs = mock_build.call_args[1]
assert call_kwargs["is_user_message"] is False
def test_message_files_loaded_with_constant_query_count(self):
"""Regression guard against N+1: message files must be batch-loaded.
Regardless of the number of messages in the thread, file loading must use a
constant number of queries (1 messages query + 2 batched file queries),
never 2 queries per message.
"""
mem = self._make_memory()
messages = [_make_message() for _ in range(5)]
for m in messages:
m.parent_message_id = None
scalars_calls = {"n": 0}
def scalars_side_effect(stmt):
r = MagicMock()
# First call returns the thread messages; the batched file queries return none.
r.all.return_value = messages if scalars_calls["n"] == 0 else []
scalars_calls["n"] += 1
return r
with (
patch("core.memory.token_buffer_memory.db") as mock_db,
patch("core.memory.token_buffer_memory.extract_thread_messages", return_value=messages),
patch("core.memory.token_buffer_memory.FileUploadConfigManager.convert", return_value=None),
):
mock_db.session.scalars.side_effect = scalars_side_effect
mem.get_history_prompt_messages()
# 1 (messages) + 2 (batched user/assistant files) = 3, independent of message count.
# Before this fix it would have been 1 + 2 * 5 = 11 (an N+1 pattern).
assert scalars_calls["n"] == 3
def test_token_pruning_removes_oldest_messages(self):
"""If tokens exceed limit, oldest messages are removed until within limit."""
conv = _make_conversation()
conv.app = MagicMock()
# Model returns tokens that decrease only after removing pairs
token_values = [3000, 1500] # first call over limit, second within
mi = MagicMock()
mi.get_llm_num_tokens.side_effect = token_values
mem = TokenBufferMemory(conversation=conv, model_instance=mi)
msg = _make_message()
msg.parent_message_id = None
call_count = {"n": 0}
def scalars_side_effect(stmt):
r = MagicMock()
if call_count["n"] == 0:
r.all.return_value = [msg]
else:
r.all.return_value = []
call_count["n"] += 1
return r
with (
patch("core.memory.token_buffer_memory.db") as mock_db,
patch(
"core.memory.token_buffer_memory.extract_thread_messages",
return_value=[msg],
),
patch(
"core.memory.token_buffer_memory.FileUploadConfigManager.convert",
return_value=None,
),
):
mock_db.session.scalars.side_effect = scalars_side_effect
result = mem.get_history_prompt_messages(max_token_limit=2000)
# After pruning, we should have fewer than the 2 initial messages
assert len(result) <= 1
def test_token_pruning_stops_at_single_message(self):
"""Pruning stops when only 1 message remains (to prevent empty list)."""
conv = _make_conversation()
conv.app = MagicMock()
# Always over limit
mi = MagicMock()
mi.get_llm_num_tokens.return_value = 99999
mem = TokenBufferMemory(conversation=conv, model_instance=mi)
msg = _make_message()
msg.parent_message_id = None
call_count = {"n": 0}
def scalars_side_effect(stmt):
r = MagicMock()
if call_count["n"] == 0:
r.all.return_value = [msg]
else:
r.all.return_value = []
call_count["n"] += 1
return r
with (
patch("core.memory.token_buffer_memory.db") as mock_db,
patch(
"core.memory.token_buffer_memory.extract_thread_messages",
return_value=[msg],
),
patch(
"core.memory.token_buffer_memory.FileUploadConfigManager.convert",
return_value=None,
),
):
mock_db.session.scalars.side_effect = scalars_side_effect
result = mem.get_history_prompt_messages(max_token_limit=1)
# At least 1 message should remain
assert len(result) >= 1
def test_no_pruning_when_within_limit(self):
"""When tokens ≤ limit, no pruning occurs."""
mem = self._make_memory()
mem.model_instance.get_llm_num_tokens.return_value = 50 # well under default 2000
msg = _make_message()
msg.parent_message_id = None
call_count = {"n": 0}
def scalars_side_effect(stmt):
r = MagicMock()
if call_count["n"] == 0:
r.all.return_value = [msg]
else:
r.all.return_value = []
call_count["n"] += 1
return r
with (
patch("core.memory.token_buffer_memory.db") as mock_db,
patch(
"core.memory.token_buffer_memory.extract_thread_messages",
return_value=[msg],
),
patch(
"core.memory.token_buffer_memory.FileUploadConfigManager.convert",
return_value=None,
),
):
mock_db.session.scalars.side_effect = scalars_side_effect
result = mem.get_history_prompt_messages(max_token_limit=2000)
assert len(result) == 2 # user + assistant
def test_plain_user_and_assistant_messages_returned(self):
"""Without files, plain UserPromptMessage and AssistantPromptMessage appear."""
mem = self._make_memory()
msg = _make_message(answer="My answer")
msg.query = "My query"
msg.parent_message_id = None
call_count = {"n": 0}
def scalars_side_effect(stmt):
r = MagicMock()
if call_count["n"] == 0:
r.all.return_value = [msg]
else:
r.all.return_value = []
call_count["n"] += 1
return r
with (
patch("core.memory.token_buffer_memory.db") as mock_db,
patch(
"core.memory.token_buffer_memory.extract_thread_messages",
return_value=[msg],
),
patch(
"core.memory.token_buffer_memory.FileUploadConfigManager.convert",
return_value=None,
),
):
mock_db.session.scalars.side_effect = scalars_side_effect
result = mem.get_history_prompt_messages()
result = mem.get_history_prompt_messages()
assert len(result) == 2
user_msg, ai_msg = result
assert isinstance(user_msg, UserPromptMessage)
assert user_msg.content == "My query"
assert isinstance(ai_msg, AssistantPromptMessage)
assert ai_msg.content == "My answer"
assert isinstance(result[0], UserPromptMessage)
assert result[0].content == "My query"
assert isinstance(result[1], AssistantPromptMessage)
assert result[1].content == "My answer"
def test_history_is_conversation_scoped(self, database: Database) -> None:
mem = self._make_memory()
_persist_message(database, mem.conversation.id, answer="visible")
_persist_message(database, "other-conversation", answer="hidden")
result = mem.get_history_prompt_messages()
assert [prompt.content for prompt in result] == ["user query", "visible"]
@pytest.mark.parametrize(
("message_limit", "expected_limit"),
[(None, 500), (9999, 500), (10, 10), (0, 500)],
)
def test_message_limit_is_applied_to_executable_query(
self,
database: Database,
message_limit: int | None,
expected_limit: int,
) -> None:
mem = self._make_memory()
before = len(database.statements)
mem.get_history_prompt_messages(message_limit=message_limit)
statements = database.statements[before:]
assert len(statements) == 1
sql, parameters = statements[0]
assert "LIMIT" in sql
assert expected_limit in parameters
@pytest.mark.parametrize(
("belongs_to", "is_user_message"),
[
(MessageFileBelongsTo.USER, True),
(None, True),
(MessageFileBelongsTo.ASSISTANT, False),
],
)
def test_message_files_use_persisted_ownership(
self,
database: Database,
belongs_to: MessageFileBelongsTo | None,
is_user_message: bool,
) -> None:
mem = self._make_memory()
message = _persist_message(database, mem.conversation.id)
message_file = _persist_message_file(database, message, belongs_to=belongs_to)
built_prompt = (
UserPromptMessage(content="built user")
if is_user_message
else AssistantPromptMessage(content="built assistant")
)
with patch.object(mem, "_build_prompt_message_with_files", return_value=built_prompt) as build_prompt:
result = mem.get_history_prompt_messages()
build_prompt.assert_called_once()
assert build_prompt.call_args.kwargs["message_files"] == [message_file]
assert build_prompt.call_args.kwargs["is_user_message"] is is_user_message
assert built_prompt in result
def test_message_files_are_batch_loaded_with_constant_query_count(self, database: Database) -> None:
mem = self._make_memory()
base_time = datetime.now(UTC).replace(tzinfo=None)
messages = [
_persist_message(
database,
mem.conversation.id,
query=f"query-{index}",
answer=f"answer-{index}",
created_at=base_time + timedelta(seconds=index),
)
for index in range(5)
]
before = len(database.statements)
with patch("core.memory.token_buffer_memory.extract_thread_messages", return_value=messages):
result = mem.get_history_prompt_messages()
selects = [sql for sql, _ in database.statements[before:] if sql.lstrip().upper().startswith("SELECT")]
assert len(selects) == 3
assert len(result) == 10
@pytest.mark.parametrize(
("token_values", "max_token_limit", "expected_length"),
[
([3000, 1500], 2000, 1),
([99999, 99999], 1, 1),
([50], 2000, 2),
],
)
def test_token_pruning_uses_persisted_history(
self,
database: Database,
token_values: list[int],
max_token_limit: int,
expected_length: int,
) -> None:
mem = self._make_memory()
mem.model_instance.get_llm_num_tokens.side_effect = token_values
_persist_message(database, mem.conversation.id)
result = mem.get_history_prompt_messages(max_token_limit=max_token_limit)
assert len(result) == expected_length
# ===========================================================================
@@ -2,13 +2,24 @@ from unittest.mock import MagicMock, patch
import pytest
from pydantic import ValidationError
from sqlalchemy.orm import Session
import core.moderation.api.api as moderation_module
from core.extension.api_based_extension_requestor import APIBasedExtensionPoint
from core.moderation.api.api import ApiModeration, ModerationInputParams, ModerationOutputParams
from core.moderation.base import ModerationAction, ModerationInputsResult, ModerationOutputsResult
from models.api_based_extension import APIBasedExtension
class _DatabaseBinding:
"""Expose the real SQLite session used by extension lookup."""
session: Session
def __init__(self, session: Session) -> None:
self.session = session
class TestApiModeration:
@pytest.fixture
def api_config(self):
@@ -165,17 +176,27 @@ class TestApiModeration:
with pytest.raises(ValueError, match="API-based Extension not found"):
api_moderation._get_config_by_requestor(APIBasedExtensionPoint.APP_MODERATION_INPUT, {})
@patch("core.moderation.api.api.db.session.scalar")
def test_get_api_based_extension(self, mock_scalar):
mock_ext = MagicMock(spec=APIBasedExtension)
mock_scalar.return_value = mock_ext
@pytest.mark.parametrize("sqlite_session", [(APIBasedExtension,)], indirect=True)
def test_get_api_based_extension(self, sqlite_session: Session, monkeypatch: pytest.MonkeyPatch) -> None:
target = APIBasedExtension(
tenant_id="tenant-1",
name="Target extension",
api_endpoint="https://example.com/moderate",
api_key="encrypted-key",
)
target.id = "ext-1"
other_tenant = APIBasedExtension(
tenant_id="tenant-2",
name="Other extension",
api_endpoint="https://example.com/other",
api_key="other-key",
)
other_tenant.id = "ext-2"
sqlite_session.add_all((target, other_tenant))
sqlite_session.commit()
monkeypatch.setattr(moderation_module, "db", _DatabaseBinding(sqlite_session))
result = ApiModeration._get_api_based_extension("tenant-1", "ext-1")
assert result == mock_ext
mock_scalar.assert_called_once()
# Verify the call has the correct filters
args, kwargs = mock_scalar.call_args
stmt = args[0]
# We can't easily inspect the statement without complex sqlalchemy tricks,
# but calling it is usually enough for unit tests if we mock the result.
assert result is target
assert ApiModeration._get_api_based_extension("tenant-1", "ext-2") is None
@@ -7,36 +7,212 @@ Covers:
- TraceTask._get_user_id_from_metadata
"""
from unittest.mock import MagicMock, patch
import uuid
from collections.abc import Iterator
from contextlib import contextmanager
from unittest.mock import PropertyMock, patch
import pytest
from sqlalchemy import Engine, event
from sqlalchemy.orm import Session
from core.tools.entities.tool_entities import ApiProviderSchemaType
from extensions.ext_database import db
from graphon.model_runtime.entities.model_entities import ModelType
from models.account import Tenant
from models.base import TypeBase
from models.model import App, AppMode, IconType
from models.provider import Provider, ProviderCredential, ProviderModel, ProviderModelCredential, ProviderType
from models.tools import ApiToolProvider, BuiltinToolProvider, MCPToolProvider, WorkflowToolProvider
# ---------------------------------------------------------------------------
# Helpers
# ---------------------------------------------------------------------------
def _make_db_and_session_patches(scalar_side_effect=None, scalar_return_value=None):
"""Return (mock_db, cm, session) ready to patch 'core.ops.ops_trace_manager.db'
and 'core.ops.ops_trace_manager.Session'.
@pytest.fixture
def orm_session(sqlite_engine: Engine) -> Iterator[Session]:
models = (
App,
Tenant,
Provider,
ProviderCredential,
ProviderModel,
ProviderModelCredential,
BuiltinToolProvider,
ApiToolProvider,
WorkflowToolProvider,
MCPToolProvider,
)
tables = [model.metadata.tables[model.__tablename__] for model in models]
TypeBase.metadata.create_all(sqlite_engine, tables=tables)
Provide either scalar_side_effect (list, for multiple calls) or
scalar_return_value (single value).
"""
mock_db = MagicMock()
mock_db.engine = MagicMock()
with patch.object(type(db), "engine", new_callable=PropertyMock, return_value=sqlite_engine):
with Session(sqlite_engine, expire_on_commit=False) as session:
yield session
session = MagicMock()
if scalar_side_effect is not None:
session.scalar.side_effect = scalar_side_effect
def _persist_app(session: Session, *, tenant_id: str, name: str = "MyApp") -> App:
app = App(
id=str(uuid.uuid4()),
tenant_id=tenant_id,
name=name,
mode=AppMode.WORKFLOW,
icon_type=IconType.EMOJI,
icon="workflow",
icon_background="#FFFFFF",
enable_site=True,
enable_api=False,
)
session.add(app)
session.commit()
return app
def _persist_tenant(session: Session, *, name: str = "MyWorkspace") -> Tenant:
tenant = Tenant(name=name)
session.add(tenant)
session.commit()
return tenant
def _persist_tool_provider(
session: Session, provider_type: str
) -> BuiltinToolProvider | ApiToolProvider | WorkflowToolProvider | MCPToolProvider:
tenant_id = str(uuid.uuid4())
user_id = str(uuid.uuid4())
if provider_type in {"builtin", "plugin"}:
provider = BuiltinToolProvider(
name="CredentialA",
tenant_id=tenant_id,
user_id=user_id,
provider="test/provider",
)
elif provider_type == "api":
provider = ApiToolProvider(
name="CredentialA",
icon="icon.svg",
schema="{}",
schema_type_str=ApiProviderSchemaType.OPENAPI,
user_id=user_id,
tenant_id=tenant_id,
description="API provider",
tools_str="[]",
credentials_str="{}",
)
elif provider_type == "workflow":
provider = WorkflowToolProvider(
name="CredentialA",
label="CredentialA",
icon="icon.svg",
app_id=str(uuid.uuid4()),
version="1",
user_id=user_id,
tenant_id=tenant_id,
description="Workflow provider",
)
elif provider_type == "mcp":
provider = MCPToolProvider(
name="CredentialA",
server_identifier="credential-a",
server_url="https://example.com/mcp",
server_url_hash="credential-a-hash",
icon="icon.svg",
tenant_id=tenant_id,
user_id=user_id,
)
else:
session.scalar.return_value = scalar_return_value
raise ValueError(f"unsupported provider type: {provider_type}")
cm = MagicMock()
cm.__enter__ = MagicMock(return_value=session)
cm.__exit__ = MagicMock(return_value=False)
session.add(provider)
session.commit()
return provider
return mock_db, cm, session
def _persist_provider_credential(
session: Session,
*,
tenant_id: str,
credential_name: str = "ProvCredName",
) -> ProviderCredential:
credential = ProviderCredential(
tenant_id=tenant_id,
provider_name="openai",
credential_name=credential_name,
encrypted_config="{}",
)
session.add(credential)
session.commit()
return credential
def _persist_model_credential(
session: Session,
*,
tenant_id: str,
credential_name: str = "ModelCredName",
) -> ProviderModelCredential:
credential = ProviderModelCredential(
tenant_id=tenant_id,
provider_name="openai",
model_name="gpt-4",
model_type=ModelType.LLM,
credential_name=credential_name,
encrypted_config="{}",
)
session.add(credential)
session.commit()
return credential
def _persist_provider(
session: Session,
*,
tenant_id: str,
credential_id: str | None,
) -> Provider:
provider = Provider(
tenant_id=tenant_id,
provider_name="openai",
provider_type=ProviderType.CUSTOM,
credential_id=credential_id,
)
session.add(provider)
session.commit()
return provider
def _persist_provider_model(
session: Session,
*,
tenant_id: str,
credential_id: str | None,
) -> ProviderModel:
model = ProviderModel(
tenant_id=tenant_id,
provider_name="openai",
model_name="gpt-4",
model_type=ModelType.LLM,
credential_id=credential_id,
)
session.add(model)
session.commit()
return model
@contextmanager
def _raise_on_table(engine: Engine, table_name: str) -> Iterator[None]:
"""Raise only when SQL targets the named table, leaving other real lookups intact."""
def fail_target_query(_conn, _cursor, statement, _parameters, _context, _executemany):
if f"FROM {table_name}" in statement:
raise RuntimeError(f"forced failure for {table_name}")
event.listen(engine, "before_cursor_execute", fail_target_query)
try:
yield
finally:
event.remove(engine, "before_cursor_execute", fail_target_query)
# ---------------------------------------------------------------------------
@@ -47,62 +223,42 @@ def _make_db_and_session_patches(scalar_side_effect=None, scalar_return_value=No
class TestLookupAppAndWorkspaceNames:
"""Tests for _lookup_app_and_workspace_names(app_id, tenant_id)."""
def test_both_found(self):
def test_both_found(self, orm_session: Session):
"""Returns (app_name, workspace_name) when both records exist."""
from core.ops.ops_trace_manager import _lookup_app_and_workspace_names
mock_db, cm, _session = _make_db_and_session_patches(scalar_side_effect=["MyApp", "MyWorkspace"])
with (
patch("core.ops.ops_trace_manager.db", mock_db),
patch("core.ops.ops_trace_manager.Session", return_value=cm),
):
app_name, workspace_name = _lookup_app_and_workspace_names("app-123", "tenant-456")
tenant = _persist_tenant(orm_session)
app = _persist_app(orm_session, tenant_id=tenant.id)
app_name, workspace_name = _lookup_app_and_workspace_names(app.id, tenant.id)
assert app_name == "MyApp"
assert workspace_name == "MyWorkspace"
def test_app_only_found(self):
def test_app_only_found(self, orm_session: Session):
"""Returns (app_name, '') when tenant record is absent."""
from core.ops.ops_trace_manager import _lookup_app_and_workspace_names
mock_db, cm, _session = _make_db_and_session_patches(scalar_side_effect=["MyApp", None])
with (
patch("core.ops.ops_trace_manager.db", mock_db),
patch("core.ops.ops_trace_manager.Session", return_value=cm),
):
app_name, workspace_name = _lookup_app_and_workspace_names("app-123", "tenant-456")
app = _persist_app(orm_session, tenant_id=str(uuid.uuid4()))
app_name, workspace_name = _lookup_app_and_workspace_names(app.id, str(uuid.uuid4()))
assert app_name == "MyApp"
assert workspace_name == ""
def test_tenant_only_found(self):
def test_tenant_only_found(self, orm_session: Session):
"""Returns ('', workspace_name) when app record is absent."""
from core.ops.ops_trace_manager import _lookup_app_and_workspace_names
mock_db, cm, _session = _make_db_and_session_patches(scalar_side_effect=[None, "MyWorkspace"])
with (
patch("core.ops.ops_trace_manager.db", mock_db),
patch("core.ops.ops_trace_manager.Session", return_value=cm),
):
app_name, workspace_name = _lookup_app_and_workspace_names("app-123", "tenant-456")
tenant = _persist_tenant(orm_session)
app_name, workspace_name = _lookup_app_and_workspace_names(str(uuid.uuid4()), tenant.id)
assert app_name == ""
assert workspace_name == "MyWorkspace"
def test_neither_found(self):
def test_neither_found(self, orm_session: Session):
"""Returns ('', '') when both DB lookups return None."""
from core.ops.ops_trace_manager import _lookup_app_and_workspace_names
mock_db, cm, _session = _make_db_and_session_patches(scalar_side_effect=[None, None])
with (
patch("core.ops.ops_trace_manager.db", mock_db),
patch("core.ops.ops_trace_manager.Session", return_value=cm),
):
app_name, workspace_name = _lookup_app_and_workspace_names("app-123", "tenant-456")
app_name, workspace_name = _lookup_app_and_workspace_names(str(uuid.uuid4()), str(uuid.uuid4()))
assert app_name == ""
assert workspace_name == ""
@@ -111,50 +267,30 @@ class TestLookupAppAndWorkspaceNames:
"""Returns ('', '') immediately when both IDs are None — no DB access."""
from core.ops.ops_trace_manager import _lookup_app_and_workspace_names
mock_db = MagicMock()
mock_session_cls = MagicMock()
app_name, workspace_name = _lookup_app_and_workspace_names(None, None)
with (
patch("core.ops.ops_trace_manager.db", mock_db),
patch("core.ops.ops_trace_manager.Session", mock_session_cls),
):
app_name, workspace_name = _lookup_app_and_workspace_names(None, None)
mock_session_cls.assert_not_called()
assert app_name == ""
assert workspace_name == ""
def test_app_id_none_only_queries_tenant(self):
def test_app_id_none_only_queries_tenant(self, orm_session: Session):
"""When app_id is None, only the tenant query is issued."""
from core.ops.ops_trace_manager import _lookup_app_and_workspace_names
mock_db, cm, session = _make_db_and_session_patches(scalar_return_value="OnlyWorkspace")
with (
patch("core.ops.ops_trace_manager.db", mock_db),
patch("core.ops.ops_trace_manager.Session", return_value=cm),
):
app_name, workspace_name = _lookup_app_and_workspace_names(None, "tenant-456")
tenant = _persist_tenant(orm_session, name="OnlyWorkspace")
app_name, workspace_name = _lookup_app_and_workspace_names(None, tenant.id)
assert app_name == ""
assert workspace_name == "OnlyWorkspace"
assert session.scalar.call_count == 1
def test_tenant_id_none_only_queries_app(self):
def test_tenant_id_none_only_queries_app(self, orm_session: Session):
"""When tenant_id is None, only the app query is issued."""
from core.ops.ops_trace_manager import _lookup_app_and_workspace_names
mock_db, cm, session = _make_db_and_session_patches(scalar_return_value="OnlyApp")
with (
patch("core.ops.ops_trace_manager.db", mock_db),
patch("core.ops.ops_trace_manager.Session", return_value=cm),
):
app_name, workspace_name = _lookup_app_and_workspace_names("app-123", None)
app = _persist_app(orm_session, tenant_id=str(uuid.uuid4()), name="OnlyApp")
app_name, workspace_name = _lookup_app_and_workspace_names(app.id, None)
assert app_name == "OnlyApp"
assert workspace_name == ""
assert session.scalar.call_count == 1
# ---------------------------------------------------------------------------
@@ -166,32 +302,20 @@ class TestLookupCredentialName:
"""Tests for _lookup_credential_name(credential_id, provider_type)."""
