version to 0.6.6 (#4050 )

test: remove explicit env settings for CI pytests (#4041 )
chore(deps): bump semver from 5.7.1 to 5.7.2 in /web (#4022 )
2024-05-02 16:06:40 +08:00 · 2024-05-02 00:49:39 +08:00 · 2024-04-30 18:47:05 +08:00 · 2024-04-30 18:39:02 +08:00 · 2024-04-30 18:11:49 +08:00 · 2024-04-30 09:39:10 +00:00
395 changed files with 12219 additions and 3496 deletions
@@ -32,8 +32,8 @@
 			]
 		}
 	},
-	"postStartCommand": "cd api && pip install -r requirements.txt",
-	"postCreateCommand": "cd web && npm install"
+	"postStartCommand": "./.devcontainer/post_start_command.sh",
+	"postCreateCommand": "./.devcontainer/post_create_command.sh"

 	// Features to add to the dev container. More info: https://containers.dev/features.
 	// "features": {},
@@ -0,0 +1,10 @@
+#!/bin/bash
+
+cd web && npm install
+
+echo 'alias start-api="cd /workspaces/dify/api && flask run --host 0.0.0.0 --port=5001 --debug"' >> ~/.bashrc
+echo 'alias start-worker="cd /workspaces/dify/api && celery -A app.celery worker -P gevent -c 1 --loglevel INFO -Q dataset,generation,mail"' >> ~/.bashrc
+echo 'alias start-web="cd /workspaces/dify/web && npm run dev"' >> ~/.bashrc
+echo 'alias start-containers="cd /workspaces/dify/docker && docker-compose -f docker-compose.middleware.yaml -p dify up -d"' >> ~/.bashrc
+
+source /home/vscode/.bashrc
@@ -0,0 +1,3 @@
+#!/bin/bash
+
+cd api && pip install -r requirements.txt
@@ -10,36 +10,14 @@ jobs:
    runs-on: ubuntu-latest
    strategy:
      matrix:
-        python-version: ["3.10", "3.11", "3.12"]
-
-    env:
-      OPENAI_API_KEY: sk-IamNotARealKeyJustForMockTestKawaiiiiiiiiii
-      AZURE_OPENAI_API_BASE: https://difyai-openai.openai.azure.com
-      AZURE_OPENAI_API_KEY: xxxxb1707exxxxxxxxxxaaxxxxxf94
-      ANTHROPIC_API_KEY: sk-ant-api11-IamNotARealKeyJustForMockTestKawaiiiiiiiiii-NotBaka-ASkksz
-      CHATGLM_API_BASE: http://a.abc.com:11451
-      XINFERENCE_SERVER_URL: http://a.abc.com:11451
-      XINFERENCE_GENERATION_MODEL_UID: generate
-      XINFERENCE_CHAT_MODEL_UID: chat
-      XINFERENCE_EMBEDDINGS_MODEL_UID: embedding
-      XINFERENCE_RERANK_MODEL_UID: rerank
-      GOOGLE_API_KEY: abcdefghijklmnopqrstuvwxyz
-      HUGGINGFACE_API_KEY: hf-awuwuwuwuwuwuwuwuwuwuwuwuwuwuwuwuwu
-      HUGGINGFACE_TEXT_GEN_ENDPOINT_URL: a
-      HUGGINGFACE_TEXT2TEXT_GEN_ENDPOINT_URL: b
-      HUGGINGFACE_EMBEDDINGS_ENDPOINT_URL: c
-      MOCK_SWITCH: true
-      CODE_MAX_STRING_LENGTH: 80000
+        python-version:
+          - "3.10"
+          - "3.11"

    steps:
      - name: Checkout code
        uses: actions/checkout@v4

-      - name: Install APT packages
-        uses: awalsh128/cache-apt-pkgs-action@v1
-        with:
-          packages: ffmpeg
-
      - name: Set up Python ${{ matrix.python-version }}
        uses: actions/setup-python@v5
        with:
@@ -52,6 +30,9 @@ jobs:
      - name: Install dependencies
        run: pip install -r ./api/requirements.txt -r ./api/requirements-dev.txt