@pytest.mark.parametrize("provider_type", ["builtin", "plugin", "api", "workflow", "mcp"])
def test_known_provider_types_return_name(self, provider_type):
def test_known_provider_types_return_name(self, provider_type: str, orm_session: Session):
"""Each valid provider_type results in a DB query and returns the credential name."""
from core.ops.ops_trace_manager import _lookup_credential_name
mock_db, cm, session = _make_db_and_session_patches(scalar_return_value="CredentialA")
with (
patch("core.ops.ops_trace_manager.db", mock_db),
patch("core.ops.ops_trace_manager.Session", return_value=cm),
):
result = _lookup_credential_name("cred-123", provider_type)
provider = _persist_tool_provider(orm_session, provider_type)
result = _lookup_credential_name(provider.id, provider_type)
assert result == "CredentialA"
session.scalar.assert_called_once()
def test_credential_not_found_returns_empty_string(self):
def test_credential_not_found_returns_empty_string(self, orm_session: Session):
"""Returns '' when DB yields None for the given credential_id."""
from core.ops.ops_trace_manager import _lookup_credential_name
mock_db, cm, _session = _make_db_and_session_patches(scalar_return_value=None)
with (
patch("core.ops.ops_trace_manager.db", mock_db),
patch("core.ops.ops_trace_manager.Session", return_value=cm),
):
result = _lookup_credential_name("cred-999", "api")
result = _lookup_credential_name(str(uuid.uuid4()), "api")
assert result == ""
@@ -199,48 +323,24 @@ class TestLookupCredentialName:
"""Returns '' immediately for an unrecognised provider_type — no DB access."""
from core.ops.ops_trace_manager import _lookup_credential_name
mock_db = MagicMock()
mock_session_cls = MagicMock()
result = _lookup_credential_name(str(uuid.uuid4()), "unknown_type")
with (
patch("core.ops.ops_trace_manager.db", mock_db),
patch("core.ops.ops_trace_manager.Session", mock_session_cls),
):
result = _lookup_credential_name("cred-123", "unknown_type")
mock_session_cls.assert_not_called()
assert result == ""
def test_none_credential_id_returns_empty_string_without_db(self):
"""Returns '' immediately when credential_id is None — no DB access."""
from core.ops.ops_trace_manager import _lookup_credential_name
mock_db = MagicMock()
mock_session_cls = MagicMock()
result = _lookup_credential_name(None, "api")
with (
patch("core.ops.ops_trace_manager.db", mock_db),
patch("core.ops.ops_trace_manager.Session", mock_session_cls),
):
result = _lookup_credential_name(None, "api")
mock_session_cls.assert_not_called()
assert result == ""
def test_none_provider_type_returns_empty_string_without_db(self):
"""Returns '' immediately when provider_type is None — no DB access."""
from core.ops.ops_trace_manager import _lookup_credential_name
mock_db = MagicMock()
mock_session_cls = MagicMock()
result = _lookup_credential_name(str(uuid.uuid4()), None)
with (
patch("core.ops.ops_trace_manager.db", mock_db),
patch("core.ops.ops_trace_manager.Session", mock_session_cls),
):
result = _lookup_credential_name("cred-123", None)
mock_session_cls.assert_not_called()
assert result == ""
def test_builtin_and_plugin_map_to_same_model(self):
@@ -281,106 +381,78 @@ class TestLookupCredentialName:
class TestLookupLlmCredentialInfo:
"""Tests for _lookup_llm_credential_info(tenant_id, provider, model, model_type)."""
def _provider_record(self, credential_id: str | None = None) -> MagicMock:
record = MagicMock()
record.credential_id = credential_id
return record
def _model_record(self, credential_id: str | None = None) -> MagicMock:
record = MagicMock()
record.credential_id = credential_id
return record
def test_model_level_credential_found(self):
def test_model_level_credential_found(self, orm_session: Session):
"""Returns model-level credential_id and name when ProviderModel has a credential."""
from core.ops.ops_trace_manager import _lookup_llm_credential_info
provider_record = self._provider_record(credential_id=None)
model_record = self._model_record(credential_id="model-cred-id")
tenant_id = str(uuid.uuid4())
model_credential = _persist_model_credential(orm_session, tenant_id=tenant_id)
_persist_provider(orm_session, tenant_id=tenant_id, credential_id=None)
_persist_provider_model(orm_session, tenant_id=tenant_id, credential_id=model_credential.id)
# scalar calls: (1) Provider, (2) ProviderModel, (3) ProviderModelCredential.credential_name
mock_db, cm, _session = _make_db_and_session_patches(
scalar_side_effect=[provider_record, model_record, "ModelCredName"]
decoy_tenant_id = str(uuid.uuid4())
decoy_credential = _persist_model_credential(
orm_session,
tenant_id=decoy_tenant_id,
credential_name="WrongTenantCredential",
)
_persist_provider(orm_session, tenant_id=decoy_tenant_id, credential_id=None)
_persist_provider_model(orm_session, tenant_id=decoy_tenant_id, credential_id=decoy_credential.id)
with (
patch("core.ops.ops_trace_manager.db", mock_db),
patch("core.ops.ops_trace_manager.Session", return_value=cm),
):
cred_id, cred_name = _lookup_llm_credential_info("tenant-1", "openai", "gpt-4")
cred_id, cred_name = _lookup_llm_credential_info(tenant_id, "openai", "gpt-4")
assert cred_id == "model-cred-id"
assert cred_id == model_credential.id
assert cred_name == "ModelCredName"
def test_provider_level_fallback_when_no_model_credential(self):
def test_provider_level_fallback_when_no_model_credential(self, orm_session: Session):
"""Falls back to provider-level credential when ProviderModel has no credential_id."""
from core.ops.ops_trace_manager import _lookup_llm_credential_info
provider_record = self._provider_record(credential_id="prov-cred-id")
model_record = self._model_record(credential_id=None)
tenant_id = str(uuid.uuid4())
provider_credential = _persist_provider_credential(orm_session, tenant_id=tenant_id)
_persist_provider(orm_session, tenant_id=tenant_id, credential_id=provider_credential.id)
_persist_provider_model(orm_session, tenant_id=tenant_id, credential_id=None)
# scalar calls: (1) Provider, (2) ProviderModel (no cred), (3) ProviderCredential.credential_name
mock_db, cm, _session = _make_db_and_session_patches(
scalar_side_effect=[provider_record, model_record, "ProvCredName"]
)
cred_id, cred_name = _lookup_llm_credential_info(tenant_id, "openai", "gpt-4")
with (
patch("core.ops.ops_trace_manager.db", mock_db),
patch("core.ops.ops_trace_manager.Session", return_value=cm),
):
cred_id, cred_name = _lookup_llm_credential_info("tenant-1", "openai", "gpt-4")
assert cred_id == "prov-cred-id"
assert cred_id == provider_credential.id
assert cred_name == "ProvCredName"
def test_provider_level_fallback_when_no_model_record(self):
def test_provider_level_fallback_when_no_model_record(self, orm_session: Session):
"""Falls back to provider-level credential when no ProviderModel row exists."""
from core.ops.ops_trace_manager import _lookup_llm_credential_info
provider_record = self._provider_record(credential_id="prov-cred-id")
tenant_id = str(uuid.uuid4())
provider_credential = _persist_provider_credential(orm_session, tenant_id=tenant_id)
_persist_provider(orm_session, tenant_id=tenant_id, credential_id=provider_credential.id)
# scalar calls: (1) Provider, (2) ProviderModel → None, (3) ProviderCredential.credential_name
mock_db, cm, _session = _make_db_and_session_patches(scalar_side_effect=[provider_record, None, "ProvCredName"])
cred_id, cred_name = _lookup_llm_credential_info(tenant_id, "openai", "gpt-4")
with (
patch("core.ops.ops_trace_manager.db", mock_db),
patch("core.ops.ops_trace_manager.Session", return_value=cm),
):
cred_id, cred_name = _lookup_llm_credential_info("tenant-1", "openai", "gpt-4")
assert cred_id == "prov-cred-id"
assert cred_id == provider_credential.id
assert cred_name == "ProvCredName"
def test_no_model_arg_uses_provider_level_only(self):
def test_no_model_arg_uses_provider_level_only(self, orm_session: Session):
"""When model is None, skips ProviderModel query and uses provider credential."""
from core.ops.ops_trace_manager import _lookup_llm_credential_info
provider_record = self._provider_record(credential_id="prov-cred-id")
tenant_id = str(uuid.uuid4())
provider_credential = _persist_provider_credential(orm_session, tenant_id=tenant_id)
_persist_provider(orm_session, tenant_id=tenant_id, credential_id=provider_credential.id)
# scalar calls: (1) Provider, (2) ProviderCredential.credential_name — no ProviderModel
mock_db, cm, session = _make_db_and_session_patches(scalar_side_effect=[provider_record, "ProvCredName"])
cred_id, cred_name = _lookup_llm_credential_info(tenant_id, "openai", None)
with (
patch("core.ops.ops_trace_manager.db", mock_db),
patch("core.ops.ops_trace_manager.Session", return_value=cm),
):
cred_id, cred_name = _lookup_llm_credential_info("tenant-1", "openai", None)
assert cred_id == "prov-cred-id"
assert cred_id == provider_credential.id
assert cred_name == "ProvCredName"
assert session.scalar.call_count == 2
def test_provider_not_found_returns_none_and_empty(self):
def test_provider_not_found_returns_none_and_empty(self, orm_session: Session):
"""Returns (None, '') when Provider record does not exist."""
from core.ops.ops_trace_manager import _lookup_llm_credential_info
mock_db, cm, _session = _make_db_and_session_patches(scalar_return_value=None)
other_tenant_id = str(uuid.uuid4())
_persist_provider(orm_session, tenant_id=other_tenant_id, credential_id=None)
tenant_id = str(uuid.uuid4())
with (
patch("core.ops.ops_trace_manager.db", mock_db),
patch("core.ops.ops_trace_manager.Session", return_value=cm),
):
cred_id, cred_name = _lookup_llm_credential_info("tenant-1", "openai", "gpt-4")
cred_id, cred_name = _lookup_llm_credential_info(tenant_id, "openai", "gpt-4")
assert cred_id is None
assert cred_name == ""
@@ -389,16 +461,8 @@ class TestLookupLlmCredentialInfo:
"""Returns (None, '') immediately when tenant_id is None — no DB access."""
from core.ops.ops_trace_manager import _lookup_llm_credential_info
mock_db = MagicMock()
mock_session_cls = MagicMock()
cred_id, cred_name = _lookup_llm_credential_info(None, "openai", "gpt-4")
with (
patch("core.ops.ops_trace_manager.db", mock_db),
patch("core.ops.ops_trace_manager.Session", mock_session_cls),
):
cred_id, cred_name = _lookup_llm_credential_info(None, "openai", "gpt-4")
mock_session_cls.assert_not_called()
assert cred_id is None
assert cred_name == ""
@@ -406,69 +470,46 @@ class TestLookupLlmCredentialInfo:
"""Returns (None, '') immediately when provider is None — no DB access."""
from core.ops.ops_trace_manager import _lookup_llm_credential_info
mock_db = MagicMock()
mock_session_cls = MagicMock()
cred_id, cred_name = _lookup_llm_credential_info(str(uuid.uuid4()), None, "gpt-4")
with (
patch("core.ops.ops_trace_manager.db", mock_db),
patch("core.ops.ops_trace_manager.Session", mock_session_cls),
):
cred_id, cred_name = _lookup_llm_credential_info("tenant-1", None, "gpt-4")
mock_session_cls.assert_not_called()
assert cred_id is None
assert cred_name == ""
def test_db_error_on_outer_query_returns_none_and_empty(self):
def test_db_error_on_outer_query_returns_none_and_empty(self, orm_session: Session, sqlite_engine: Engine):
"""Returns (None, '') and logs a warning when the outer DB query raises."""
from core.ops.ops_trace_manager import _lookup_llm_credential_info
mock_db, cm, session = _make_db_and_session_patches()
session.scalar.side_effect = Exception("DB connection failed")
with (
patch("core.ops.ops_trace_manager.db", mock_db),
patch("core.ops.ops_trace_manager.Session", return_value=cm),
):
cred_id, cred_name = _lookup_llm_credential_info("tenant-1", "openai", "gpt-4")
with _raise_on_table(sqlite_engine, "providers"):
cred_id, cred_name = _lookup_llm_credential_info(str(uuid.uuid4()), "openai", "gpt-4")
assert cred_id is None
assert cred_name == ""
def test_credential_name_lookup_failure_returns_id_with_empty_name(self):
def test_credential_name_lookup_failure_returns_id_with_empty_name(
self, orm_session: Session, sqlite_engine: Engine
):
"""When credential name sub-query fails, returns cred_id but '' for name."""
from core.ops.ops_trace_manager import _lookup_llm_credential_info
provider_record = self._provider_record(credential_id="prov-cred-id")
tenant_id = str(uuid.uuid4())
provider_credential = _persist_provider_credential(orm_session, tenant_id=tenant_id)
_persist_provider(orm_session, tenant_id=tenant_id, credential_id=provider_credential.id)
# Provider found, no model record, then name lookup raises
mock_db, cm, _session = _make_db_and_session_patches(
scalar_side_effect=[provider_record, None, Exception("deleted")]
)
with _raise_on_table(sqlite_engine, "provider_credentials"):
cred_id, cred_name = _lookup_llm_credential_info(tenant_id, "openai", "gpt-4")
with (
patch("core.ops.ops_trace_manager.db", mock_db),
patch("core.ops.ops_trace_manager.Session", return_value=cm),
):
cred_id, cred_name = _lookup_llm_credential_info("tenant-1", "openai", "gpt-4")
assert cred_id == "prov-cred-id"
assert cred_id == provider_credential.id
assert cred_name == ""
def test_no_credential_on_provider_or_model_returns_none_id(self):
def test_no_credential_on_provider_or_model_returns_none_id(self, orm_session: Session):
"""Returns (None, '') when neither provider nor model has a credential_id."""
from core.ops.ops_trace_manager import _lookup_llm_credential_info
provider_record = self._provider_record(credential_id=None)
model_record = self._model_record(credential_id=None)
tenant_id = str(uuid.uuid4())
_persist_provider(orm_session, tenant_id=tenant_id, credential_id=None)
_persist_provider_model(orm_session, tenant_id=tenant_id, credential_id=None)
mock_db, cm, _session = _make_db_and_session_patches(scalar_side_effect=[provider_record, model_record])
with (
patch("core.ops.ops_trace_manager.db", mock_db),
patch("core.ops.ops_trace_manager.Session", return_value=cm),
):
cred_id, cred_name = _lookup_llm_credential_info("tenant-1", "openai", "gpt-4")
cred_id, cred_name = _lookup_llm_credential_info(tenant_id, "openai", "gpt-4")
assert cred_id is None
assert cred_name == ""
+47 -11
View File
@@ -1,9 +1,11 @@
import re
from datetime import datetime
from unittest.mock import MagicMock, patch
from decimal import Decimal
import pytest
from sqlalchemy.orm import Session
import core.ops.utils as utils_module
from core.ops.utils import (
filter_none_values,
generate_dotted_order,
@@ -15,6 +17,42 @@ from core.ops.utils import (
validate_url,
validate_url_with_path,
)
from models.enums import ConversationFromSource
from models.model import Message
class _DatabaseBinding:
"""Expose the real SQLite session used by the message lookup helper."""
session: Session
def __init__(self, session: Session) -> None:
self.session = session
@pytest.fixture
def message_session(sqlite_session: Session, monkeypatch: pytest.MonkeyPatch) -> Session:
"""Bind the message lookup helper to the shared SQLite test session."""
monkeypatch.setattr(utils_module, "db", _DatabaseBinding(sqlite_session))
return sqlite_session
def _message(message_id: str) -> Message:
message = Message(
id=message_id,
app_id="app-id",
conversation_id="conversation-id",
query="question",
message={"role": "user", "content": "question"},
answer="answer",
message_unit_price=Decimal("0.0001"),
answer_unit_price=Decimal("0.0001"),
currency="USD",
from_source=ConversationFromSource.API,
)
message._inputs = {}
return message
class TestValidateUrl:
@@ -220,22 +258,20 @@ class TestFilterNoneValues:
assert filter_none_values({}) == {}
@pytest.mark.parametrize("sqlite_session", [(Message,)], indirect=True)
class TestGetMessageData:
"""Test cases for get_message_data function"""
@patch("core.ops.utils.db")
@patch("core.ops.utils.Message")
@patch("core.ops.utils.select")
def test_get_message_data(self, mock_select, mock_message, mock_db):
mock_scalar = mock_db.session.scalar
mock_msg_instance = MagicMock()
mock_scalar.return_value = mock_msg_instance
def test_get_message_data(self, message_session: Session):
target = _message("message-id")
unrelated = _message("other-message-id")
message_session.add_all((target, unrelated))
message_session.commit()
result = get_message_data("message-id")
assert result == mock_msg_instance
mock_select.assert_called_once()
mock_scalar.assert_called_once()
assert result is target
assert result.id == "message-id"
class TestMeasureTime:
@@ -1,9 +1,42 @@
from unittest.mock import MagicMock
from datetime import datetime, timedelta
from decimal import Decimal
from uuid import uuid4
import pytest
from sqlalchemy.orm import Session
from constants import UUID_NIL
from core.prompt.utils.extract_thread_messages import extract_thread_messages
from core.prompt.utils.get_thread_messages_length import get_thread_messages_length
from models.enums import ConversationFromSource
from models.model import Message
def _persisted_message(
*,
message_id: str,
conversation_id: str,
parent_message_id: str,
answer: str,
created_at: datetime,
) -> Message:
message = Message(
id=message_id,
app_id="app-id",
conversation_id=conversation_id,
query="question",
message={"role": "user", "content": "question"},
answer=answer,
message_unit_price=Decimal("0.0001"),
answer_unit_price=Decimal("0.0001"),
currency="USD",
from_source=ConversationFromSource.API,
parent_message_id=parent_message_id,
created_at=created_at,
updated_at=created_at,
)
message._inputs = {}
return message
class MockMessage:
@@ -104,33 +137,64 @@ def test_extract_thread_messages_breaks_when_parent_is_none():
assert result[0].id == id2
def test_get_thread_messages_length_excludes_newly_created_empty_answer():
@pytest.mark.parametrize("sqlite_session", [(Message,)], indirect=True)
def test_get_thread_messages_length_excludes_newly_created_empty_answer(sqlite_session: Session):
id1, id2 = str(uuid4()), str(uuid4())
now = datetime.now()
messages = [
MockMessage(id2, id1, answer=""), # newest generated message should be excluded
MockMessage(id1, UUID_NIL, answer="ok"),
_persisted_message(
message_id=id2,
conversation_id="conversation-1",
parent_message_id=id1,
answer="",
created_at=now,
),
_persisted_message(
message_id=id1,
conversation_id="conversation-1",
parent_message_id=UUID_NIL,
answer="ok",
created_at=now - timedelta(seconds=1),
),
_persisted_message(
message_id=str(uuid4()),
conversation_id="other-conversation",
parent_message_id=UUID_NIL,
answer="unrelated",
created_at=now + timedelta(seconds=1),
),
]
sqlite_session.add_all(messages)
sqlite_session.commit()
session = MagicMock()
session.scalars.return_value.all.return_value = messages
length = get_thread_messages_length("conversation-1", session=session)
length = get_thread_messages_length("conversation-1", session=sqlite_session)
assert length == 1
session.scalars.assert_called_once()
def test_get_thread_messages_length_keeps_non_empty_latest_answer():
@pytest.mark.parametrize("sqlite_session", [(Message,)], indirect=True)
def test_get_thread_messages_length_keeps_non_empty_latest_answer(sqlite_session: Session):
id1, id2 = str(uuid4()), str(uuid4())
now = datetime.now()
messages = [
MockMessage(id2, id1, answer="latest-answer"),
MockMessage(id1, UUID_NIL, answer="older-answer"),
_persisted_message(
message_id=id2,
conversation_id="conversation-2",
parent_message_id=id1,
answer="latest-answer",
created_at=now,
),
_persisted_message(
message_id=id1,
conversation_id="conversation-2",
parent_message_id=UUID_NIL,
answer="older-answer",
created_at=now - timedelta(seconds=1),
),
]
sqlite_session.add_all(messages)
sqlite_session.commit()
session = MagicMock()
session.scalars.return_value.all.return_value = messages
length = get_thread_messages_length("conversation-2", session=session)
length = get_thread_messages_length("conversation-2", session=sqlite_session)
assert length == 2
session.scalars.assert_called_once()
File diff suppressed because it is too large Load Diff
@@ -1,16 +1,16 @@
"""Unit tests for workflow-as-tool behavior.
StubSession/StubScalars emulate SQLAlchemy session/scalars with minimal methods
(`scalar`, `scalars`, `expunge`, `commit`, `refresh`, context manager) to keep
database access mocked and predictable in tests.
"""
"""Unit tests for workflow-as-tool behavior with real SQLite ORM boundaries."""
import json
import uuid
from collections.abc import Iterator
from dataclasses import dataclass
from types import SimpleNamespace
from typing import Any
from unittest.mock import MagicMock, Mock, patch
import pytest
from sqlalchemy import Engine, inspect
from sqlalchemy.orm import Session, sessionmaker
from core.app.entities.app_invoke_entities import InvokeFrom
from core.tools.__base.tool_runtime import ToolRuntime
@@ -23,74 +23,142 @@ from core.tools.entities.tool_entities import (
ToolProviderType,
)
from core.tools.errors import ToolInvokeError
from core.tools.workflow_as_tool import tool as workflow_tool_module
from core.tools.workflow_as_tool.tool import WorkflowTool
from graphon.file import FILE_MODEL_IDENTITY, FileTransferMethod, FileType
from models.account import Account, Tenant, TenantAccountJoin, TenantAccountRole
from models.base import TypeBase
from models.enums import EndUserType
from models.model import App, AppMode, EndUser
from models.workflow import Workflow, WorkflowType
TENANT_ID = "00000000-0000-0000-0000-000000000001"
OTHER_TENANT_ID = "00000000-0000-0000-0000-000000000002"
APP_ID = "00000000-0000-0000-0000-000000000003"
ACCOUNT_ID = "00000000-0000-0000-0000-000000000004"
END_USER_ID = "00000000-0000-0000-0000-000000000005"
CREATOR_ID = "00000000-0000-0000-0000-000000000006"
class StubScalars:
"""Minimal stub for SQLAlchemy scalar results."""
_value: Any
def __init__(self, value: Any) -> None:
self._value = value
def first(self) -> Any:
return self._value
@dataclass(frozen=True)
class SqliteToolDb:
engine: Engine
session_maker: sessionmaker[Session]
caller_session: Session
class StubSession:
"""Minimal stub for session_factory-created sessions."""
@pytest.fixture
def sqlite_tool_db(
monkeypatch: pytest.MonkeyPatch,
sqlite_engine: Engine,
) -> Iterator[SqliteToolDb]:
"""Bind service-owned sessions and Account tenant reloads to SQLite."""
models = (App, Workflow, EndUser, Account, Tenant, TenantAccountJoin)
TypeBase.metadata.create_all(sqlite_engine, tables=[model.__table__ for model in models])
session_maker = sessionmaker(bind=sqlite_engine, expire_on_commit=False)
monkeypatch.setattr(workflow_tool_module.session_factory, "create_session", session_maker)
scalar_results: list[Any]
scalars_results: list[Any]
expunge_calls: list[object]
from models import account as account_module
def __init__(self, *, scalar_results: list[Any] | None = None, scalars_results: list[Any] | None = None) -> None:
self.scalar_results = list(scalar_results or [])
self.scalars_results = list(scalars_results or [])
self.expunge_calls: list[object] = []
def scalar(self, _stmt: Any) -> Any:
return self.scalar_results.pop(0)
def scalars(self, _stmt: Any) -> StubScalars:
return StubScalars(self.scalars_results.pop(0))
def expunge(self, value: Any) -> None:
self.expunge_calls.append(value)
def begin(self) -> "StubSession":
return self
def commit(self) -> None:
pass
def refresh(self, _value: Any) -> None:
pass
def close(self) -> None:
pass
def __enter__(self) -> "StubSession":
return self
def __exit__(self, exc_type: Any, exc: Any, tb: Any) -> bool:
return False
monkeypatch.setattr(account_module, "db", SimpleNamespace(engine=sqlite_engine))
with session_maker() as caller_session:
yield SqliteToolDb(engine=sqlite_engine, session_maker=session_maker, caller_session=caller_session)
def _build_tool() -> WorkflowTool:
def _persist_tenant(db: SqliteToolDb, *, tenant_id: str = TENANT_ID) -> Tenant:
tenant = Tenant(name="Tenant")
tenant.id = tenant_id
db.caller_session.add(tenant)
db.caller_session.commit()
return tenant
def _persist_account(db: SqliteToolDb, *, tenant_id: str = TENANT_ID) -> Account:
account = Account(name="Account", email="account@example.com")
account.id = ACCOUNT_ID
join = TenantAccountJoin(
tenant_id=tenant_id,
account_id=account.id,
current=True,
role=TenantAccountRole.NORMAL,
)
db.caller_session.add_all([account, join])
db.caller_session.commit()
return account
def _persist_end_user(
db: SqliteToolDb,
*,
end_user_id: str = END_USER_ID,
tenant_id: str = TENANT_ID,
) -> EndUser:
end_user = EndUser(
id=end_user_id,
tenant_id=tenant_id,
app_id=APP_ID,
type=EndUserType.SERVICE_API,
name="End user",
session_id="end-user-session",
)
db.caller_session.add(end_user)
db.caller_session.commit()
return end_user
def _persist_app(db: SqliteToolDb) -> App:
app = App(
id=APP_ID,
tenant_id=TENANT_ID,
name="Workflow app",
description="",
mode=AppMode.WORKFLOW,
icon_type=None,
icon="",
icon_background=None,
app_model_config_id=None,
workflow_id=None,
enable_site=False,
enable_api=True,
max_active_requests=None,
created_by=CREATOR_ID,
)
db.caller_session.add(app)
db.caller_session.commit()
return app
def _persist_workflow(db: SqliteToolDb, *, version: str, workflow_id: str | None = None) -> Workflow:
workflow = Workflow.new(
tenant_id=TENANT_ID,
app_id=APP_ID,
type=WorkflowType.WORKFLOW.value,
version=version,
graph=json.dumps({"nodes": [], "edges": []}),
features="{}",
created_by=CREATOR_ID,
environment_variables=[],
conversation_variables=[],
rag_pipeline_variables=[],
)
workflow.id = workflow_id or str(uuid.uuid4())
db.caller_session.add(workflow)
db.caller_session.commit()
return workflow
def _build_tool(*, tenant_id: str = "test_tool", workflow_app_id: str = "app-1", version: str = "1") -> WorkflowTool:
entity = ToolEntity(
identity=ToolIdentity(author="test", name="test tool", label=I18nObject(en_US="test tool"), provider="test"),
parameters=[],
description=None,
has_runtime_parameters=False,
)
runtime = ToolRuntime(tenant_id="test_tool", invoke_from=InvokeFrom.EXPLORE)
runtime = ToolRuntime(tenant_id=tenant_id, invoke_from=InvokeFrom.EXPLORE)
return WorkflowTool(
workflow_app_id="app-1",
workflow_app_id=workflow_app_id,
workflow_as_tool_id="wf-tool-1",
version="1",
version=version,
workflow_entities={},
workflow_call_depth=1,
entity=entity,
@@ -98,7 +166,10 @@ def _build_tool() -> WorkflowTool:
)
def test_workflow_tool_should_raise_tool_invoke_error_when_result_has_error_field(monkeypatch: pytest.MonkeyPatch):
def test_workflow_tool_should_raise_tool_invoke_error_when_result_has_error_field(
monkeypatch: pytest.MonkeyPatch,
sqlite_tool_db: SqliteToolDb,
):
"""Ensure that WorkflowTool will throw a `ToolInvokeError` exception when
`WorkflowAppGenerator.generate` returns a result with `error` key inside
the `data` element.
@@ -122,11 +193,14 @@ def test_workflow_tool_should_raise_tool_invoke_error_when_result_has_error_fiel
with pytest.raises(ToolInvokeError) as exc_info:
# WorkflowTool always returns a generator, so we need to iterate to
# actually `run` the tool.
list(tool.invoke(MagicMock(), "test_user", {}))
list(tool.invoke(sqlite_tool_db.caller_session, "test_user", {}))
assert exc_info.value.args == ("oops",)
def test_workflow_tool_does_not_use_pause_state_config(monkeypatch: pytest.MonkeyPatch):
def test_workflow_tool_does_not_use_pause_state_config(
monkeypatch: pytest.MonkeyPatch,
sqlite_tool_db: SqliteToolDb,
):
"""Ensure pause_state_config is passed as None."""
tool = _build_tool()
@@ -140,14 +214,17 @@ def test_workflow_tool_does_not_use_pause_state_config(monkeypatch: pytest.Monke
monkeypatch.setattr("core.app.apps.workflow.app_generator.WorkflowAppGenerator.generate", generate_mock)
monkeypatch.setattr("libs.login.current_user", lambda *args, **kwargs: None)
list(tool.invoke(MagicMock(), "test_user", {}))
list(tool.invoke(sqlite_tool_db.caller_session, "test_user", {}))
call_kwargs = generate_mock.call_args.kwargs
assert "pause_state_config" in call_kwargs
assert call_kwargs["pause_state_config"] is None
def test_workflow_tool_passes_parent_trace_context_from_runtime(monkeypatch: pytest.MonkeyPatch):
def test_workflow_tool_passes_parent_trace_context_from_runtime(
monkeypatch: pytest.MonkeyPatch,
sqlite_tool_db: SqliteToolDb,
):
"""Ensure nested workflow runtime metadata is forwarded as parent trace context."""
tool = _build_tool()
tool.set_parent_trace_context(
@@ -165,7 +242,7 @@ def test_workflow_tool_passes_parent_trace_context_from_runtime(monkeypatch: pyt
monkeypatch.setattr("core.app.apps.workflow.app_generator.WorkflowAppGenerator.generate", generate_mock)
monkeypatch.setattr("libs.login.current_user", lambda *args, **kwargs: None)
list(tool.invoke(MagicMock(), "test_user", {}))
list(tool.invoke(sqlite_tool_db.caller_session, "test_user", {}))
call_kwargs = generate_mock.call_args.kwargs
assert call_kwargs["args"]["parent_trace_context"].model_dump() == {
@@ -174,7 +251,10 @@ def test_workflow_tool_passes_parent_trace_context_from_runtime(monkeypatch: pyt
}
def test_workflow_tool_passes_parent_trace_session_id(monkeypatch: pytest.MonkeyPatch):
def test_workflow_tool_passes_parent_trace_session_id(
monkeypatch: pytest.MonkeyPatch,
sqlite_tool_db: SqliteToolDb,
):
"""Ensure nested workflows inherit the parent observability session ID."""
tool = _build_tool()
tool.entity.parameters = [
@@ -197,14 +277,17 @@ def test_workflow_tool_passes_parent_trace_session_id(monkeypatch: pytest.Monkey
monkeypatch.setattr("core.app.apps.workflow.app_generator.WorkflowAppGenerator.generate", generate_mock)
monkeypatch.setattr("libs.login.current_user", lambda *args, **kwargs: None)
list(tool.invoke(MagicMock(), "test_user", {"trace_session_id": "user-input-session"}))
list(tool.invoke(sqlite_tool_db.caller_session, "test_user", {"trace_session_id": "user-input-session"}))
call_kwargs = generate_mock.call_args.kwargs
assert call_kwargs["args"]["inputs"]["trace_session_id"] == "user-input-session"
assert call_kwargs["args"]["trace_session_id"] == "session-1"
def test_workflow_tool_keeps_user_inputs_named_like_trace_runtime_keys(monkeypatch: pytest.MonkeyPatch):
def test_workflow_tool_keeps_user_inputs_named_like_trace_runtime_keys(
monkeypatch: pytest.MonkeyPatch,
sqlite_tool_db: SqliteToolDb,
):
"""Ensure private trace context does not overwrite same-named workflow inputs."""
tool = _build_tool()
tool.entity.parameters = [
@@ -238,7 +321,7 @@ def test_workflow_tool_keeps_user_inputs_named_like_trace_runtime_keys(monkeypat
list(
tool.invoke(
MagicMock(),
sqlite_tool_db.caller_session,
"test_user",
{
"outer_workflow_run_id": "user-workflow-input",
@@ -256,7 +339,10 @@ def test_workflow_tool_keeps_user_inputs_named_like_trace_runtime_keys(monkeypat
}
def test_workflow_tool_can_clear_parent_trace_context(monkeypatch: pytest.MonkeyPatch):
def test_workflow_tool_can_clear_parent_trace_context(
monkeypatch: pytest.MonkeyPatch,
sqlite_tool_db: SqliteToolDb,
):
"""Ensure reused WorkflowTool instances do not keep stale parent trace context."""
tool = _build_tool()
tool.set_parent_trace_context(
@@ -275,13 +361,16 @@ def test_workflow_tool_can_clear_parent_trace_context(monkeypatch: pytest.Monkey
monkeypatch.setattr("core.app.apps.workflow.app_generator.WorkflowAppGenerator.generate", generate_mock)
monkeypatch.setattr("libs.login.current_user", lambda *args, **kwargs: None)
list(tool.invoke(MagicMock(), "test_user", {}))
list(tool.invoke(sqlite_tool_db.caller_session, "test_user", {}))
call_kwargs = generate_mock.call_args.kwargs
assert "parent_trace_context" not in call_kwargs["args"]
def test_workflow_tool_can_clear_trace_session_id(monkeypatch: pytest.MonkeyPatch):
def test_workflow_tool_can_clear_trace_session_id(
monkeypatch: pytest.MonkeyPatch,
sqlite_tool_db: SqliteToolDb,
):
"""Ensure reused WorkflowTool instances do not keep stale trace session IDs."""
tool = _build_tool()
tool.set_trace_session_id("session-1")
@@ -297,7 +386,7 @@ def test_workflow_tool_can_clear_trace_session_id(monkeypatch: pytest.MonkeyPatc
monkeypatch.setattr("core.app.apps.workflow.app_generator.WorkflowAppGenerator.generate", generate_mock)
monkeypatch.setattr("libs.login.current_user", lambda *args, **kwargs: None)
list(tool.invoke(MagicMock(), "test_user", {}))
list(tool.invoke(sqlite_tool_db.caller_session, "test_user", {}))
call_kwargs = generate_mock.call_args.kwargs
assert "trace_session_id" not in call_kwargs["args"]
@@ -315,6 +404,7 @@ def test_workflow_tool_can_clear_trace_session_id(monkeypatch: pytest.MonkeyPatc
def test_workflow_tool_omits_parent_trace_context_when_runtime_is_incomplete(
monkeypatch: pytest.MonkeyPatch,
runtime_parameters: dict[str, Any],
sqlite_tool_db: SqliteToolDb,
):
"""Ensure incomplete runtime metadata does not leak parent trace context into generator args."""
tool = _build_tool()
@@ -330,13 +420,16 @@ def test_workflow_tool_omits_parent_trace_context_when_runtime_is_incomplete(
monkeypatch.setattr("core.app.apps.workflow.app_generator.WorkflowAppGenerator.generate", generate_mock)
monkeypatch.setattr("libs.login.current_user", lambda *args, **kwargs: None)
list(tool.invoke(MagicMock(), "test_user", {}))
list(tool.invoke(sqlite_tool_db.caller_session, "test_user", {}))
call_kwargs = generate_mock.call_args.kwargs
assert "parent_trace_context" not in call_kwargs["args"]
def test_workflow_tool_should_generate_variable_messages_for_outputs(monkeypatch: pytest.MonkeyPatch):
def test_workflow_tool_should_generate_variable_messages_for_outputs(
monkeypatch: pytest.MonkeyPatch,
sqlite_tool_db: SqliteToolDb,
):
"""Test that WorkflowTool should generate variable messages when there are outputs"""
tool = _build_tool()
@@ -359,7 +452,7 @@ def test_workflow_tool_should_generate_variable_messages_for_outputs(monkeypatch
monkeypatch.setattr("libs.login.current_user", lambda *args, **kwargs: None)
# Execute tool invocation
messages = list(tool.invoke(MagicMock(), "test_user", {}))
messages = list(tool.invoke(sqlite_tool_db.caller_session, "test_user", {}))
# Verify variable messages
variable_messages = [msg for msg in messages if msg.type == ToolInvokeMessage.MessageType.VARIABLE]
@@ -382,7 +475,10 @@ def test_workflow_tool_should_generate_variable_messages_for_outputs(monkeypatch
assert json_messages[0].message.json_object == mock_outputs
def test_workflow_tool_should_handle_empty_outputs(monkeypatch: pytest.MonkeyPatch):
def test_workflow_tool_should_handle_empty_outputs(
monkeypatch: pytest.MonkeyPatch,
sqlite_tool_db: SqliteToolDb,
):
"""Test that WorkflowTool should handle empty outputs correctly"""
tool = _build_tool()
@@ -402,7 +498,7 @@ def test_workflow_tool_should_handle_empty_outputs(monkeypatch: pytest.MonkeyPat
monkeypatch.setattr("libs.login.current_user", lambda *args, **kwargs: None)
# Execute tool invocation
messages = list(tool.invoke(MagicMock(), "test_user", {}))
messages = list(tool.invoke(sqlite_tool_db.caller_session, "test_user", {}))
# Verify generated messages
# Should contain: 0 variable messages + 1 text message + 1 JSON message = 2 messages
@@ -458,41 +554,32 @@ def test_create_file_message_should_include_file_marker():
assert message.meta == {"file": file_obj}
def test_resolve_user_from_database_falls_back_to_end_user(monkeypatch: pytest.MonkeyPatch):
def test_resolve_user_from_database_falls_back_to_end_user(sqlite_tool_db: SqliteToolDb):
"""Ensure worker context can resolve EndUser when Account is missing."""
tenant = SimpleNamespace(id="tenant_id")
end_user = SimpleNamespace(id="end_user_id", tenant_id="tenant_id")
# Monkeypatch session factory to return our stub session
stub_session = StubSession(scalar_results=[tenant, None, end_user])
monkeypatch.setattr(
"core.tools.workflow_as_tool.tool.session_factory.create_session",
lambda: stub_session,
_persist_tenant(sqlite_tool_db)
end_user = _persist_end_user(sqlite_tool_db)
other_tenant_end_user = _persist_end_user(
sqlite_tool_db,
end_user_id="00000000-0000-0000-0000-000000000007",
tenant_id=OTHER_TENANT_ID,
)
tool = _build_tool()
tool = _build_tool(tenant_id=TENANT_ID)
tool.runtime.invoke_from = InvokeFrom.SERVICE_API
tool.runtime.tenant_id = "tenant_id"
resolved_user = tool._resolve_user_from_database(user_id=end_user.id)
assert resolved_user is end_user
assert stub_session.expunge_calls == [end_user]
assert isinstance(resolved_user, EndUser)
assert resolved_user.id == end_user.id
assert resolved_user.tenant_id == TENANT_ID
assert inspect(resolved_user).detached is True
assert tool._resolve_user_from_database(user_id=other_tenant_end_user.id) is None
def test_resolve_user_from_database_returns_none_when_no_tenant(monkeypatch: pytest.MonkeyPatch):
def test_resolve_user_from_database_returns_none_when_no_tenant(sqlite_tool_db: SqliteToolDb):
"""Return None if tenant cannot be found in worker context."""
# Monkeypatch session factory to return our stub session with no tenant
monkeypatch.setattr(
"core.tools.workflow_as_tool.tool.session_factory.create_session",
lambda: StubSession(scalar_results=[None]),
)
tool = _build_tool()
tool = _build_tool(tenant_id=OTHER_TENANT_ID)
tool.runtime.invoke_from = InvokeFrom.SERVICE_API
tool.runtime.tenant_id = "missing_tenant"
resolved_user = tool._resolve_user_from_database(user_id="any")
@@ -544,7 +631,10 @@ def test_extract_usage_from_nested():
assert nested == {"total_tokens": 3}
def test_invoke_raises_when_user_not_found(monkeypatch: pytest.MonkeyPatch):
def test_invoke_raises_when_user_not_found(
monkeypatch: pytest.MonkeyPatch,
sqlite_tool_db: SqliteToolDb,
):
"""Raise ToolInvokeError when user resolution fails."""
tool = _build_tool()
monkeypatch.setattr(tool, "_get_app", lambda *args, **kwargs: None)
@@ -552,58 +642,45 @@ def test_invoke_raises_when_user_not_found(monkeypatch: pytest.MonkeyPatch):
monkeypatch.setattr(tool, "_resolve_user", lambda *args, **kwargs: None)
with pytest.raises(ToolInvokeError, match="User not found"):
list(tool.invoke(MagicMock(), "missing", {}))
list(tool.invoke(sqlite_tool_db.caller_session, "missing", {}))
def test_resolve_user_from_database_returns_account(monkeypatch: pytest.MonkeyPatch):
def test_resolve_user_from_database_returns_account(sqlite_tool_db: SqliteToolDb):
"""Resolve Account and set tenant in worker context."""
tenant = SimpleNamespace(id="tenant_id")
account = SimpleNamespace(id="account_id", current_tenant=None)
set_current_tenant = Mock(side_effect=lambda tenant, *, session: setattr(account, "current_tenant", tenant))
account.set_current_tenant_with_session = set_current_tenant
session = StubSession(scalar_results=[tenant, account])
tenant = _persist_tenant(sqlite_tool_db)
account = _persist_account(sqlite_tool_db)
tool = _build_tool(tenant_id=TENANT_ID)
monkeypatch.setattr("core.tools.workflow_as_tool.tool.session_factory.create_session", lambda: session)
tool = _build_tool()
tool.runtime.tenant_id = "tenant_id"
resolved = tool._resolve_user_from_database(user_id="account_id")
assert resolved is account
assert account.current_tenant is tenant
set_current_tenant.assert_called_once_with(tenant, session=session)
assert session.expunge_calls == [account]
resolved = tool._resolve_user_from_database(user_id=account.id)
assert isinstance(resolved, Account)
assert resolved.id == account.id
assert resolved.current_tenant_id == tenant.id
assert inspect(resolved).detached is True
def test_get_workflow_and_get_app_db_branches(monkeypatch: pytest.MonkeyPatch):
def test_get_workflow_and_get_app_db_branches(sqlite_tool_db: SqliteToolDb):
"""Cover workflow/app retrieval branches and error cases."""
tool = _build_tool()
latest_workflow = SimpleNamespace(id="wf-latest")
specific_workflow = SimpleNamespace(id="wf-v1")
app = SimpleNamespace(id="app-1")
sessions = iter(
[
StubSession(scalar_results=[], scalars_results=[latest_workflow]),
StubSession(scalar_results=[specific_workflow], scalars_results=[]),
StubSession(scalar_results=[app], scalars_results=[]),
]
)
monkeypatch.setattr(
"core.tools.workflow_as_tool.tool.session_factory.create_session",
lambda: next(sessions),
)
app = _persist_app(sqlite_tool_db)
specific_workflow = _persist_workflow(sqlite_tool_db, version="1")
latest_workflow = _persist_workflow(sqlite_tool_db, version="2")
_persist_workflow(sqlite_tool_db, version=Workflow.VERSION_DRAFT)
tool = _build_tool(tenant_id=TENANT_ID, workflow_app_id=APP_ID)
assert tool._get_workflow("app-1", "") is latest_workflow
assert tool._get_workflow("app-1", "1") is specific_workflow
assert tool._get_app("app-1") is app
latest = tool._get_workflow(APP_ID, "")
specific = tool._get_workflow(APP_ID, "1")
resolved_app = tool._get_app(APP_ID)
assert latest.id == latest_workflow.id
assert specific.id == specific_workflow.id
assert resolved_app.id == app.id
assert inspect(latest).detached is True
assert inspect(specific).detached is True
assert inspect(resolved_app).detached is True
monkeypatch.setattr(
"core.tools.workflow_as_tool.tool.session_factory.create_session",
lambda: StubSession(scalar_results=[None, None], scalars_results=[None]),
)
with pytest.raises(ValueError, match="workflow not found"):
tool._get_workflow("app-1", "1")
tool._get_workflow(APP_ID, "missing")
with pytest.raises(ValueError, match="app not found"):
tool._get_app("app-1")
tool._get_app("00000000-0000-0000-0000-000000000099")
def _setup_transform_args_tool(monkeypatch: pytest.MonkeyPatch) -> WorkflowTool:
@@ -722,7 +799,10 @@ def test_transform_args_normalizes_optional_files_parameter(
assert files == []
def test_workflow_tool_invocation_normalizes_optional_files_parameter(monkeypatch: pytest.MonkeyPatch):
def test_workflow_tool_invocation_normalizes_optional_files_parameter(
monkeypatch: pytest.MonkeyPatch,
sqlite_tool_db: SqliteToolDb,
):
"""Ensure casted empty FILES values do not reach workflow input validation as [None]."""
tool = _build_tool()
images_param = ToolParameter.get_simple_instance(
@@ -741,7 +821,7 @@ def test_workflow_tool_invocation_normalizes_optional_files_parameter(monkeypatc
generate_mock = MagicMock(return_value={"data": {}})
monkeypatch.setattr("core.app.apps.workflow.app_generator.WorkflowAppGenerator.generate", generate_mock)
list(tool.invoke(MagicMock(), "test_user", {"images": None}))
list(tool.invoke(sqlite_tool_db.caller_session, "test_user", {"images": None}))
call_kwargs = generate_mock.call_args.kwargs
assert call_kwargs["args"]["inputs"]["images"] == []
@@ -26,6 +26,7 @@ class TestPlannerSystemPrompt:
"""Auto-mode resolution rides on the planner echoing its mode choice."""
assert '"mode": "workflow | advanced-chat"' in PLANNER_SYSTEM_PROMPT
assert "When the ``# Mode`` section says auto, YOU decide" in PLANNER_SYSTEM_PROMPT
assert "userinput.files" in PLANNER_SYSTEM_PROMPT
class TestFormatIdealOutputSection:
@@ -75,6 +76,9 @@ class TestNodeBuilderPrompt:
assert '"viewport":' not in prompt
assert '"positionAbsolute":' not in prompt
def test_start_uses_user_input_files(self):
assert "userinput.files" in get_node_builder_system_prompt("start")
def test_supports_main_human_input_and_assigner_contracts(self):
human_input = get_node_builder_system_prompt("human-input")
assigner = get_node_builder_system_prompt("assigner")
@@ -130,17 +134,19 @@ class TestNodeBuilderUserSections:
class TestModeSection:
def test_advanced_chat_documents_system_variables(self):
def test_advanced_chat_documents_built_in_variables(self):
out = format_mode_section("advanced-chat")
assert "sys.query" in out
assert '["sys", "query"]' in out
assert "userinput.files" in out
assert '["userinput", "files"]' in out
assert "do NOT invent start-node variables" in out
def test_workflow_mode_forbids_system_variables(self):
def test_workflow_mode_documents_file_input(self):
out = format_mode_section("workflow")
assert "NO automatic system variables" in out
assert "userinput.files" in out
assert "start node's declared variables" in out
class TestExistingGraphSection:
@@ -228,12 +228,12 @@ def _previous_node_prompt_payload(result, selector: str) -> object:
def _uploaded_workflow_files_prompt_payload(result) -> object:
prefix = " - sys.files: "
prefix = " - userinput.files: "
user_prompt = _workflow_user_prompt(result)
for line in user_prompt.splitlines():
if line.startswith(prefix):
return json.loads(line.removeprefix(prefix))
raise AssertionError("missing prompt payload for sys.files")
raise AssertionError("missing prompt payload for userinput.files")
def test_builds_create_run_request_from_agent_soul_and_node_job():
@@ -0,0 +1,161 @@
from core.workflow.legacy_system_files import (
LegacySysFilesCompatVariable,
attach_legacy_sys_files_warning,
migrate_legacy_sys_files_graph_with_result,
normalize_legacy_sys_files_args,
resolve_legacy_sys_files_compat_variable,
)
_LEGACY_NODE_ID = "sys"
_LEGACY_ALIAS_NODE_ID = "userinput"
_LEGACY_VARIABLE_NAME = "files"
_LEGACY_SELECTOR = [_LEGACY_NODE_ID, _LEGACY_VARIABLE_NAME]
_LEGACY_TEMPLATE = "{{#" + ".".join((_LEGACY_NODE_ID, _LEGACY_VARIABLE_NAME)) + "#}}"
_LEGACY_ALIAS_SELECTOR = [_LEGACY_ALIAS_NODE_ID, _LEGACY_VARIABLE_NAME]
_LEGACY_ALIAS_TEMPLATE = "{{#" + ".".join((_LEGACY_ALIAS_NODE_ID, _LEGACY_VARIABLE_NAME)) + "#}}"
_LEGACY_ALIAS_INPUT_KEY = ".".join((_LEGACY_ALIAS_NODE_ID, _LEGACY_VARIABLE_NAME))
def test_migrate_legacy_sys_files_graph_ignores_invalid_or_unrelated_graphs():
assert not migrate_legacy_sys_files_graph_with_result({}).changed
assert not migrate_legacy_sys_files_graph_with_result({"nodes": [], "edges": [_LEGACY_SELECTOR]}).changed
assert not migrate_legacy_sys_files_graph_with_result({"nodes": [{"data": {"value": ["sys", "query"]}}]}).changed
def test_migrate_legacy_sys_files_graph_rewrites_sys_files_to_userinput_files_without_start_variable():
graph = {
"nodes": [
{"id": "start", "data": {"type": "start", "variables": [{"variable": "sys_files"}]}},
{
"id": "answer",
"data": {
"type": "answer",
"answer": _LEGACY_SELECTOR,
"template": _LEGACY_TEMPLATE,
},
},
],
}
result = migrate_legacy_sys_files_graph_with_result(graph)
assert result.changed
start_data = result.graph["nodes"][0]["data"]
assert start_data["variables"] == [{"variable": "sys_files"}]
assert result.graph["nodes"][1]["data"]["answer"] == _LEGACY_ALIAS_SELECTOR
assert result.graph["nodes"][1]["data"]["template"] == _LEGACY_ALIAS_TEMPLATE
assert graph["nodes"][1]["data"]["answer"] == _LEGACY_SELECTOR
assert graph["nodes"][1]["data"]["template"] == _LEGACY_TEMPLATE
def test_migrate_legacy_sys_files_graph_leaves_userinput_files_target_unchanged():
graph = {
"nodes": [
{"id": "start", "data": {"type": "start", "variables": []}},
{
"id": "answer",
"data": {
"type": "answer",
"answer": _LEGACY_ALIAS_SELECTOR,
"template": _LEGACY_ALIAS_TEMPLATE,
},
},
],
}
result = migrate_legacy_sys_files_graph_with_result(graph)
assert not result.changed
assert result.graph == graph
def test_resolve_legacy_sys_files_compat_variable_returns_userinput_files_target():
assert resolve_legacy_sys_files_compat_variable({}) is None
assert resolve_legacy_sys_files_compat_variable({"nodes": [{"data": {"value": ["sys", "query"]}}]}) is None
compat_variable = resolve_legacy_sys_files_compat_variable({"nodes": [{"data": {"value": _LEGACY_SELECTOR}}]})
assert compat_variable == LegacySysFilesCompatVariable(
node_id=_LEGACY_ALIAS_NODE_ID,
variable_name=_LEGACY_VARIABLE_NAME,
)
assert (
resolve_legacy_sys_files_compat_variable({"nodes": [{"data": {"value": _LEGACY_ALIAS_SELECTOR}}]})
== compat_variable
)
def test_normalize_legacy_sys_files_args_handles_no_compat_and_top_level_files():
args_without_legacy, compat_without_legacy = normalize_legacy_sys_files_args(
graph={"nodes": []},
args={"inputs": {}},
)
assert args_without_legacy == {"inputs": {}}
assert compat_without_legacy is None
files = [{"id": "file-1"}]
graph = {
"nodes": [
{"id": "start", "data": {"type": "start", "variables": []}},
{"id": "answer", "data": {"type": "answer", "answer": _LEGACY_TEMPLATE}},
],
}
normalized_args, compat_variable = normalize_legacy_sys_files_args(
graph=graph,
args={"inputs": {}, "files": files},
)
assert compat_variable is not None
assert normalized_args["files"] == files
assert normalized_args["inputs"][".".join((compat_variable.node_id, compat_variable.variable_name))] == files
def test_normalize_legacy_sys_files_args_maps_userinput_files_to_top_level_files_without_warning():
files = [{"id": "file-1"}]
normalized_args, compat_variable = normalize_legacy_sys_files_args(
graph={"nodes": []},
args={"inputs": {_LEGACY_ALIAS_INPUT_KEY: files}},
)
assert compat_variable is None
assert normalized_args["files"] == files
assert normalized_args["inputs"] == {_LEGACY_ALIAS_INPUT_KEY: files}
def test_normalize_legacy_sys_files_args_prefers_userinput_files_over_legacy_files():
legacy_files = [{"id": "legacy-file"}]
userinput_files = [{"id": "userinput-file"}]
normalized_args, compat_variable = normalize_legacy_sys_files_args(
graph={"nodes": [{"data": {"type": "answer", "answer": _LEGACY_ALIAS_TEMPLATE}}]},
args={
"inputs": {_LEGACY_ALIAS_INPUT_KEY: userinput_files},
"files": legacy_files,
},
)
assert compat_variable is None
assert normalized_args["files"] == userinput_files
def test_attach_legacy_sys_files_warning_wraps_stream_and_closes_source():
class CloseableStream:
closed = False
def __iter__(self):
yield "data: payload\n\n"
def close(self):
self.closed = True
stream = CloseableStream()
wrapped = attach_legacy_sys_files_warning(
stream,
LegacySysFilesCompatVariable(node_id=_LEGACY_ALIAS_NODE_ID, variable_name=_LEGACY_VARIABLE_NAME),
)
chunks = list(wrapped)
assert "warning" in chunks[0]
assert chunks[1] == "data: payload\n\n"
assert stream.closed
@@ -1,6 +1,7 @@
from types import SimpleNamespace
from core.workflow.system_variables import (
build_bootstrap_variables,
build_system_variables,
default_system_variables,
get_node_creation_preload_selectors,
@@ -56,6 +57,25 @@ def test_build_system_variables_preserves_file_values():
assert system_values["files"] == [file]
def test_build_bootstrap_variables_adds_userinput_files_alias():
file = File(
file_type=FileType.DOCUMENT,
transfer_method=FileTransferMethod.LOCAL_FILE,
related_id="file-id",
filename="test.txt",
extension=".txt",
mime_type="text/plain",
size=1,
storage_key="storage-key",
)
bootstrap_variables = build_bootstrap_variables(system_variables=build_system_variables(files=[file]))
file_variables_by_selector = {tuple(variable.selector): variable for variable in bootstrap_variables}
assert file_variables_by_selector[("sys", "files")].value == [file]
assert file_variables_by_selector[("userinput", "files")].value == [file]
def test_default_system_variables_generates_workflow_run_id():
system_variables = default_system_variables()
system_values = system_variables_to_mapping(system_variables)
@@ -18,7 +18,6 @@ from models.provider import ProviderType
@pytest.fixture
def credit_pool_session_factory(sqlite_engine: Engine) -> Iterator[sessionmaker[Session]]:
"""Bind message-created accounting to fixture-owned SQLite sessions."""
TenantCreditPool.__table__.create(sqlite_engine)
session_factory = sessionmaker(bind=sqlite_engine, expire_on_commit=False)
with patch("events.event_handlers.update_provider_when_message_created.db.session", session_factory):
yield session_factory
@@ -18,6 +18,11 @@ from models.workflow import (
is_system_variable_editable,
)
_LEGACY_FILE_TEMPLATE = "{{#" + ".".join(("sys", "files")) + "#}}"
_LEGACY_FILE_SELECTOR = ["sys", "files"]
_USER_INPUT_FILE_TEMPLATE = "{{#" + ".".join(("userinput", "files")) + "#}}"
_USER_INPUT_FILE_SELECTOR = ["userinput", "files"]
def test_environment_variables():
# tenant_id context variable removed - using current_user.current_tenant_id directly
@@ -245,6 +250,144 @@ class TestIsSystemVariableEditable:
assert is_system_variable_editable("invalid_or_new_system_variable") == False
class TestWorkflowLegacySysFilesCompatibility:
def _make_workflow(self, graph: dict, *, features: dict | None = None) -> Workflow:
return Workflow(
tenant_id="tenant_id",
app_id="app_id",
type="workflow",
version="draft",
graph=json.dumps(graph),
features=json.dumps(features or {}),
created_by="account_id",
environment_variables=[],
conversation_variables=[],
)
def test_graph_dict_rewrites_legacy_sys_files_references_to_userinput_files(self):
workflow = self._make_workflow(
{
"nodes": [
{
"id": "start",
"data": {
"type": "start",
"title": "Start",
"variables": [],
},
},
{
"id": "llm",
"data": {
"type": "llm",
"prompt_template": [{"role": "user", "text": f"files: {_LEGACY_FILE_TEMPLATE}"}],
"context": {"variable_selector": _LEGACY_FILE_SELECTOR},
},
},
],
"edges": [],
}
)
stored_graph_before_read = workflow.graph
graph = workflow.graph_dict
start_node = next(node for node in graph["nodes"] if node["id"] == "start")
llm_node = next(node for node in graph["nodes"] if node["id"] == "llm")
assert start_node["data"]["variables"] == []
assert llm_node["data"]["prompt_template"][0]["text"] == f"files: {_USER_INPUT_FILE_TEMPLATE}"
assert llm_node["data"]["context"]["variable_selector"] == _USER_INPUT_FILE_SELECTOR
assert workflow.graph == stored_graph_before_read
def test_migrate_legacy_sys_files_graph_in_place_updates_stored_graph(self):
workflow = self._make_workflow(
{
"nodes": [
{"id": "answer", "data": {"type": "answer", "answer": _LEGACY_FILE_TEMPLATE}},
],
"edges": [],
}
)
assert workflow.migrate_legacy_sys_files_graph_in_place()
assert _LEGACY_FILE_TEMPLATE not in workflow.graph
assert _USER_INPUT_FILE_TEMPLATE in workflow.graph
def test_graph_dict_preserves_existing_start_variables_when_migrating_legacy_sys_files(self):
workflow = self._make_workflow(
{
"nodes": [
{
"id": "start",
"data": {
"type": "start",
"title": "Start",
"variables": [
{"variable": "sys_files", "label": "Existing", "type": "text-input"},
],
},
},
{
"id": "answer",
"data": {
"type": "answer",
"answer": _LEGACY_FILE_TEMPLATE,
},
},
],
"edges": [],
}
)
graph = workflow.graph_dict
start_node = next(node for node in graph["nodes"] if node["id"] == "start")
answer_node = next(node for node in graph["nodes"] if node["id"] == "answer")
assert [variable["variable"] for variable in start_node["data"]["variables"]] == ["sys_files"]
assert answer_node["data"]["answer"] == _USER_INPUT_FILE_TEMPLATE
def test_graph_dict_leaves_userinput_files_references_unchanged(self):
workflow = self._make_workflow(
{
"nodes": [
{
"id": "start",
"data": {
"type": "start",
"title": "Start",
"variables": [],
},
},
{
"id": "answer",
"data": {
"type": "answer",
"answer": _USER_INPUT_FILE_TEMPLATE,
},
},
],
"edges": [],
},
features={
"file_upload": {
"enabled": True,
"allowed_file_upload_methods": ["remote_url"],
"allowed_file_types": ["document", "custom"],
"allowed_file_extensions": [".pdf"],
"number_limits": 8,
}
},
)
graph = workflow.graph_dict
start_node = next(node for node in graph["nodes"] if node["id"] == "start")
assert start_node["data"]["variables"] == []
assert json.loads(workflow.graph) == graph
class TestWorkflowDraftVariableGetValue:
def test_get_value_by_case(self):
@dataclasses.dataclass
@@ -1,89 +1,105 @@
import logging
from types import SimpleNamespace
from unittest.mock import MagicMock, patch
from uuid import uuid4
import pytest
from sqlalchemy import select
from sqlalchemy.orm import Session
from models.account import (
TenantPluginAutoUpgradeCategory,
TenantPluginAutoUpgradeMode,
TenantPluginAutoUpgradeStrategy,
TenantPluginAutoUpgradeStrategySetting,
)
from services.plugin.plugin_auto_upgrade_service import PluginAutoUpgradeService
MODULE = "services.plugin.plugin_auto_upgrade_service"
PLUGIN_CATEGORY = TenantPluginAutoUpgradeCategory.TOOL
STRATEGY_MODELS = (TenantPluginAutoUpgradeStrategy,)
def _patched_session():
"""Return a mock SQLAlchemy session for service calls."""
session = MagicMock()
return session
def _strategy(
tenant_id: str,
*,
category: TenantPluginAutoUpgradeCategory = PLUGIN_CATEGORY,
setting: TenantPluginAutoUpgradeStrategySetting = TenantPluginAutoUpgradeStrategySetting.FIX_ONLY,
mode: TenantPluginAutoUpgradeMode = TenantPluginAutoUpgradeMode.EXCLUDE,
exclude: list[str] | None = None,
include: list[str] | None = None,
upgrade_time: int = 0,
) -> TenantPluginAutoUpgradeStrategy:
return TenantPluginAutoUpgradeStrategy(
tenant_id=tenant_id,
category=category,
strategy_setting=setting,
upgrade_time_of_day=upgrade_time,
upgrade_mode=mode,
exclude_plugins=exclude or [],
include_plugins=include or [],
)
class TestGetStrategy:
def test_returns_strategy_when_found(self):
session = _patched_session()
strategy = MagicMock()
session.scalar.return_value = strategy
@pytest.mark.parametrize("sqlite_session", [STRATEGY_MODELS], indirect=True)
def test_returns_strategy_when_found(self, sqlite_session: Session) -> None:
tenant_id = str(uuid4())
strategy = _strategy(tenant_id)
sqlite_session.add(strategy)
sqlite_session.commit()
from services.plugin.plugin_auto_upgrade_service import PluginAutoUpgradeService
result = PluginAutoUpgradeService.get_strategy("t1", PLUGIN_CATEGORY, session=session)
result = PluginAutoUpgradeService.get_strategy(tenant_id, PLUGIN_CATEGORY, session=sqlite_session)
assert result is strategy
def test_returns_none_when_not_found(self):
session = _patched_session()
session.scalar.return_value = None
from services.plugin.plugin_auto_upgrade_service import PluginAutoUpgradeService
result = PluginAutoUpgradeService.get_strategy("t1", PLUGIN_CATEGORY, session=session)
assert result is None
@pytest.mark.parametrize("sqlite_session", [STRATEGY_MODELS], indirect=True)
def test_returns_none_when_not_found(self, sqlite_session: Session) -> None:
assert PluginAutoUpgradeService.get_strategy(str(uuid4()), PLUGIN_CATEGORY, session=sqlite_session) is None
class TestChangeStrategy:
def test_creates_new_strategy(self):
session = _patched_session()
session.scalar.return_value = None
with patch(f"{MODULE}.select"), patch(f"{MODULE}.TenantPluginAutoUpgradeStrategy") as strat_cls:
strat_cls.return_value = MagicMock()
from services.plugin.plugin_auto_upgrade_service import PluginAutoUpgradeService
result = PluginAutoUpgradeService.change_strategy(
"t1",
TenantPluginAutoUpgradeStrategySetting.FIX_ONLY,
3,
TenantPluginAutoUpgradeMode.ALL,
[],
[],
category=PLUGIN_CATEGORY,
session=session,
)
assert result is True
session.add.assert_called_once()
def test_updates_existing_strategy(self):
session = _patched_session()
existing = MagicMock()
session.scalar.return_value = existing
from services.plugin.plugin_auto_upgrade_service import PluginAutoUpgradeService
@pytest.mark.parametrize("sqlite_session", [STRATEGY_MODELS], indirect=True)
def test_creates_new_strategy(self, sqlite_session: Session) -> None:
tenant_id = str(uuid4())
result = PluginAutoUpgradeService.change_strategy(
"t1",
tenant_id,
TenantPluginAutoUpgradeStrategySetting.FIX_ONLY,
3,
TenantPluginAutoUpgradeMode.ALL,
[],
[],
category=PLUGIN_CATEGORY,
session=sqlite_session,
)
strategy = sqlite_session.scalar(select(TenantPluginAutoUpgradeStrategy))
assert result is True
assert strategy is not None
assert strategy.tenant_id == tenant_id
assert strategy.upgrade_time_of_day == 3
assert strategy.upgrade_mode == TenantPluginAutoUpgradeMode.ALL
@pytest.mark.parametrize("sqlite_session", [STRATEGY_MODELS], indirect=True)
def test_updates_existing_strategy(self, sqlite_session: Session) -> None:
tenant_id = str(uuid4())
existing = _strategy(tenant_id)
sqlite_session.add(existing)
sqlite_session.commit()
result = PluginAutoUpgradeService.change_strategy(
tenant_id,
TenantPluginAutoUpgradeStrategySetting.LATEST,
5,
TenantPluginAutoUpgradeMode.PARTIAL,
["p1"],
["p2"],
category=PLUGIN_CATEGORY,
session=session,
session=sqlite_session,
)
sqlite_session.refresh(existing)
assert result is True
assert existing.strategy_setting == TenantPluginAutoUpgradeStrategySetting.LATEST
assert existing.upgrade_time_of_day == 5
@@ -93,157 +109,115 @@ class TestChangeStrategy:
class TestExcludePlugin:
def test_creates_default_strategy_when_none_exists(self):
session = _patched_session()
session.scalar.return_value = None
@pytest.mark.parametrize("sqlite_session", [STRATEGY_MODELS], indirect=True)
def test_creates_default_strategy_when_none_exists(self, sqlite_session: Session) -> None:
tenant_id = str(uuid4())
with (
patch(f"{MODULE}.select"),
patch(f"{MODULE}.TenantPluginAutoUpgradeStrategy"),
):
from services.plugin.plugin_auto_upgrade_service import PluginAutoUpgradeService
result = PluginAutoUpgradeService.exclude_plugin(
"t1",
"plugin-1",
PLUGIN_CATEGORY,
session=session,
)
result = PluginAutoUpgradeService.exclude_plugin(tenant_id, "plugin-1", PLUGIN_CATEGORY, session=sqlite_session)
strategy = sqlite_session.scalar(select(TenantPluginAutoUpgradeStrategy))
assert result is True
session.add.assert_called_once()
assert strategy is not None
assert strategy.exclude_plugins == ["plugin-1"]
def test_appends_to_exclude_list_in_exclude_mode(self):
session = _patched_session()
existing = MagicMock()
existing.upgrade_mode = TenantPluginAutoUpgradeMode.EXCLUDE
existing.exclude_plugins = ["p-existing"]
session.scalar.return_value = existing
@pytest.mark.parametrize("sqlite_session", [STRATEGY_MODELS], indirect=True)
def test_appends_to_exclude_list_in_exclude_mode(self, sqlite_session: Session) -> None:
tenant_id = str(uuid4())
existing = _strategy(tenant_id, exclude=["p-existing"])
sqlite_session.add(existing)
sqlite_session.commit()
with patch(f"{MODULE}.select"), patch(f"{MODULE}.TenantPluginAutoUpgradeStrategy") as strat_cls:
strat_cls.UpgradeMode.EXCLUDE = "exclude"
strat_cls.UpgradeMode.PARTIAL = "partial"
strat_cls.UpgradeMode.ALL = "all"
from services.plugin.plugin_auto_upgrade_service import PluginAutoUpgradeService
PluginAutoUpgradeService.exclude_plugin(tenant_id, "p-new", PLUGIN_CATEGORY, session=sqlite_session)
result = PluginAutoUpgradeService.exclude_plugin("t1", "p-new", PLUGIN_CATEGORY, session=session)
assert result is True
sqlite_session.refresh(existing)
assert existing.exclude_plugins == ["p-existing", "p-new"]
def test_removes_from_include_list_in_partial_mode(self):
session = _patched_session()
existing = MagicMock()
existing.upgrade_mode = TenantPluginAutoUpgradeMode.PARTIAL
existing.include_plugins = ["p1", "p2"]
session.scalar.return_value = existing
@pytest.mark.parametrize("sqlite_session", [STRATEGY_MODELS], indirect=True)
def test_removes_from_include_list_in_partial_mode(self, sqlite_session: Session) -> None:
tenant_id = str(uuid4())
existing = _strategy(tenant_id, mode=TenantPluginAutoUpgradeMode.PARTIAL, include=["p1", "p2"])
sqlite_session.add(existing)
sqlite_session.commit()
with patch(f"{MODULE}.select"), patch(f"{MODULE}.TenantPluginAutoUpgradeStrategy") as strat_cls:
strat_cls.UpgradeMode.EXCLUDE = "exclude"
strat_cls.UpgradeMode.PARTIAL = "partial"
strat_cls.UpgradeMode.ALL = "all"
from services.plugin.plugin_auto_upgrade_service import PluginAutoUpgradeService
PluginAutoUpgradeService.exclude_plugin(tenant_id, "p1", PLUGIN_CATEGORY, session=sqlite_session)
result = PluginAutoUpgradeService.exclude_plugin("t1", "p1", PLUGIN_CATEGORY, session=session)
assert result is True
sqlite_session.refresh(existing)
assert existing.include_plugins == ["p2"]
def test_switches_to_exclude_mode_from_all(self):
session = _patched_session()
existing = MagicMock()
existing.upgrade_mode = TenantPluginAutoUpgradeMode.ALL
session.scalar.return_value = existing
@pytest.mark.parametrize("sqlite_session", [STRATEGY_MODELS], indirect=True)
def test_switches_to_exclude_mode_from_all(self, sqlite_session: Session) -> None:
tenant_id = str(uuid4())
existing = _strategy(tenant_id, mode=TenantPluginAutoUpgradeMode.ALL)
sqlite_session.add(existing)
sqlite_session.commit()
with patch(f"{MODULE}.select"), patch(f"{MODULE}.TenantPluginAutoUpgradeStrategy") as strat_cls:
strat_cls.UpgradeMode.EXCLUDE = "exclude"
strat_cls.UpgradeMode.PARTIAL = "partial"
strat_cls.UpgradeMode.ALL = "all"
from services.plugin.plugin_auto_upgrade_service import PluginAutoUpgradeService
PluginAutoUpgradeService.exclude_plugin(tenant_id, "p1", PLUGIN_CATEGORY, session=sqlite_session)
result = PluginAutoUpgradeService.exclude_plugin("t1", "p1", PLUGIN_CATEGORY, session=session)
assert result is True
sqlite_session.refresh(existing)
assert existing.upgrade_mode == TenantPluginAutoUpgradeMode.EXCLUDE
assert existing.exclude_plugins == ["p1"]
def test_no_duplicate_in_exclude_list(self):
session = _patched_session()
existing = MagicMock()
existing.upgrade_mode = TenantPluginAutoUpgradeMode.EXCLUDE
existing.exclude_plugins = ["p1"]
session.scalar.return_value = existing
@pytest.mark.parametrize("sqlite_session", [STRATEGY_MODELS], indirect=True)
def test_no_duplicate_in_exclude_list(self, sqlite_session: Session) -> None:
tenant_id = str(uuid4())
existing = _strategy(tenant_id, exclude=["p1"])
sqlite_session.add(existing)
sqlite_session.commit()
with patch(f"{MODULE}.select"), patch(f"{MODULE}.TenantPluginAutoUpgradeStrategy") as strat_cls:
strat_cls.UpgradeMode.EXCLUDE = "exclude"
strat_cls.UpgradeMode.PARTIAL = "partial"
strat_cls.UpgradeMode.ALL = "all"
from services.plugin.plugin_auto_upgrade_service import PluginAutoUpgradeService
PluginAutoUpgradeService.exclude_plugin("t1", "p1", PLUGIN_CATEGORY, session=session)
PluginAutoUpgradeService.exclude_plugin(tenant_id, "p1", PLUGIN_CATEGORY, session=sqlite_session)
sqlite_session.refresh(existing)
assert existing.exclude_plugins == ["p1"]
class TestBackfillStrategyCategories:
def test_creates_default_missing_categories_without_fetching_daemon(self):
session = _patched_session()
tool_strategy = SimpleNamespace(
category=TenantPluginAutoUpgradeCategory.TOOL,
strategy_setting=TenantPluginAutoUpgradeStrategySetting.FIX_ONLY,
upgrade_time_of_day=0,
upgrade_mode=TenantPluginAutoUpgradeMode.EXCLUDE,
exclude_plugins=[],
include_plugins=[],
)
session.scalars.return_value.all.return_value = [tool_strategy]
@pytest.mark.parametrize("sqlite_session", [STRATEGY_MODELS], indirect=True)
def test_creates_default_missing_categories_without_fetching_daemon(self, sqlite_session: Session) -> None:
tenant_id = str(uuid4())
tool_strategy = _strategy(tenant_id)
sqlite_session.add(tool_strategy)
sqlite_session.commit()
installer = MagicMock()
with patch(f"{MODULE}.PluginInstaller", return_value=installer):
from services.plugin.plugin_auto_upgrade_service import PluginAutoUpgradeService
result = PluginAutoUpgradeService.backfill_strategy_categories("t1", session=session)
expected_time = PluginAutoUpgradeService.default_upgrade_time_of_day("t1")
result = PluginAutoUpgradeService.backfill_strategy_categories(tenant_id, session=sqlite_session)
expected_time = PluginAutoUpgradeService.default_upgrade_time_of_day(tenant_id)
strategies = list(sqlite_session.scalars(select(TenantPluginAutoUpgradeStrategy)).all())
assert result.created_count == len(TenantPluginAutoUpgradeCategory) - 1
assert result.normalized is False
installer.list_plugins.assert_not_called()
assert len(strategies) == len(TenantPluginAutoUpgradeCategory)
assert tool_strategy.upgrade_time_of_day == expected_time
created_strategies = [call.args[0] for call in session.add.call_args_list]
model_strategy = next(
strategy for strategy in created_strategies if strategy.category == TenantPluginAutoUpgradeCategory.MODEL
strategy for strategy in strategies if strategy.category == TenantPluginAutoUpgradeCategory.MODEL
)
assert model_strategy.strategy_setting == TenantPluginAutoUpgradeStrategySetting.LATEST
assert model_strategy.upgrade_time_of_day == expected_time
def test_default_upgrade_time_is_aligned_to_fifteen_minutes(self):
from services.plugin.plugin_auto_upgrade_service import PluginAutoUpgradeService
default_time = PluginAutoUpgradeService.default_upgrade_time_of_day("t1")
def test_default_upgrade_time_is_aligned_to_fifteen_minutes(self) -> None:
default_time = PluginAutoUpgradeService.default_upgrade_time_of_day(str(uuid4()))
assert default_time % (15 * 60) == 0
assert 0 <= default_time < 24 * 60 * 60
def test_creates_missing_categories_and_splits_known_plugins(self, caplog: pytest.LogCaptureFixture):
session = _patched_session()
tool_strategy = SimpleNamespace(
category=TenantPluginAutoUpgradeCategory.TOOL,
strategy_setting=TenantPluginAutoUpgradeStrategySetting.FIX_ONLY,
upgrade_time_of_day=0,
upgrade_mode=TenantPluginAutoUpgradeMode.EXCLUDE,
exclude_plugins=["tool-plugin", "model-plugin", "unknown-plugin"],
include_plugins=["model-plugin", "tool-plugin"],
@pytest.mark.parametrize("sqlite_session", [STRATEGY_MODELS], indirect=True)
def test_creates_missing_categories_and_splits_known_plugins(
self, sqlite_session: Session, caplog: pytest.LogCaptureFixture
) -> None:
tenant_id = str(uuid4())
tool_strategy = _strategy(
tenant_id,
exclude=["tool-plugin", "model-plugin", "unknown-plugin"],
include=["model-plugin", "tool-plugin"],
)
model_strategy = SimpleNamespace(
model_strategy = _strategy(
tenant_id,
category=TenantPluginAutoUpgradeCategory.MODEL,
strategy_setting=TenantPluginAutoUpgradeStrategySetting.FIX_ONLY,
upgrade_time_of_day=0,
upgrade_mode=TenantPluginAutoUpgradeMode.EXCLUDE,
exclude_plugins=["tool-plugin", "model-plugin", "unknown-plugin"],
include_plugins=["model-plugin", "tool-plugin"],
exclude=["tool-plugin", "model-plugin", "unknown-plugin"],
include=["model-plugin", "tool-plugin"],
)
session.scalars.return_value.all.return_value = [tool_strategy, model_strategy]
sqlite_session.add_all([tool_strategy, model_strategy])
sqlite_session.commit()
installed_plugins = [
SimpleNamespace(
plugin_id="tool-plugin",
@@ -261,18 +235,17 @@ class TestBackfillStrategyCategories:
patch(f"{MODULE}.PluginInstaller", return_value=installer),
caplog.at_level(logging.WARNING, logger=MODULE),
):
from services.plugin.plugin_auto_upgrade_service import PluginAutoUpgradeService
result = PluginAutoUpgradeService.backfill_strategy_categories("t1", session=session)
result = PluginAutoUpgradeService.backfill_strategy_categories(tenant_id, session=sqlite_session)
strategies = list(sqlite_session.scalars(select(TenantPluginAutoUpgradeStrategy)).all())
assert result.created_count == len(TenantPluginAutoUpgradeCategory) - 2
assert result.normalized is True
assert session.add.call_count == len(TenantPluginAutoUpgradeCategory) - 2
assert len(strategies) == len(TenantPluginAutoUpgradeCategory)
assert tool_strategy.exclude_plugins == ["tool-plugin"]
assert tool_strategy.include_plugins == ["tool-plugin"]
assert model_strategy.exclude_plugins == ["model-plugin"]
assert model_strategy.include_plugins == ["model-plugin"]
assert (
"Skipped unknown plugin IDs while backfilling plugin auto-upgrade strategies: "
"tenant_id=t1, field=exclude_plugins, plugin_ids=['unknown-plugin']" in caplog.messages
f"tenant_id={tenant_id}, field=exclude_plugins, plugin_ids=['unknown-plugin']" in caplog.messages
)
@@ -7,28 +7,19 @@ import pytest
import zstandard
from pydantic import TypeAdapter
from redis import RedisError
from sqlalchemy.orm import Session
from core.helper.model_provider_cache import ProviderCredentialsCacheType
from core.plugin.entities.plugin import PluginCategory, PluginInstallationSource
from core.plugin.entities.plugin_daemon import PluginInstallTask, PluginInstallTaskStatus, PluginModelProviderEntity
from graphon.model_runtime.entities.common_entities import I18nObject
from graphon.model_runtime.entities.provider_entities import ConfigurateMethod, ProviderEntity
from models.provider import Provider, ProviderCredential, ProviderType, TenantPreferredModelProvider
MODULE = "core.plugin.plugin_service"
class _FakeSession:
def __init__(self) -> None:
self.execute = Mock()
self.scalars = Mock(return_value=SimpleNamespace(all=Mock(return_value=[])))
def __enter__(self) -> "_FakeSession":
return self
def __exit__(self, exc_type, exc, traceback) -> None:
return None
def begin(self) -> "_FakeSession":
return self
TENANT_ID = "11111111-1111-1111-1111-111111111111"
OTHER_TENANT_ID = "22222222-2222-2222-2222-222222222222"
USER_ID = "33333333-3333-3333-3333-333333333333"
def _build_provider_entity(provider: str = "openai") -> ProviderEntity:
@@ -246,6 +237,26 @@ class TestPluginModelProviderCache:
call([cache_key]),
]
def test_fetch_plugin_model_providers_bypasses_redis_when_cache_disabled(self) -> None:
"""With the cache disabled the daemon is the only source, and Redis is never touched."""
with patch(f"{MODULE}.redis_client") as redis_client, patch(f"{MODULE}.dify_config") as config:
config.PLUGIN_MODEL_PROVIDERS_CACHE_ENABLED = False
client = Mock()
client.fetch_model_providers.return_value = [_build_plugin_model_provider()]
from core.plugin.plugin_service import PluginService
first = PluginService.fetch_plugin_model_providers(tenant_id="tenant-1", client=client)
second = PluginService.fetch_plugin_model_providers(tenant_id="tenant-1", client=client)
assert [provider.provider for provider in first] == ["langgenius/openai/openai"]
assert [provider.provider for provider in second] == ["langgenius/openai/openai"]
assert client.fetch_model_providers.call_count == 2
redis_client.get.assert_not_called()
redis_client.mget.assert_not_called()
redis_client.setex.assert_not_called()
redis_client.lock.assert_not_called()
def test_fetch_plugin_model_providers_refetches_when_cache_read_fails(self) -> None:
"""Redis read failures do not block provider discovery for the tenant."""
with patch(f"{MODULE}.redis_client") as redis_client:
@@ -1146,19 +1157,72 @@ class TestPluginModelProviderCacheInvalidation:
assert result is True
invalidate_cache.assert_called_once_with("tenant-1")
def test_uninstall_existing_plugin_invalidates_cache_after_credential_cleanup(self) -> None:
@pytest.mark.parametrize(
"sqlite_session", [(Provider, ProviderCredential, TenantPreferredModelProvider)], indirect=True
)
def test_uninstall_existing_plugin_invalidates_cache_after_credential_cleanup(
self, sqlite_session: Session
) -> None:
"""Successful uninstall with plugin metadata also invalidates the mutated tenant provider cache."""
plugin_id = "langgenius/openai"
provider_name = f"{plugin_id}/openai"
plugin = SimpleNamespace(
installation_id="installation-1",
plugin_id="langgenius/openai",
plugin_id=plugin_id,
plugin_unique_identifier="langgenius/openai:1.0.0",
)
session = _FakeSession()
credential = ProviderCredential(
tenant_id=TENANT_ID,
provider_name=provider_name,
credential_name="Target credential",
encrypted_config="{}",
user_id=USER_ID,
)
other_credential = ProviderCredential(
tenant_id=OTHER_TENANT_ID,
provider_name=provider_name,
credential_name="Other credential",
encrypted_config="{}",
user_id=USER_ID,
)
sqlite_session.add_all([credential, other_credential])
sqlite_session.flush()
provider = Provider(
tenant_id=TENANT_ID,
provider_name=provider_name,
provider_type=ProviderType.CUSTOM,
credential_id=credential.id,
)
other_provider = Provider(
tenant_id=OTHER_TENANT_ID,
provider_name=provider_name,
provider_type=ProviderType.CUSTOM,
credential_id=other_credential.id,
)
preferred_provider = TenantPreferredModelProvider(
tenant_id=TENANT_ID,
provider_name=provider_name,
preferred_provider_type=ProviderType.CUSTOM,
)
other_preferred_provider = TenantPreferredModelProvider(
tenant_id=OTHER_TENANT_ID,
provider_name=provider_name,
preferred_provider_type=ProviderType.CUSTOM,
)
sqlite_session.add_all([provider, other_provider, preferred_provider, other_preferred_provider])
sqlite_session.commit()
credential_id = credential.id
other_credential_id = other_credential.id
provider_id = provider.id
other_provider_id = other_provider.id
preferred_provider_id = preferred_provider.id
other_preferred_provider_id = other_preferred_provider.id
with (
patch(f"{MODULE}.db", SimpleNamespace(engine=object())),
patch(f"{MODULE}.db", SimpleNamespace(engine=sqlite_session.get_bind())),
patch(f"{MODULE}.dify_config") as mock_config,
patch(f"{MODULE}.PluginInstaller") as installer_cls,
patch(f"{MODULE}.Session", return_value=session),
patch(f"{MODULE}.ProviderCredentialsCache") as credentials_cache,
patch(f"{MODULE}.PluginService.invalidate_plugin_model_providers_cache") as invalidate_cache,
):
mock_config.ENTERPRISE_ENABLED = False
@@ -1168,8 +1232,26 @@ class TestPluginModelProviderCacheInvalidation:
from core.plugin.plugin_service import PluginService
result = PluginService.uninstall("tenant-1", "installation-1")
result = PluginService.uninstall(TENANT_ID, "installation-1")
assert result is True
installer.uninstall.assert_called_once_with("tenant-1", "installation-1")
invalidate_cache.assert_called_once_with("tenant-1")
installer.uninstall.assert_called_once_with(TENANT_ID, "installation-1")
invalidate_cache.assert_called_once_with(TENANT_ID)
credentials_cache.assert_called_once_with(
tenant_id=TENANT_ID,
identity_id=provider_id,
cache_type=ProviderCredentialsCacheType.PROVIDER,
)
credentials_cache.return_value.delete.assert_called_once_with()
sqlite_session.expunge_all()
assert sqlite_session.get(ProviderCredential, credential_id) is None
persisted_provider = sqlite_session.get(Provider, provider_id)
assert persisted_provider is not None
assert persisted_provider.credential_id is None
assert sqlite_session.get(TenantPreferredModelProvider, preferred_provider_id) is None
assert sqlite_session.get(ProviderCredential, other_credential_id) is not None
persisted_other_provider = sqlite_session.get(Provider, other_provider_id)
assert persisted_other_provider is not None
assert persisted_other_provider.credential_id == other_credential_id
assert sqlite_session.get(TenantPreferredModelProvider, other_preferred_provider_id) is not None
@@ -8,6 +8,7 @@ verification, marketplace upgrade flows, and uninstall with credential cleanup.
from __future__ import annotations
from collections.abc import Iterator
from typing import cast
from unittest.mock import MagicMock, patch
from uuid import uuid4
@@ -19,7 +20,6 @@ from sqlalchemy.orm import Session
from core.plugin.entities.plugin import PluginInstallationSource
from core.plugin.entities.plugin_daemon import PluginVerification
from core.plugin.plugin_service import PluginService
from enums.deployment_edition import DeploymentEdition
from models import ProviderType
from models.engine import db
from models.provider import Provider, ProviderCredential, TenantPreferredModelProvider
@@ -27,20 +27,16 @@ from services.errors.plugin import PluginInstallationForbiddenError
from services.feature_service import (
PluginInstallationPermissionModel,
PluginInstallationScope,
SystemFeatureModel,
)
def _make_features(
def _make_permission(
restrict_to_marketplace: bool = False,
scope: PluginInstallationScope = PluginInstallationScope.ALL,
) -> SystemFeatureModel:
return SystemFeatureModel(
deployment_edition=DeploymentEdition.COMMUNITY,
plugin_installation_permission=PluginInstallationPermissionModel(
restrict_to_marketplace_only=restrict_to_marketplace,
plugin_installation_scope=scope,
),
) -> PluginInstallationPermissionModel:
return PluginInstallationPermissionModel(
restrict_to_marketplace_only=restrict_to_marketplace,
plugin_installation_scope=scope,
)
@@ -119,22 +115,31 @@ class TestFetchLatestPluginVersion:
class TestCheckMarketplaceOnlyPermission:
@patch("core.plugin.plugin_service.FeatureService")
def test_raises_when_restricted(self, mock_fs):
mock_fs.get_system_features.return_value = _make_features(restrict_to_marketplace=True)
mock_fs.get_plugin_installation_permission.return_value = _make_permission(restrict_to_marketplace=True)
with pytest.raises(PluginInstallationForbiddenError):
PluginService._check_marketplace_only_permission()
@patch("core.plugin.plugin_service.FeatureService")
def test_passes_when_not_restricted(self, mock_fs):
mock_fs.get_system_features.return_value = _make_features(restrict_to_marketplace=False)
mock_fs.get_plugin_installation_permission.return_value = _make_permission(restrict_to_marketplace=False)
PluginService._check_marketplace_only_permission() # should not raise
@patch("core.plugin.plugin_service.FeatureService")
def test_raises_when_scope_denies_all(self, mock_fs):
mock_fs.get_plugin_installation_permission.return_value = _make_permission(scope=PluginInstallationScope.NONE)
with pytest.raises(PluginInstallationForbiddenError, match="not allowed"):
PluginService._check_marketplace_only_permission()
class TestCheckPluginInstallationScope:
@patch("core.plugin.plugin_service.FeatureService")
def test_official_only_allows_langgenius(self, mock_fs):
mock_fs.get_system_features.return_value = _make_features(scope=PluginInstallationScope.OFFICIAL_ONLY)
mock_fs.get_plugin_installation_permission.return_value = _make_permission(
scope=PluginInstallationScope.OFFICIAL_ONLY
)
verification = MagicMock()
verification.authorized_category = PluginVerification.AuthorizedCategory.Langgenius
@@ -142,14 +147,16 @@ class TestCheckPluginInstallationScope:
@patch("core.plugin.plugin_service.FeatureService")
def test_official_only_rejects_third_party(self, mock_fs):
mock_fs.get_system_features.return_value = _make_features(scope=PluginInstallationScope.OFFICIAL_ONLY)
mock_fs.get_plugin_installation_permission.return_value = _make_permission(
scope=PluginInstallationScope.OFFICIAL_ONLY
)
with pytest.raises(PluginInstallationForbiddenError):
PluginService._check_plugin_installation_scope(None)
@patch("core.plugin.plugin_service.FeatureService")
def test_official_and_partners_allows_partner(self, mock_fs):
mock_fs.get_system_features.return_value = _make_features(
mock_fs.get_plugin_installation_permission.return_value = _make_permission(
scope=PluginInstallationScope.OFFICIAL_AND_SPECIFIC_PARTNERS
)
verification = MagicMock()
@@ -159,7 +166,7 @@ class TestCheckPluginInstallationScope:
@patch("core.plugin.plugin_service.FeatureService")
def test_official_and_partners_rejects_none(self, mock_fs):
mock_fs.get_system_features.return_value = _make_features(
mock_fs.get_plugin_installation_permission.return_value = _make_permission(
scope=PluginInstallationScope.OFFICIAL_AND_SPECIFIC_PARTNERS
)
@@ -168,7 +175,7 @@ class TestCheckPluginInstallationScope:
@patch("core.plugin.plugin_service.FeatureService")
def test_none_scope_always_raises(self, mock_fs):
mock_fs.get_system_features.return_value = _make_features(scope=PluginInstallationScope.NONE)
mock_fs.get_plugin_installation_permission.return_value = _make_permission(scope=PluginInstallationScope.NONE)
verification = MagicMock()
verification.authorized_category = PluginVerification.AuthorizedCategory.Langgenius
@@ -177,10 +184,19 @@ class TestCheckPluginInstallationScope:
@patch("core.plugin.plugin_service.FeatureService")
def test_all_scope_passes_any(self, mock_fs):
mock_fs.get_system_features.return_value = _make_features(scope=PluginInstallationScope.ALL)
mock_fs.get_plugin_installation_permission.return_value = _make_permission(scope=PluginInstallationScope.ALL)
PluginService._check_plugin_installation_scope(None) # should not raise
@patch("core.plugin.plugin_service.FeatureService")
def test_unknown_scope_always_raises(self, mock_fs):
permission = _make_permission()
permission.plugin_installation_scope = cast(PluginInstallationScope, "unknown-scope")
mock_fs.get_plugin_installation_permission.return_value = permission
with pytest.raises(PluginInstallationForbiddenError, match="policy is invalid"):
PluginService._check_plugin_installation_scope(None)
class TestGetPluginIconUrl:
@patch("core.plugin.plugin_service.dify_config")
@@ -248,7 +264,7 @@ class TestUpgradePluginWithMarketplace:
@patch("core.plugin.plugin_service.dify_config")
def test_skips_download_when_already_installed(self, mock_config, mock_installer_cls, mock_fs, mock_marketplace):
mock_config.MARKETPLACE_ENABLED = True
mock_fs.get_system_features.return_value = _make_features()
mock_fs.get_plugin_installation_permission.return_value = _make_permission()
installer = mock_installer_cls.return_value
installer.fetch_plugin_manifest.return_value = MagicMock()
installer.upgrade_plugin.return_value = MagicMock()
@@ -264,7 +280,7 @@ class TestUpgradePluginWithMarketplace:
@patch("core.plugin.plugin_service.dify_config")
def test_downloads_when_not_installed(self, mock_config, mock_installer_cls, mock_fs, mock_download):
mock_config.MARKETPLACE_ENABLED = True
mock_fs.get_system_features.return_value = _make_features()
mock_fs.get_plugin_installation_permission.return_value = _make_permission()
installer = mock_installer_cls.return_value
installer.fetch_plugin_manifest.side_effect = RuntimeError("not found")
mock_download.return_value = b"pkg-bytes"
@@ -283,7 +299,7 @@ class TestUpgradePluginWithGithub:
@patch("core.plugin.plugin_service.FeatureService")
@patch("core.plugin.plugin_service.PluginInstaller")
def test_checks_marketplace_permission_and_delegates(self, mock_installer_cls: MagicMock, mock_fs: MagicMock):
mock_fs.get_system_features.return_value = _make_features()
mock_fs.get_plugin_installation_permission.return_value = _make_permission()
installer = mock_installer_cls.return_value
installer.upgrade_plugin.return_value = MagicMock()
@@ -298,7 +314,7 @@ class TestUploadPkg:
@patch("core.plugin.plugin_service.FeatureService")
@patch("core.plugin.plugin_service.PluginInstaller")
def test_runs_permission_and_scope_checks(self, mock_installer_cls: MagicMock, mock_fs: MagicMock):
mock_fs.get_system_features.return_value = _make_features()
mock_fs.get_plugin_installation_permission.return_value = _make_permission()
upload_resp = MagicMock()
upload_resp.verification = None
mock_installer_cls.return_value.upload_pkg.return_value = upload_resp
@@ -322,7 +338,7 @@ class TestInstallFromMarketplacePkg:
@patch("core.plugin.plugin_service.dify_config")
def test_downloads_when_not_cached(self, mock_config, mock_installer_cls, mock_fs, mock_download):
mock_config.MARKETPLACE_ENABLED = True
mock_fs.get_system_features.return_value = _make_features()
mock_fs.get_plugin_installation_permission.return_value = _make_permission()
installer = mock_installer_cls.return_value
installer.fetch_plugin_manifest.side_effect = RuntimeError("not found")
mock_download.return_value = b"pkg"
@@ -344,7 +360,7 @@ class TestInstallFromMarketplacePkg:
@patch("core.plugin.plugin_service.dify_config")
def test_uses_cached_when_already_downloaded(self, mock_config, mock_installer_cls: MagicMock, mock_fs: MagicMock):
mock_config.MARKETPLACE_ENABLED = True
mock_fs.get_system_features.return_value = _make_features()
mock_fs.get_plugin_installation_permission.return_value = _make_permission()
installer = mock_installer_cls.return_value
installer.fetch_plugin_manifest.return_value = MagicMock()
decode_resp = MagicMock()
@@ -1,5 +1,4 @@
import json
from collections.abc import Iterator
from datetime import datetime, timedelta
from unittest.mock import MagicMock, patch
from uuid import UUID
@@ -11,7 +10,6 @@ from sqlalchemy.orm import Session
from configs import dify_config
from models.account import (
Account,
AccountIntegrate,
AccountStatus,
Tenant,
TenantAccountJoin,
@@ -114,22 +112,6 @@ class TestAccountService:
- Error conditions and edge cases
"""
@pytest.fixture
def sqlite_session(self, sqlite_engine) -> Iterator[Session]:
"""SQLite session with the account/workspace tables these service tests touch."""
tables = [
model.metadata.tables[model.__tablename__]
for model in (
Account,
Tenant,
TenantAccountJoin,
TenantPluginAutoUpgradeStrategy,
)
]
Account.metadata.create_all(sqlite_engine, tables=tables)
with Session(sqlite_engine, expire_on_commit=False) as session:
yield session
@pytest.fixture
def mock_password_dependencies(self):
"""Mock setup for password-related functions."""
@@ -1264,24 +1246,6 @@ class TestRegisterService:
- Error conditions and edge cases
"""
@pytest.fixture
def sqlite_session(self, sqlite_engine) -> Iterator[Session]:
"""SQLite session with the account/workspace tables registration flows touch."""
tables = [
model.metadata.tables[model.__tablename__]
for model in (
Account,
AccountIntegrate,
Tenant,
TenantAccountJoin,
TenantPluginAutoUpgradeStrategy,
DifySetup,
)
]
Account.metadata.create_all(sqlite_engine, tables=tables)
with Session(sqlite_engine, expire_on_commit=False) as session:
yield session
@pytest.fixture
def mock_redis_dependencies(self):
"""Mock setup for Redis-related functions."""
@@ -1,19 +1,23 @@
from __future__ import annotations
from collections.abc import Callable
from datetime import datetime
from types import SimpleNamespace
from typing import cast
from unittest.mock import MagicMock, patch
from uuid import uuid4
import pytest
from sqlalchemy import event
from sqlalchemy.exc import IntegrityError
from sqlalchemy.orm import Session
from graphon.model_runtime.entities.model_entities import ModelType
from models import Account
from models.model import App, AppMode, AppModelConfig
from models.model import App, AppMode, AppModelConfig, IconType
from models.workflow import Workflow
from services.agent.errors import AgentNameConflictError
from services.app_service import AppService, CreateAppParams
from services.app_service import AppListParams, AppService, CreateAppParams
class TestCreateAppTransactionBoundary:
@@ -236,6 +240,92 @@ class TestOpenapiVisibilityHelpers:
mock_session.execute.assert_called_once()
@pytest.mark.parametrize("sqlite_session", [(Account, App, AppModelConfig)], indirect=True)
def test_get_recent_apps_uses_one_tenant_scoped_projection_query(sqlite_session: Session) -> None:
tenant_id = str(uuid4())
other_tenant_id = str(uuid4())
account = Account(name="Recent Apps Author", email="recent-apps@example.com")
sqlite_session.add(account)
sqlite_session.flush()
def create_app(*, name: str, tenant_id: str, updated_at: datetime, mode: AppMode = AppMode.CHAT) -> App:
app = App()
app.id = str(uuid4())
app.tenant_id = tenant_id
app.name = name
app.description = ""
app.mode = mode
app.icon_type = IconType.EMOJI
app.icon = "🚀"
app.icon_background = "#FFFFFF"
app.enable_site = False
app.enable_api = False
app.created_by = account.id
app.maintainer = account.id
app.created_at = updated_at
app.updated_at = updated_at
app.use_icon_as_answer_icon = False
return app
newest = create_app(name="Newest", tenant_id=tenant_id, updated_at=datetime(2026, 7, 3))
legacy_agent = AppModelConfig(app_id=newest.id)
legacy_agent.agent_mode = '{"enabled": true, "strategy": "react"}'
newest.app_model_config_id = legacy_agent.id
second = create_app(
name="Second",
tenant_id=tenant_id,
updated_at=datetime(2026, 7, 2),
mode=AppMode.WORKFLOW,
)
second.icon_type = None
second.icon = None
second.icon_background = None
second.created_by = None
second.maintainer = None
channel = create_app(
name="Channel",
tenant_id=tenant_id,
updated_at=datetime(2026, 7, 5),
mode=AppMode.CHANNEL,
)
rag_pipeline = create_app(
name="RAG Pipeline",
tenant_id=tenant_id,
updated_at=datetime(2026, 7, 4),
mode=AppMode.RAG_PIPELINE,
)
oldest = create_app(name="Oldest", tenant_id=tenant_id, updated_at=datetime(2026, 7, 1))
foreign = create_app(name="Foreign", tenant_id=other_tenant_id, updated_at=datetime(2026, 7, 4))
sqlite_session.add_all([newest, legacy_agent, second, channel, rag_pipeline, oldest, foreign])
sqlite_session.commit()
statements: list[str] = []
bind = sqlite_session.get_bind()
def record_sql(_conn, _cursor, statement, _parameters, _context, _executemany) -> None:
statements.append(statement)
event.listen(bind, "before_cursor_execute", record_sql)
try:
recent_apps = AppService().get_recent_apps(
account.id,
tenant_id,
AppListParams(limit=2),
sqlite_session,
)
finally:
event.remove(bind, "before_cursor_execute", record_sql)
assert [(app.name, app.mode, app.icon_type, app.author_name, app.maintainer) for app in recent_apps] == [
("Newest", AppMode.CHAT, IconType.EMOJI, "Recent Apps Author", account.id),
("Second", AppMode.WORKFLOW, None, None, None),
]
select_statements = [statement for statement in statements if statement.lstrip().upper().startswith("SELECT")]
assert len(select_statements) == 1
assert "count(" not in select_statements[0].lower()
assert "app_model_configs" not in select_statements[0].lower()
class TestAppMeta:
def test_loads_workflow_with_caller_session(self):
session = MagicMock()
@@ -0,0 +1,92 @@
import logging
import pytest
from enums.deployment_edition import DeploymentEdition
from services import feature_service as feature_service_module
from services.feature_service import FeatureService, PluginInstallationScope, SystemFeatureModel
def test_get_plugin_installation_permission_defaults_to_all_for_non_enterprise(
monkeypatch: pytest.MonkeyPatch,
) -> None:
monkeypatch.setattr(feature_service_module.dify_config, "ENTERPRISE_ENABLED", False)
permission = FeatureService.get_plugin_installation_permission()
assert permission.plugin_installation_scope is PluginInstallationScope.ALL
assert permission.restrict_to_marketplace_only is False
def test_get_plugin_installation_permission_parses_enterprise_policy(
monkeypatch: pytest.MonkeyPatch,
) -> None:
monkeypatch.setattr(feature_service_module.dify_config, "ENTERPRISE_ENABLED", True)
monkeypatch.setattr(
feature_service_module.EnterpriseService,
"get_info",
staticmethod(
lambda: {
"PluginInstallationPermission": {
"pluginInstallationScope": "official_only",
"restrictToMarketplaceOnly": True,
}
}
),
)
permission = FeatureService.get_plugin_installation_permission()
assert permission.plugin_installation_scope is PluginInstallationScope.OFFICIAL_ONLY
assert permission.restrict_to_marketplace_only is True
@pytest.mark.parametrize(
"invalid_permission",
[
{
"pluginInstallationScope": "unknown-scope",
"restrictToMarketplaceOnly": False,
},
{
"pluginInstallationScope": "all",
"restrictToMarketplaceOnly": "false",
},
],
ids=["unknown_scope", "non_boolean_marketplace_restriction"],
)
def test_invalid_enterprise_policy_denies_all_plugin_installations(
caplog: pytest.LogCaptureFixture,
invalid_permission: dict[str, object],
) -> None:
with caplog.at_level(logging.ERROR, logger="services.feature_service"):
permission = FeatureService._resolve_plugin_installation_permission(
{"PluginInstallationPermission": invalid_permission}
)
assert permission.plugin_installation_scope is PluginInstallationScope.NONE
assert permission.restrict_to_marketplace_only is True
assert "denying all plugin installations" in caplog.text
def test_system_features_exposes_only_validated_plugin_installation_policy(
monkeypatch: pytest.MonkeyPatch,
) -> None:
monkeypatch.setattr(
feature_service_module.EnterpriseService,
"get_info",
staticmethod(
lambda: {
"PluginInstallationPermission": {
"pluginInstallationScope": "unknown-scope",
"restrictToMarketplaceOnly": False,
}
}
),
)
features = SystemFeatureModel(deployment_edition=DeploymentEdition.ENTERPRISE)
FeatureService._fulfill_params_from_enterprise(features)
assert features.plugin_installation_permission.plugin_installation_scope is PluginInstallationScope.NONE
assert features.plugin_installation_permission.restrict_to_marketplace_only is True
+111 -124
View File
@@ -1,6 +1,8 @@
import base64
import hashlib
import os
from collections.abc import Iterator
from datetime import UTC, datetime
from unittest.mock import MagicMock, patch
import pytest
@@ -9,6 +11,8 @@ from sqlalchemy.orm import Session, sessionmaker
from werkzeug.exceptions import NotFound
from configs import dify_config
from extensions.storage.storage_type import StorageType
from models.base import TypeBase
from models.enums import CreatorUserRole
from models.model import Account, EndUser, UploadFile
from services.errors.file import BlockedFileExtensionError, FileTooLargeError, UnsupportedFileTypeError
@@ -17,31 +21,54 @@ from services.file_service import FileService
class TestFileService:
@pytest.fixture
def mock_db_session(self):
session = MagicMock(spec=Session)
# Mock context manager behavior
session.__enter__.return_value = session
return session
def sqlite_session_maker(self, sqlite_engine: Engine) -> sessionmaker[Session]:
TypeBase.metadata.create_all(sqlite_engine, tables=[TypeBase.metadata.tables[UploadFile.__tablename__]])
return sessionmaker(bind=sqlite_engine, expire_on_commit=False)
@pytest.fixture
def mock_session_maker(self, mock_db_session):
maker = MagicMock(spec=sessionmaker)
maker.return_value = mock_db_session
return maker
def db_session(self, sqlite_session_maker: sessionmaker[Session]) -> Iterator[Session]:
with sqlite_session_maker() as session:
yield session
@pytest.fixture
def file_service(self, mock_session_maker):
return FileService(session_factory=mock_session_maker)
def file_service(self, sqlite_session_maker: sessionmaker[Session]) -> FileService:
return FileService(session_factory=sqlite_session_maker)
def test_init_with_engine(self):
engine = MagicMock(spec=Engine)
service = FileService(session_factory=engine)
@staticmethod
def _persist_upload_file(
session: Session,
*,
file_id: str = "file_id",
tenant_id: str = "tenant_id",
extension: str = "txt",
mime_type: str = "text/plain",
key: str = "key",
) -> UploadFile:
upload_file = UploadFile(
tenant_id=tenant_id,
storage_type=StorageType.LOCAL,
key=key,
name=f"test.{extension}",
size=10,
extension=extension,
mime_type=mime_type,
created_by_role=CreatorUserRole.ACCOUNT,
created_by="user_id",
created_at=datetime(2024, 1, 1, tzinfo=UTC),
used=False,
)
upload_file.id = file_id
session.add(upload_file)
session.commit()
return upload_file
def test_init_with_engine(self, sqlite_engine: Engine):
service = FileService(session_factory=sqlite_engine)
assert isinstance(service._session_maker, sessionmaker)
def test_init_with_sessionmaker(self):
maker = MagicMock(spec=sessionmaker)
service = FileService(session_factory=maker)
assert service._session_maker == maker
def test_init_with_sessionmaker(self, sqlite_session_maker: sessionmaker[Session]):
service = FileService(session_factory=sqlite_session_maker)
assert service._session_maker == sqlite_session_maker
def test_init_invalid_factory(self):
with pytest.raises(AssertionError, match="must be a sessionmaker or an Engine."):
@@ -52,11 +79,11 @@ class TestFileService:
@patch("services.file_service.extract_tenant_id")
@patch("services.file_service.file_helpers.get_signed_file_url")
def test_upload_file_success(
self, mock_get_url, mock_tenant_id, mock_now, mock_storage, file_service: FileService, mock_db_session
self, mock_get_url, mock_tenant_id, mock_now, mock_storage, file_service: FileService, db_session: Session
):
# Setup
mock_tenant_id.return_value = "tenant_id"
mock_now.return_value = "2024-01-01"
mock_now.return_value = datetime(2024, 1, 1, tzinfo=UTC)
mock_get_url.return_value = "http://signed-url"
user = MagicMock(spec=Account)
@@ -81,8 +108,9 @@ class TestFileService:
assert result.source_url == "http://signed-url"
mock_storage.save.assert_called_once()
mock_db_session.add.assert_called_once_with(result)
mock_db_session.commit.assert_called_once()
persisted = db_session.get(UploadFile, result.id)
assert persisted is not None
assert persisted.hash == result.hash
def test_upload_file_uses_explicit_resource_tenant(self, file_service: FileService):
user = MagicMock(spec=Account)
@@ -109,7 +137,7 @@ class TestFileService:
with pytest.raises(ValueError, match="Filename contains invalid characters"):
file_service.upload_file(filename="invalid/file.txt", content=b"", mimetype="text/plain", user=MagicMock())
def test_upload_file_long_filename(self, file_service: FileService, mock_db_session):
def test_upload_file_long_filename(self, file_service: FileService, db_session: Session):
# Setup
long_name = "a" * 210 + ".txt"
user = MagicMock(spec=Account)
@@ -124,6 +152,7 @@ class TestFileService:
result = file_service.upload_file(filename=long_name, content=b"test", mimetype="text/plain", user=user)
assert len(result.name) <= 205 # 200 + . + extension
assert result.name.endswith(".txt")
assert db_session.get(UploadFile, result.id) is not None
def test_upload_file_blocked_extension(self, file_service):
with patch.object(dify_config, "inner_UPLOAD_FILE_EXTENSION_BLACKLIST", "exe"):
@@ -145,7 +174,7 @@ class TestFileService:
with pytest.raises(FileTooLargeError):
file_service.upload_file(filename="test.jpg", content=content, mimetype="image/jpeg", user=MagicMock())
def test_upload_file_end_user(self, file_service: FileService, mock_db_session):
def test_upload_file_end_user(self, file_service: FileService, db_session: Session):
user = MagicMock(spec=EndUser)
user.id = "end_user_id"
@@ -157,6 +186,7 @@ class TestFileService:
mock_tenant.return_value = "tenant"
result = file_service.upload_file(filename="test.txt", content=b"test", mimetype="text/plain", user=user)
assert result.created_by_role == CreatorUserRole.END_USER
assert db_session.get(UploadFile, result.id) is not None
def test_is_file_size_within_limit(self):
with (
@@ -181,12 +211,8 @@ class TestFileService:
assert FileService.is_file_size_within_limit(extension="txt", file_size=5 * 1024 * 1024) is True
assert FileService.is_file_size_within_limit(extension="pdf", file_size=6 * 1024 * 1024) is False
def test_get_file_base64_success(self, file_service: FileService, mock_db_session):
# Setup
upload_file = MagicMock(spec=UploadFile)
upload_file.id = "file_id"
upload_file.key = "test_key"
mock_db_session.scalar.return_value = upload_file
def test_get_file_base64_success(self, file_service: FileService, db_session: Session):
self._persist_upload_file(db_session, key="test_key")
with patch("services.file_service.storage") as mock_storage:
mock_storage.load_once.return_value = b"test content"
@@ -198,16 +224,17 @@ class TestFileService:
assert result == base64.b64encode(b"test content").decode()
mock_storage.load_once.assert_called_once_with("test_key")
def test_get_file_base64_not_found(self, file_service: FileService, mock_db_session):
mock_db_session.scalar.return_value = None
def test_get_file_base64_not_found(self, file_service: FileService):
with pytest.raises(NotFound, match="File not found"):
file_service.get_file_base64("non_existent")
def test_get_file_presigned_url_success(self, file_service: FileService, mock_db_session):
upload_file = MagicMock(spec=UploadFile)
upload_file.key = "upload_files/tenant_id/icon.png"
upload_file.mime_type = "image/png"
mock_db_session.scalar.return_value = upload_file
def test_get_file_presigned_url_success(self, file_service: FileService, db_session: Session):
self._persist_upload_file(
db_session,
extension="png",
mime_type="image/png",
key="upload_files/tenant_id/icon.png",
)
with (
patch.object(dify_config, "FILES_ACCESS_TIMEOUT", 300),
@@ -224,13 +251,11 @@ class TestFileService:
content_type="image/png",
)
def test_get_file_presigned_url_not_found(self, file_service: FileService, mock_db_session):
mock_db_session.scalar.return_value = None
def test_get_file_presigned_url_not_found(self, file_service: FileService):
with pytest.raises(NotFound, match="File not found"):
file_service.get_file_presigned_url(file_id="file_id", tenant_id="tenant_id")
def test_upload_text_success(self, file_service: FileService, mock_db_session):
def test_upload_text_success(self, file_service: FileService, db_session: Session):
# Setup
text = "sample text"
text_name = "test.txt"
@@ -249,21 +274,17 @@ class TestFileService:
assert result.used is True
assert result.extension == "txt"
mock_storage.save.assert_called_once()
mock_db_session.add.assert_called_once()
mock_db_session.commit.assert_called_once()
assert db_session.get(UploadFile, result.id) is not None
def test_upload_text_long_name(self, file_service: FileService, mock_db_session):
def test_upload_text_long_name(self, file_service: FileService, db_session: Session):
long_name = "a" * 210
with patch("services.file_service.storage"):
result = file_service.upload_text("text", long_name, "user", "tenant")
assert len(result.name) == 200
assert db_session.get(UploadFile, result.id) is not None
def test_get_file_preview_success(self, file_service: FileService, mock_db_session):
# Setup
upload_file = MagicMock(spec=UploadFile)
upload_file.id = "file_id"
upload_file.extension = "pdf"
mock_db_session.scalar.return_value = upload_file
def test_get_file_preview_success(self, file_service: FileService, db_session: Session):
self._persist_upload_file(db_session, extension="pdf", mime_type="application/pdf")
with patch("services.file_service.ExtractProcessor.load_from_upload_file") as mock_extract:
mock_extract.return_value = "Extracted text content"
@@ -274,27 +295,17 @@ class TestFileService:
# Assert
assert result == "Extracted text content"
def test_get_file_preview_not_found(self, file_service: FileService, mock_db_session):
mock_db_session.scalar.return_value = None
def test_get_file_preview_not_found(self, file_service: FileService):
with pytest.raises(NotFound, match="File not found"):
file_service.get_file_preview("non_existent", "tenant_id")
def test_get_file_preview_unsupported_type(self, file_service: FileService, mock_db_session):
upload_file = MagicMock(spec=UploadFile)
upload_file.id = "file_id"
upload_file.extension = "exe"
mock_db_session.scalar.return_value = upload_file
def test_get_file_preview_unsupported_type(self, file_service: FileService, db_session: Session):
self._persist_upload_file(db_session, extension="exe", mime_type="application/octet-stream")
with pytest.raises(UnsupportedFileTypeError):
file_service.get_file_preview("file_id", "tenant_id")
def test_get_image_preview_success(self, file_service: FileService, mock_db_session):
# Setup
upload_file = MagicMock(spec=UploadFile)
upload_file.id = "file_id"
upload_file.extension = "jpg"
upload_file.mime_type = "image/jpeg"
upload_file.key = "key"
mock_db_session.scalar.return_value = upload_file
def test_get_image_preview_success(self, file_service: FileService, db_session: Session):
self._persist_upload_file(db_session, extension="jpg", mime_type="image/jpeg")
with (
patch("services.file_service.file_helpers.verify_image_signature") as mock_verify,
@@ -316,28 +327,21 @@ class TestFileService:
with pytest.raises(NotFound, match="File not found or signature is invalid"):
file_service.get_image_preview("file_id", "ts", "nonce", "sign")
def test_get_image_preview_not_found(self, file_service: FileService, mock_db_session):
mock_db_session.scalar.return_value = None
def test_get_image_preview_not_found(self, file_service: FileService):
with patch("services.file_service.file_helpers.verify_image_signature") as mock_verify:
mock_verify.return_value = True
with pytest.raises(NotFound, match="File not found or signature is invalid"):
file_service.get_image_preview("file_id", "ts", "nonce", "sign")
def test_get_image_preview_unsupported_type(self, file_service: FileService, mock_db_session):
upload_file = MagicMock(spec=UploadFile)
upload_file.id = "file_id"
upload_file.extension = "txt"
mock_db_session.scalar.return_value = upload_file
def test_get_image_preview_unsupported_type(self, file_service: FileService, db_session: Session):
self._persist_upload_file(db_session)
with patch("services.file_service.file_helpers.verify_image_signature") as mock_verify:
mock_verify.return_value = True
with pytest.raises(UnsupportedFileTypeError):
file_service.get_image_preview("file_id", "ts", "nonce", "sign")
def test_get_file_generator_by_file_id_success(self, file_service: FileService, mock_db_session):
upload_file = MagicMock(spec=UploadFile)
upload_file.id = "file_id"
upload_file.key = "key"
mock_db_session.scalar.return_value = upload_file
def test_get_file_generator_by_file_id_success(self, file_service: FileService, db_session: Session):
upload_file = self._persist_upload_file(db_session)
with (
patch("services.file_service.file_helpers.verify_file_signature") as mock_verify,
@@ -348,7 +352,8 @@ class TestFileService:
gen, file = file_service.get_file_generator_by_file_id("file_id", "ts", "nonce", "sign")
assert list(gen) == [b"chunk"]
assert file == upload_file
assert file.id == upload_file.id
assert file.key == upload_file.key
def test_get_file_generator_by_file_id_invalid_sig(self, file_service):
with patch("services.file_service.file_helpers.verify_file_signature") as mock_verify:
@@ -356,20 +361,14 @@ class TestFileService:
with pytest.raises(NotFound, match="File not found or signature is invalid"):
file_service.get_file_generator_by_file_id("file_id", "ts", "nonce", "sign")
def test_get_file_generator_by_file_id_not_found(self, file_service: FileService, mock_db_session):
mock_db_session.scalar.return_value = None
def test_get_file_generator_by_file_id_not_found(self, file_service: FileService):
with patch("services.file_service.file_helpers.verify_file_signature") as mock_verify:
mock_verify.return_value = True
with pytest.raises(NotFound, match="File not found or signature is invalid"):
file_service.get_file_generator_by_file_id("file_id", "ts", "nonce", "sign")
def test_get_public_image_preview_success(self, file_service: FileService, mock_db_session):
upload_file = MagicMock(spec=UploadFile)
upload_file.id = "file_id"
upload_file.extension = "png"
upload_file.mime_type = "image/png"
upload_file.key = "key"
mock_db_session.scalar.return_value = upload_file
def test_get_public_image_preview_success(self, file_service: FileService, db_session: Session):
self._persist_upload_file(db_session, extension="png", mime_type="image/png")
with patch("services.file_service.storage") as mock_storage:
mock_storage.load.return_value = b"image content"
@@ -377,66 +376,56 @@ class TestFileService:
assert gen == b"image content"
assert mime == "image/png"
def test_get_public_image_preview_not_found(self, file_service: FileService, mock_db_session):
mock_db_session.scalar.return_value = None
def test_get_public_image_preview_not_found(self, file_service: FileService):
with pytest.raises(NotFound, match="File not found or signature is invalid"):
file_service.get_public_image_preview("file_id")
def test_get_public_image_preview_unsupported_type(self, file_service: FileService, mock_db_session):
upload_file = MagicMock(spec=UploadFile)
upload_file.id = "file_id"
upload_file.extension = "txt"
mock_db_session.scalar.return_value = upload_file
def test_get_public_image_preview_unsupported_type(self, file_service: FileService, db_session: Session):
self._persist_upload_file(db_session)
with pytest.raises(UnsupportedFileTypeError):
file_service.get_public_image_preview("file_id")
def test_get_file_content_success(self, file_service: FileService, mock_db_session):
upload_file = MagicMock(spec=UploadFile)
upload_file.id = "file_id"
upload_file.key = "key"
mock_db_session.scalar.return_value = upload_file
def test_get_file_content_success(self, file_service: FileService, db_session: Session):
self._persist_upload_file(db_session)
with patch("services.file_service.storage") as mock_storage:
mock_storage.load.return_value = b"hello world"
result = file_service.get_file_content("file_id")
assert result == "hello world"
def test_get_file_content_not_found(self, file_service: FileService, mock_db_session):
mock_db_session.scalar.return_value = None
def test_get_file_content_not_found(self, file_service: FileService):
with pytest.raises(NotFound, match="File not found"):
file_service.get_file_content("file_id")
def test_delete_file_success(self, file_service: FileService, mock_db_session):
upload_file = MagicMock(spec=UploadFile)
upload_file.id = "file_id"
upload_file.key = "key"
# For session.scalar(select(...))
mock_db_session.scalar.return_value = upload_file
def test_delete_file_success(self, file_service: FileService, db_session: Session):
self._persist_upload_file(db_session)
with patch("services.file_service.storage") as mock_storage:
file_service.delete_file("file_id")
mock_storage.delete.assert_called_once_with("key")
mock_db_session.delete.assert_called_once_with(upload_file)
db_session.expire_all()
assert db_session.get(UploadFile, "file_id") is None
def test_delete_file_not_found(self, file_service: FileService, mock_db_session):
mock_db_session.scalar.return_value = None
def test_delete_file_not_found(self, file_service: FileService):
file_service.delete_file("file_id")
# Should return without doing anything
def test_get_upload_files_by_ids_empty(self):
session = MagicMock()
result = FileService.get_upload_files_by_ids("tenant_id", [], session=session)
def test_get_upload_files_by_ids_empty(self, db_session: Session):
result = FileService.get_upload_files_by_ids("tenant_id", [], session=db_session)
assert result == {}
def test_get_upload_files_by_ids(self):
upload_file = MagicMock(spec=UploadFile)
upload_file.id = "550e8400-e29b-41d4-a716-446655440000"
upload_file.tenant_id = "tenant_id"
session = MagicMock()
session.scalars().all.return_value = [upload_file]
def test_get_upload_files_by_ids(self, db_session: Session):
upload_file = self._persist_upload_file(db_session, file_id="550e8400-e29b-41d4-a716-446655440000")
self._persist_upload_file(
db_session,
file_id="550e8400-e29b-41d4-a716-446655440001",
tenant_id="other-tenant",
)
result = FileService.get_upload_files_by_ids(
"tenant_id", ["550e8400-e29b-41d4-a716-446655440000"], session=session
"tenant_id",
["550e8400-e29b-41d4-a716-446655440000", "550e8400-e29b-41d4-a716-446655440001"],
session=db_session,
)
assert result["550e8400-e29b-41d4-a716-446655440000"] == upload_file
@@ -453,10 +442,8 @@ class TestFileService:
used.add("a (1).txt")
assert FileService._dedupe_zip_entry_name("a.txt", used) == "a (2).txt"
def test_build_upload_files_zip_tempfile(self):
upload_file = MagicMock(spec=UploadFile)
upload_file.name = "test.txt"
upload_file.key = "key"
def test_build_upload_files_zip_tempfile(self, db_session: Session):
upload_file = self._persist_upload_file(db_session)
with (
patch("services.file_service.storage") as mock_storage,
@@ -1,6 +1,5 @@
import dataclasses
import logging
from collections.abc import Iterator
from datetime import datetime, timedelta
from unittest.mock import MagicMock
@@ -42,14 +41,13 @@ from services.human_input_service import (
@pytest.fixture
def sqlite_session_factory(sqlite_engine: Engine) -> Iterator[tuple[sessionmaker[Session], Session]]:
factory = sessionmaker(bind=sqlite_engine, expire_on_commit=False)
with factory() as session:
yield factory, session
def unbound_session_factory() -> sessionmaker[Session]:
"""Supply the required constructor dependency without enabling database access."""
return sessionmaker()
def _persist_app(sqlite_session: Session, mode: AppMode) -> App:
app = App(
def _make_app(mode: AppMode) -> App:
return App(
id="app-id",
tenant_id="tenant-id",
name="Test App",
@@ -60,9 +58,6 @@ def _persist_app(sqlite_session: Session, mode: AppMode) -> App:
enable_api=True,
max_active_requests=0,
)
sqlite_session.add(app)
sqlite_session.commit()
return app
@pytest.fixture
@@ -97,14 +92,11 @@ def sample_form_record():
)
@pytest.mark.parametrize("sqlite_session", [(App,)], indirect=True)
def test_enqueue_resume_dispatches_task_for_workflow(
mocker: MockerFixture,
sqlite_session_factory,
sqlite_session: Session,
sqlite_session_factory: sessionmaker[Session],
):
session_factory, _ = sqlite_session_factory
service = HumanInputService(session_factory)
service = HumanInputService(sqlite_session_factory)
workflow_run = MagicMock()
workflow_run.app_id = "app-id"
@@ -116,7 +108,8 @@ def test_enqueue_resume_dispatches_task_for_workflow(
return_value=workflow_run_repo,
)
_persist_app(sqlite_session, AppMode.WORKFLOW)
with sqlite_session_factory.begin() as arrange_session:
arrange_session.add(_make_app(AppMode.WORKFLOW))
resume_task = mocker.patch("services.human_input_service.resume_app_execution")
@@ -128,10 +121,9 @@ def test_enqueue_resume_dispatches_task_for_workflow(
def test_ensure_form_active_respects_global_timeout(
monkeypatch, sample_form_record: HumanInputFormRecord, sqlite_session_factory
monkeypatch, sample_form_record: HumanInputFormRecord, unbound_session_factory
):
session_factory, _ = sqlite_session_factory
service = HumanInputService(session_factory)
service = HumanInputService(unbound_session_factory)
expired_record = dataclasses.replace(
sample_form_record,
created_at=naive_utc_now() - timedelta(hours=2),
@@ -143,14 +135,11 @@ def test_ensure_form_active_respects_global_timeout(
service.ensure_form_active(Form(expired_record))
@pytest.mark.parametrize("sqlite_session", [(App,)], indirect=True)
def test_enqueue_resume_dispatches_task_for_advanced_chat(
mocker: MockerFixture,
sqlite_session_factory,
sqlite_session: Session,
sqlite_session_factory: sessionmaker[Session],
):
session_factory, _ = sqlite_session_factory
service = HumanInputService(session_factory)
service = HumanInputService(sqlite_session_factory)
workflow_run = MagicMock()
workflow_run.app_id = "app-id"
@@ -162,7 +151,8 @@ def test_enqueue_resume_dispatches_task_for_advanced_chat(
return_value=workflow_run_repo,
)
_persist_app(sqlite_session, AppMode.ADVANCED_CHAT)
with sqlite_session_factory.begin() as arrange_session:
arrange_session.add(_make_app(AppMode.ADVANCED_CHAT))
resume_task = mocker.patch("services.human_input_service.resume_app_execution")
@@ -173,14 +163,11 @@ def test_enqueue_resume_dispatches_task_for_advanced_chat(
assert call_kwargs["kwargs"]["payload"]["workflow_run_id"] == "workflow-run-id"
@pytest.mark.parametrize("sqlite_session", [(App,)], indirect=True)
def test_enqueue_resume_skips_unsupported_app_mode(
mocker: MockerFixture,
sqlite_session_factory,
sqlite_session: Session,
sqlite_session_factory: sessionmaker[Session],
):
session_factory, _ = sqlite_session_factory
service = HumanInputService(session_factory)
service = HumanInputService(sqlite_session_factory)
workflow_run = MagicMock()
workflow_run.app_id = "app-id"
@@ -192,7 +179,8 @@ def test_enqueue_resume_skips_unsupported_app_mode(
return_value=workflow_run_repo,
)
_persist_app(sqlite_session, AppMode.COMPLETION)
with sqlite_session_factory.begin() as arrange_session:
arrange_session.add(_make_app(AppMode.COMPLETION))
resume_task = mocker.patch("services.human_input_service.resume_app_execution")
@@ -202,14 +190,13 @@ def test_enqueue_resume_skips_unsupported_app_mode(
def test_get_form_definition_by_token_for_console_uses_repository(
sample_form_record: HumanInputFormRecord, sqlite_session_factory
sample_form_record: HumanInputFormRecord, unbound_session_factory
):
session_factory, _ = sqlite_session_factory
repo = MagicMock(spec=HumanInputFormSubmissionRepository)
console_record = dataclasses.replace(sample_form_record, recipient_type=RecipientType.CONSOLE)
repo.get_by_token.return_value = console_record
service = HumanInputService(session_factory, form_repository=repo)
service = HumanInputService(unbound_session_factory, form_repository=repo)
form = service.get_form_definition_by_token_for_console("token")
repo.get_by_token.assert_called_once_with("token")
@@ -245,9 +232,8 @@ def _build_resumption_context_state(*, options: list[str], workflow_run_id: str)
def test_resolve_form_inputs_uses_runtime_select_options(
sample_form_record: HumanInputFormRecord, sqlite_session_factory, mocker: MockerFixture
sample_form_record: HumanInputFormRecord, unbound_session_factory, mocker: MockerFixture
):
session_factory, _ = sqlite_session_factory
configured_input = SelectInputConfig(
output_variable_name="decision",
option_source=StringListSource(
@@ -272,7 +258,7 @@ def test_resolve_form_inputs_uses_runtime_select_options(
"services.human_input_service.DifyAPIRepositoryFactory.create_api_workflow_run_repository",
return_value=workflow_run_repo,
)
service = HumanInputService(session_factory)
service = HumanInputService(unbound_session_factory)
resolved_inputs = service.resolve_form_inputs(Form(record))
@@ -284,13 +270,12 @@ def test_resolve_form_inputs_uses_runtime_select_options(
def test_submit_form_by_token_calls_repository_and_enqueue(
sample_form_record: HumanInputFormRecord, sqlite_session_factory, mocker: MockerFixture
sample_form_record: HumanInputFormRecord, unbound_session_factory, mocker: MockerFixture
):
session_factory, _ = sqlite_session_factory
repo = MagicMock(spec=HumanInputFormSubmissionRepository)
repo.get_by_token.return_value = sample_form_record
repo.mark_submitted.return_value = sample_form_record
service = HumanInputService(session_factory, form_repository=repo)
service = HumanInputService(unbound_session_factory, form_repository=repo)
enqueue_spy = mocker.patch.object(service, "enqueue_resume")
service.submit_form_by_token(
@@ -313,11 +298,10 @@ def test_submit_form_by_token_calls_repository_and_enqueue(
def test_submit_form_by_token_enqueues_agent_app_resume_for_conversation_form(
sample_form_record, sqlite_session_factory, mocker: MockerFixture
sample_form_record, unbound_session_factory, mocker: MockerFixture
):
# ENG-635: a conversation-owned (Agent v2 chat) form routes to the chat
# resume, not the workflow resume.
session_factory, _ = sqlite_session_factory
repo = MagicMock(spec=HumanInputFormSubmissionRepository)
conversation_record = dataclasses.replace(
sample_form_record,
@@ -326,7 +310,7 @@ def test_submit_form_by_token_enqueues_agent_app_resume_for_conversation_form(
)
repo.get_by_token.return_value = conversation_record
repo.mark_submitted.return_value = conversation_record
service = HumanInputService(session_factory, form_repository=repo)
service = HumanInputService(unbound_session_factory, form_repository=repo)
workflow_enqueue_spy = mocker.patch.object(service, "enqueue_resume")
chat_enqueue_spy = mocker.patch.object(service, "enqueue_agent_app_resume")
@@ -343,9 +327,8 @@ def test_submit_form_by_token_enqueues_agent_app_resume_for_conversation_form(
def test_submit_form_by_token_skips_enqueue_for_delivery_test(
sample_form_record: HumanInputFormRecord, sqlite_session_factory, mocker: MockerFixture
sample_form_record: HumanInputFormRecord, unbound_session_factory, mocker: MockerFixture
):
session_factory, _ = sqlite_session_factory
repo = MagicMock(spec=HumanInputFormSubmissionRepository)
test_record = dataclasses.replace(
sample_form_record,
@@ -354,7 +337,7 @@ def test_submit_form_by_token_skips_enqueue_for_delivery_test(
)
repo.get_by_token.return_value = test_record
repo.mark_submitted.return_value = test_record
service = HumanInputService(session_factory, form_repository=repo)
service = HumanInputService(unbound_session_factory, form_repository=repo)
enqueue_spy = mocker.patch.object(service, "enqueue_resume")
service.submit_form_by_token(
@@ -368,13 +351,12 @@ def test_submit_form_by_token_skips_enqueue_for_delivery_test(
def test_submit_form_by_token_passes_submission_user_id(
sample_form_record: HumanInputFormRecord, sqlite_session_factory, mocker: MockerFixture
sample_form_record: HumanInputFormRecord, unbound_session_factory, mocker: MockerFixture
):
session_factory, _ = sqlite_session_factory
repo = MagicMock(spec=HumanInputFormSubmissionRepository)
repo.get_by_token.return_value = sample_form_record
repo.mark_submitted.return_value = sample_form_record
service = HumanInputService(session_factory, form_repository=repo)
service = HumanInputService(unbound_session_factory, form_repository=repo)
enqueue_spy = mocker.patch.object(service, "enqueue_resume")
service.submit_form_by_token(
@@ -391,11 +373,10 @@ def test_submit_form_by_token_passes_submission_user_id(
enqueue_spy.assert_called_once_with(sample_form_record.workflow_run_id)
def test_submit_form_by_token_invalid_action(sample_form_record: HumanInputFormRecord, sqlite_session_factory):
session_factory, _ = sqlite_session_factory
def test_submit_form_by_token_invalid_action(sample_form_record: HumanInputFormRecord, unbound_session_factory):
repo = MagicMock(spec=HumanInputFormSubmissionRepository)
repo.get_by_token.return_value = dataclasses.replace(sample_form_record)
service = HumanInputService(session_factory, form_repository=repo)
service = HumanInputService(unbound_session_factory, form_repository=repo)
with pytest.raises(InvalidFormDataError) as exc_info:
service.submit_form_by_token(
@@ -409,8 +390,7 @@ def test_submit_form_by_token_invalid_action(sample_form_record: HumanInputFormR
repo.mark_submitted.assert_not_called()
def test_submit_form_by_token_missing_inputs(sample_form_record: HumanInputFormRecord, sqlite_session_factory):
session_factory, _ = sqlite_session_factory
def test_submit_form_by_token_missing_inputs(sample_form_record: HumanInputFormRecord, unbound_session_factory):
repo = MagicMock(spec=HumanInputFormSubmissionRepository)
definition_with_input = FormDefinition(
@@ -422,7 +402,7 @@ def test_submit_form_by_token_missing_inputs(sample_form_record: HumanInputFormR
)
form_with_input = dataclasses.replace(sample_form_record, definition=definition_with_input)
repo.get_by_token.return_value = form_with_input
service = HumanInputService(session_factory, form_repository=repo)
service = HumanInputService(unbound_session_factory, form_repository=repo)
with pytest.raises(InvalidFormDataError) as exc_info:
service.submit_form_by_token(
@@ -436,42 +416,6 @@ def test_submit_form_by_token_missing_inputs(sample_form_record: HumanInputFormR
repo.mark_submitted.assert_not_called()
def test_validate_human_input_submission_accepts_select_file_and_file_list(sqlite_session_factory):
session_factory, _ = sqlite_session_factory
service = HumanInputService(session_factory)
definition = FormDefinition.model_validate(
{
"form_content": "Pick one and upload files",
"inputs": [
{
"type": "select",
"output_variable_name": "decision",
"option_source": {
"type": "constant",
"value": ["approve", "reject"],
},
},
{
"type": "file",
"output_variable_name": "attachment",
"allowed_file_types": ["document"],
"allowed_file_upload_methods": ["remote_url"],
},
{
"type": "file-list",
"output_variable_name": "attachments",
"allowed_file_types": ["document"],
"allowed_file_upload_methods": ["remote_url"],
"number_limits": 3,
},
],
"user_actions": [{"id": "submit", "title": "Submit"}],
"rendered_content": "<p>Pick one and upload files</p>",
"expiration_time": naive_utc_now() + timedelta(hours=1),
}
)
@pytest.mark.parametrize(
("input_definition", "submitted_value", "expected_message"),
[
@@ -522,12 +466,11 @@ def test_validate_human_input_submission_accepts_select_file_and_file_list(sqlit
)
def test_validate_human_input_submission_rejects_invalid_select_and_file_payloads(
sample_form_record,
sqlite_session_factory,
unbound_session_factory,
input_definition,
submitted_value,
expected_message,
):
session_factory, _ = sqlite_session_factory
repo = MagicMock(spec=HumanInputFormSubmissionRepository)
definition = FormDefinition.model_validate(
{
@@ -539,7 +482,7 @@ def test_validate_human_input_submission_rejects_invalid_select_and_file_payload
}
)
repo.get_by_token.return_value = dataclasses.replace(sample_form_record, definition=definition)
service = HumanInputService(session_factory, form_repository=repo)
service = HumanInputService(unbound_session_factory, form_repository=repo)
with pytest.raises(InvalidFormDataError) as exc_info:
service.submit_form_by_token(
@@ -569,7 +512,7 @@ def test_form_properties(sample_form_record: HumanInputFormRecord):
def test_form_submitted_error_init():
error = FormSubmittedError(form_id="test-form")
assert "form_id=test-form" in error.description
assert error.description == "This form has already been submitted by another user, form_id=test-form"
assert error.code == 412
@@ -580,61 +523,55 @@ def test_human_input_service_init_with_engine(sqlite_engine: Engine):
assert service._session_factory.kw["bind"] is sqlite_engine
def test_get_form_by_token_none(sqlite_session_factory):
session_factory, _ = sqlite_session_factory
def test_get_form_by_token_none(unbound_session_factory):
repo = MagicMock(spec=HumanInputFormSubmissionRepository)
repo.get_by_token.return_value = None
service = HumanInputService(session_factory, form_repository=repo)
service = HumanInputService(unbound_session_factory, form_repository=repo)
assert service.get_form_by_token("invalid") is None
def test_get_form_definition_by_token_mismatch(sample_form_record: HumanInputFormRecord, sqlite_session_factory):
session_factory, _ = sqlite_session_factory
def test_get_form_definition_by_token_mismatch(sample_form_record: HumanInputFormRecord, unbound_session_factory):
repo = MagicMock(spec=HumanInputFormSubmissionRepository)
repo.get_by_token.return_value = sample_form_record
service = HumanInputService(session_factory, form_repository=repo)
service = HumanInputService(unbound_session_factory, form_repository=repo)
# RecipientType mismatch
assert service.get_form_definition_by_token(RecipientType.CONSOLE, "token") is None
def test_get_form_definition_by_token_success(sample_form_record: HumanInputFormRecord, sqlite_session_factory):
session_factory, _ = sqlite_session_factory
def test_get_form_definition_by_token_success(sample_form_record: HumanInputFormRecord, unbound_session_factory):
repo = MagicMock(spec=HumanInputFormSubmissionRepository)
repo.get_by_token.return_value = sample_form_record
service = HumanInputService(session_factory, form_repository=repo)
service = HumanInputService(unbound_session_factory, form_repository=repo)
form = service.get_form_definition_by_token(RecipientType.STANDALONE_WEB_APP, "token")
assert form is not None
assert form.id == sample_form_record.form_id
def test_get_form_definition_by_token_for_console_mismatch(
sample_form_record: HumanInputFormRecord, sqlite_session_factory
sample_form_record: HumanInputFormRecord, unbound_session_factory
):
session_factory, _ = sqlite_session_factory
repo = MagicMock(spec=HumanInputFormSubmissionRepository)
repo.get_by_token.return_value = sample_form_record # is STANDALONE_WEB_APP
service = HumanInputService(session_factory, form_repository=repo)
service = HumanInputService(unbound_session_factory, form_repository=repo)
assert service.get_form_definition_by_token_for_console("token") is None
def test_submit_form_by_token_delivery_not_enabled(sqlite_session_factory):
session_factory, _ = sqlite_session_factory
def test_submit_form_by_token_delivery_not_enabled(unbound_session_factory):
repo = MagicMock(spec=HumanInputFormSubmissionRepository)
repo.get_by_token.return_value = None
service = HumanInputService(session_factory, form_repository=repo)
service = HumanInputService(unbound_session_factory, form_repository=repo)
with pytest.raises(human_input_service_module.WebAppDeliveryNotEnabledError):
service.submit_form_by_token(RecipientType.STANDALONE_WEB_APP, "token", "action", {})
def test_submit_form_by_token_no_workflow_run_id(
sample_form_record: HumanInputFormRecord, sqlite_session_factory, mocker: MockerFixture
sample_form_record: HumanInputFormRecord, unbound_session_factory, mocker: MockerFixture
):
session_factory, _ = sqlite_session_factory
repo = MagicMock(spec=HumanInputFormSubmissionRepository)
repo.get_by_token.return_value = sample_form_record
@@ -642,16 +579,15 @@ def test_submit_form_by_token_no_workflow_run_id(
result_record = dataclasses.replace(sample_form_record, workflow_run_id=None)
repo.mark_submitted.return_value = result_record
service = HumanInputService(session_factory, form_repository=repo)
service = HumanInputService(unbound_session_factory, form_repository=repo)
enqueue_spy = mocker.patch.object(service, "enqueue_resume")
service.submit_form_by_token(RecipientType.STANDALONE_WEB_APP, "token", "submit", {})
enqueue_spy.assert_not_called()
def test_ensure_form_active_errors(sample_form_record: HumanInputFormRecord, sqlite_session_factory):
session_factory, _ = sqlite_session_factory
service = HumanInputService(session_factory)
def test_ensure_form_active_errors(sample_form_record: HumanInputFormRecord, unbound_session_factory):
service = HumanInputService(unbound_session_factory)
# Submitted
submitted_record = dataclasses.replace(sample_form_record, submitted_at=naive_utc_now())
@@ -671,18 +607,16 @@ def test_ensure_form_active_errors(sample_form_record: HumanInputFormRecord, sql
service.ensure_form_active(Form(expired_time_record))
def test_ensure_not_submitted_raises(sample_form_record: HumanInputFormRecord, sqlite_session_factory):
session_factory, _ = sqlite_session_factory
service = HumanInputService(session_factory)
def test_ensure_not_submitted_raises(sample_form_record: HumanInputFormRecord, unbound_session_factory):
service = HumanInputService(unbound_session_factory)
submitted_record = dataclasses.replace(sample_form_record, submitted_at=naive_utc_now())
with pytest.raises(human_input_service_module.FormSubmittedError):
service._ensure_not_submitted(Form(submitted_record))
def test_enqueue_resume_workflow_not_found(mocker: MockerFixture, sqlite_session_factory):
session_factory, _ = sqlite_session_factory
service = HumanInputService(session_factory)
def test_enqueue_resume_workflow_not_found(mocker: MockerFixture, unbound_session_factory):
service = HumanInputService(unbound_session_factory)
workflow_run_repo = MagicMock()
workflow_run_repo.get_workflow_run_by_id_without_tenant.return_value = None
@@ -696,15 +630,12 @@ def test_enqueue_resume_workflow_not_found(mocker: MockerFixture, sqlite_session
assert "WorkflowRun not found" in str(excinfo.value)
@pytest.mark.parametrize("sqlite_session", [(App,)], indirect=True)
def test_enqueue_resume_app_not_found(
mocker,
sqlite_session_factory,
sqlite_session: Session,
sqlite_session_factory: sessionmaker[Session],
caplog: pytest.LogCaptureFixture,
):
session_factory, _ = sqlite_session_factory
service = HumanInputService(session_factory)
service = HumanInputService(sqlite_session_factory)
workflow_run = MagicMock()
workflow_run.app_id = "app-id"
@@ -715,26 +646,31 @@ def test_enqueue_resume_app_not_found(
"services.human_input_service.DifyAPIRepositoryFactory.create_api_workflow_run_repository",
return_value=workflow_run_repo,
)
resume_task = mocker.patch("services.human_input_service.resume_app_execution")
with caplog.at_level(logging.ERROR, logger="services.human_input_service"):
service.enqueue_resume("workflow-run-id")
assert any(r.levelno >= logging.ERROR for r in caplog.records)
assert (
"services.human_input_service",
logging.ERROR,
"App not found for WorkflowRun, workflow_run_id=workflow-run-id, app_id=app-id",
) in caplog.record_tuples
resume_task.apply_async.assert_not_called()
def test_is_globally_expired_zero_timeout(
monkeypatch: pytest.MonkeyPatch, sample_form_record: HumanInputFormRecord, sqlite_session_factory
monkeypatch: pytest.MonkeyPatch, sample_form_record: HumanInputFormRecord, unbound_session_factory
):
session_factory, _ = sqlite_session_factory
service = HumanInputService(session_factory)
service = HumanInputService(unbound_session_factory)
monkeypatch.setattr(human_input_service_module.dify_config, "HUMAN_INPUT_GLOBAL_TIMEOUT_SECONDS", 0)
assert service._is_globally_expired(Form(sample_form_record)) is False
def test_submit_form_by_token_normalizes_select_and_files(
sample_form_record: HumanInputFormRecord, sqlite_session_factory, mocker: MockerFixture
sample_form_record: HumanInputFormRecord, unbound_session_factory, mocker: MockerFixture
) -> None:
session_factory, _ = sqlite_session_factory
repo = MagicMock(spec=HumanInputFormSubmissionRepository)
definition = FormDefinition(
form_content="hello",
@@ -753,7 +689,7 @@ def test_submit_form_by_token_normalizes_select_and_files(
form_with_inputs = dataclasses.replace(sample_form_record, definition=definition)
repo.get_by_token.return_value = form_with_inputs
repo.mark_submitted.return_value = form_with_inputs
service = HumanInputService(session_factory, form_repository=repo)
service = HumanInputService(unbound_session_factory, form_repository=repo)
single_file = File(
file_id="file-1",
@@ -815,9 +751,8 @@ def test_submit_form_by_token_normalizes_select_and_files(
def test_submit_form_by_token_invalid_select_value(
sample_form_record: HumanInputFormRecord, sqlite_session_factory
sample_form_record: HumanInputFormRecord, unbound_session_factory
) -> None:
session_factory, _ = sqlite_session_factory
repo = MagicMock(spec=HumanInputFormSubmissionRepository)
definition = FormDefinition(
form_content="hello",
@@ -832,7 +767,7 @@ def test_submit_form_by_token_invalid_select_value(
expiration_time=sample_form_record.expiration_time,
)
repo.get_by_token.return_value = dataclasses.replace(sample_form_record, definition=definition)
service = HumanInputService(session_factory, form_repository=repo)
service = HumanInputService(unbound_session_factory, form_repository=repo)
with pytest.raises(InvalidFormDataError, match="Invalid value for select input 'decision'"):
service.submit_form_by_token(
@@ -844,9 +779,8 @@ def test_submit_form_by_token_invalid_select_value(
def test_submit_form_by_token_invalid_file_list_item(
sample_form_record: HumanInputFormRecord, sqlite_session_factory
sample_form_record: HumanInputFormRecord, unbound_session_factory
) -> None:
session_factory, _ = sqlite_session_factory
repo = MagicMock(spec=HumanInputFormSubmissionRepository)
definition = FormDefinition(
form_content="hello",
@@ -856,7 +790,7 @@ def test_submit_form_by_token_invalid_file_list_item(
expiration_time=sample_form_record.expiration_time,
)
repo.get_by_token.return_value = dataclasses.replace(sample_form_record, definition=definition)
service = HumanInputService(session_factory, form_repository=repo)
service = HumanInputService(unbound_session_factory, form_repository=repo)
with pytest.raises(
InvalidFormDataError,
@@ -871,9 +805,8 @@ def test_submit_form_by_token_invalid_file_list_item(
def test_submit_form_by_token_rejects_cross_tenant_file(
sample_form_record: HumanInputFormRecord, sqlite_session_factory, mocker: MockerFixture
sample_form_record: HumanInputFormRecord, unbound_session_factory, mocker: MockerFixture
) -> None:
session_factory, _ = sqlite_session_factory
repo = MagicMock(spec=HumanInputFormSubmissionRepository)
definition = FormDefinition(
form_content="hello",
@@ -883,7 +816,7 @@ def test_submit_form_by_token_rejects_cross_tenant_file(
expiration_time=sample_form_record.expiration_time,
)
repo.get_by_token.return_value = dataclasses.replace(sample_form_record, definition=definition)
service = HumanInputService(session_factory, form_repository=repo)
service = HumanInputService(unbound_session_factory, form_repository=repo)
mocker.patch("services.human_input_service.build_from_mapping", side_effect=ValueError("Invalid upload file"))
with pytest.raises(InvalidFormDataError, match="Invalid value for file input 'attachment'"):
@@ -904,9 +837,8 @@ def test_submit_form_by_token_rejects_cross_tenant_file(
def test_submit_form_by_token_rejects_cross_tenant_file_list(
sample_form_record: HumanInputFormRecord, sqlite_session_factory, mocker: MockerFixture
sample_form_record: HumanInputFormRecord, unbound_session_factory, mocker: MockerFixture
) -> None:
session_factory, _ = sqlite_session_factory
repo = MagicMock(spec=HumanInputFormSubmissionRepository)
definition = FormDefinition(
form_content="hello",
@@ -916,7 +848,7 @@ def test_submit_form_by_token_rejects_cross_tenant_file_list(
expiration_time=sample_form_record.expiration_time,
)
repo.get_by_token.return_value = dataclasses.replace(sample_form_record, definition=definition)
service = HumanInputService(session_factory, form_repository=repo)
service = HumanInputService(unbound_session_factory, form_repository=repo)
mocker.patch("services.human_input_service.build_from_mappings", side_effect=ValueError("Invalid upload file"))
with pytest.raises(
@@ -1,16 +1,20 @@
"""Testcontainers integration tests for OAuthServerService."""
"""Unit tests for OAuthServerService with SQLite-backed database access."""
from __future__ import annotations
import uuid
from collections.abc import Iterator
from typing import cast
from unittest.mock import MagicMock, patch
from uuid import uuid4
import pytest
from flask import Flask
from sqlalchemy import Engine
from sqlalchemy.orm import Session
from werkzeug.exceptions import BadRequest
from models.engine import db
from models.model import OAuthProviderApp
from services.oauth_server import (
OAUTH_ACCESS_TOKEN_EXPIRES_IN,
@@ -23,10 +27,24 @@ from services.oauth_server import (
)
class TestOAuthServerServiceGetProviderApp:
"""DB-backed tests for get_oauth_provider_app."""
@pytest.fixture
def oauth_db() -> Iterator[Session]:
"""Provide the production database extension with an isolated SQLite provider table."""
app = Flask(__name__)
app.config["SQLALCHEMY_DATABASE_URI"] = "sqlite:///:memory:"
db.init_app(app)
def _create_oauth_provider_app(self, db_session_with_containers: Session, *, client_id: str) -> OAuthProviderApp:
with app.app_context():
OAuthProviderApp.__table__.create(db.engine)
with Session(db.engine, expire_on_commit=False) as session:
yield session
class TestOAuthServerServiceGetProviderApp:
"""Verify provider lookup against a real SQLAlchemy database."""
def test_get_oauth_provider_app_returns_app_when_exists(self, oauth_db: Session) -> None:
client_id = f"client-{uuid4()}"
app = OAuthProviderApp(
app_icon="icon.png",
client_id=client_id,
@@ -35,35 +53,30 @@ class TestOAuthServerServiceGetProviderApp:
redirect_uris=["https://example.com/callback"],
scope="read",
)
db_session_with_containers.add(app)
db_session_with_containers.commit()
return app
def test_get_oauth_provider_app_returns_app_when_exists(self, db_session_with_containers: Session):
client_id = f"client-{uuid4()}"
created = self._create_oauth_provider_app(db_session_with_containers, client_id=client_id)
oauth_db.add(app)
oauth_db.commit()
result = OAuthServerService.get_oauth_provider_app(client_id)
assert result is not None
assert result.client_id == client_id
assert result.id == created.id
assert result.id == app.id
def test_get_oauth_provider_app_returns_none_when_not_exists(self, db_session_with_containers: Session):
def test_get_oauth_provider_app_returns_none_when_not_exists(self, oauth_db: Session) -> None:
result = OAuthServerService.get_oauth_provider_app(f"nonexistent-{uuid4()}")
assert result is None
class TestOAuthServerServiceTokenOperations:
"""Redis-backed tests for token sign/validate operations."""
"""Verify Redis-backed token signing and validation branches."""
@pytest.fixture
def mock_redis(self):
with patch("services.oauth_server.redis_client") as mock:
yield mock
def test_sign_authorization_code_stores_and_returns_code(self, mock_redis):
def test_sign_authorization_code_stores_and_returns_code(self, mock_redis) -> None:
deterministic_uuid = uuid.UUID("00000000-0000-0000-0000-000000000111")
with patch("services.oauth_server.uuid.uuid4", return_value=deterministic_uuid):
code = OAuthServerService.sign_oauth_authorization_code("client-1", "user-1")
@@ -75,7 +88,7 @@ class TestOAuthServerServiceTokenOperations:
ex=600,
)
def test_sign_access_token_raises_bad_request_for_invalid_code(self, mock_redis):
def test_sign_access_token_raises_bad_request_for_invalid_code(self, mock_redis) -> None:
mock_redis.get.return_value = None
with pytest.raises(BadRequest, match="invalid code"):
@@ -85,14 +98,13 @@ class TestOAuthServerServiceTokenOperations:
client_id="client-1",
)
def test_sign_access_token_issues_tokens_for_valid_code(self, mock_redis):
def test_sign_access_token_issues_tokens_for_valid_code(self, mock_redis) -> None:
token_uuids = [
uuid.UUID("00000000-0000-0000-0000-000000000201"),
uuid.UUID("00000000-0000-0000-0000-000000000202"),
]
with patch("services.oauth_server.uuid.uuid4", side_effect=token_uuids):
mock_redis.get.return_value = b"user-1"
access_token, refresh_token = OAuthServerService.sign_oauth_access_token(
grant_type=OAuthGrantType.AUTHORIZATION_CODE,
code="code-1",
@@ -114,7 +126,7 @@ class TestOAuthServerServiceTokenOperations:
ex=OAUTH_REFRESH_TOKEN_EXPIRES_IN,
)
def test_sign_access_token_raises_bad_request_for_invalid_refresh_token(self, mock_redis):
def test_sign_access_token_raises_bad_request_for_invalid_refresh_token(self, mock_redis) -> None:
mock_redis.get.return_value = None
with pytest.raises(BadRequest, match="invalid refresh token"):
@@ -124,11 +136,10 @@ class TestOAuthServerServiceTokenOperations:
client_id="client-1",
)
def test_sign_access_token_issues_new_token_for_valid_refresh(self, mock_redis):
def test_sign_access_token_issues_new_token_for_valid_refresh(self, mock_redis) -> None:
deterministic_uuid = uuid.UUID("00000000-0000-0000-0000-000000000301")
with patch("services.oauth_server.uuid.uuid4", return_value=deterministic_uuid):
mock_redis.get.return_value = b"user-1"
access_token, returned_refresh = OAuthServerService.sign_oauth_access_token(
grant_type=OAuthGrantType.REFRESH_TOKEN,
refresh_token="refresh-1",
@@ -138,14 +149,14 @@ class TestOAuthServerServiceTokenOperations:
assert access_token == str(deterministic_uuid)
assert returned_refresh == "refresh-1"
def test_sign_access_token_returns_none_for_unknown_grant_type(self, mock_redis):
def test_sign_access_token_returns_none_for_unknown_grant_type(self, mock_redis) -> None:
grant_type = cast(OAuthGrantType, "invalid-grant-type")
result = OAuthServerService.sign_oauth_access_token(grant_type=grant_type, client_id="client-1")
assert result is None
def test_sign_refresh_token_stores_with_expected_expiry(self, mock_redis):
def test_sign_refresh_token_stores_with_expected_expiry(self, mock_redis) -> None:
deterministic_uuid = uuid.UUID("00000000-0000-0000-0000-000000000401")
with patch("services.oauth_server.uuid.uuid4", return_value=deterministic_uuid):
refresh_token = OAuthServerService._sign_oauth_refresh_token("client-2", "user-2")
@@ -157,22 +168,21 @@ class TestOAuthServerServiceTokenOperations:
ex=OAUTH_REFRESH_TOKEN_EXPIRES_IN,
)
def test_validate_access_token_returns_none_when_not_found(self, mock_redis, db_session_with_containers: Session):
def test_validate_access_token_returns_none_when_not_found(self, mock_redis, sqlite_engine: Engine) -> None:
mock_redis.get.return_value = None
session = MagicMock()
result = OAuthServerService.validate_oauth_access_token("client-1", "missing-token", db_session_with_containers)
with Session(sqlite_engine) as session:
result = OAuthServerService.validate_oauth_access_token("client-1", "missing-token", session)
assert result is None
def test_validate_access_token_loads_user_when_exists(self, mock_redis, db_session_with_containers: Session):
def test_validate_access_token_loads_user_when_exists(self, mock_redis, sqlite_engine: Engine) -> None:
mock_redis.get.return_value = b"user-88"
expected_user = MagicMock()
with patch("services.oauth_server.AccountService.load_user", return_value=expected_user) as mock_load:
result = OAuthServerService.validate_oauth_access_token(
"client-1", "access-token", db_session_with_containers
)
with Session(sqlite_engine) as session:
with patch("services.oauth_server.AccountService.load_user", return_value=expected_user) as mock_load:
result = OAuthServerService.validate_oauth_access_token("client-1", "access-token", session)
mock_load.assert_called_once_with("user-88", session)
assert result is expected_user
mock_load.assert_called_once_with("user-88", db_session_with_containers)
File diff suppressed because it is too large Load Diff

Some files were not shown because too many files have changed in this diff Show More