+      - name: Run Unit tests
+        run: dev/pytest/pytest_unit_tests.sh
+
      - name: Run ModelRuntime
        run: dev/pytest/pytest_model_runtime.sh

@@ -60,3 +41,22 @@ jobs:

      - name: Run Workflow
        run: dev/pytest/pytest_workflow.sh
+
+      - name: Set up Vector Stores (Weaviate, Qdrant, Milvus, PgVecto-RS)
+        uses: hoverkraft-tech/compose-action@v2.0.0
+        with:
+          compose-file: |
+            docker/docker-compose.middleware.yaml
+            docker/docker-compose.qdrant.yaml
+            docker/docker-compose.milvus.yaml
+            docker/docker-compose.pgvecto-rs.yaml
+          services: |
+            weaviate
+            qdrant
+            etcd
+            minio
+            milvus-standalone
+            pgvecto-rs
+
+      - name: Test Vector Stores
+        run: dev/pytest/pytest_vdb.sh
@@ -29,12 +29,12 @@
 </p>

 <p align="center">
-  <a href="./README.md"><img alt="Commits last month" src="https://img.shields.io/badge/English-d9d9d9"></a>
-  <a href="./README_CN.md"><img alt="Commits last month" src="https://img.shields.io/badge/简体中文-d9d9d9"></a>
-  <a href="./README_JA.md"><img alt="Commits last month" src="https://img.shields.io/badge/日本語-d9d9d9"></a>
-  <a href="./README_ES.md"><img alt="Commits last month" src="https://img.shields.io/badge/Español-d9d9d9"></a>
-  <a href="./README_FR.md"><img alt="Commits last month" src="https://img.shields.io/badge/Français-d9d9d9"></a>
-  <a href="./README_KL.md"><img alt="Commits last month" src="https://img.shields.io/badge/Klingon-d9d9d9"></a>
+  <a href="./README.md"><img alt="README in English" src="https://img.shields.io/badge/English-d9d9d9"></a>
+  <a href="./README_CN.md"><img alt="简体中文版自述文件" src="https://img.shields.io/badge/简体中文-d9d9d9"></a>
+  <a href="./README_JA.md"><img alt="日本語のREADME" src="https://img.shields.io/badge/日本語-d9d9d9"></a>
+  <a href="./README_ES.md"><img alt="README en Español" src="https://img.shields.io/badge/Español-d9d9d9"></a>
+  <a href="./README_FR.md"><img alt="README en Français" src="https://img.shields.io/badge/Français-d9d9d9"></a>
+  <a href="./README_KL.md"><img alt="README tlhIngan Hol" src="https://img.shields.io/badge/Klingon-d9d9d9"></a>
 </p>

 #
@@ -54,7 +54,7 @@ Dify is an open-source LLM app development platform. Its intuitive interface com


 **2. Comprehensive model support**: 
-  Seamless integration with hundreds of proprietary / open-source LLMs from dozens of inference providers and self-hosted solutions, covering GPT, Mistral, Llama2, and any OpenAI API-compatible models. A full list of supported model providers can be found [here](https://docs.dify.ai/getting-started/readme/model-providers).
+  Seamless integration with hundreds of proprietary / open-source LLMs from dozens of inference providers and self-hosted solutions, covering GPT, Mistral, Llama3, and any OpenAI API-compatible models. A full list of supported model providers can be found [here](https://docs.dify.ai/getting-started/readme/model-providers).

 ![providers-v5](https://github.com/langgenius/dify/assets/13230914/5a17bdbe-097a-4100-8363-40255b70f6e3)

@@ -54,7 +54,7 @@ Dify es una plataforma de desarrollo de aplicaciones de LLM de código abierto.


 **2. Soporte de modelos completo**: 
-  Integración perfecta con cientos de LLMs propietarios / de código abierto de docenas de proveedores de inferencia y soluciones auto-alojadas, que cubren GPT, Mistral, Llama2 y cualquier modelo compatible con la API de OpenAI. Se puede encontrar una lista completa de proveedores de modelos admitidos [aquí](https://docs.dify.ai/getting-started/readme/model-providers).
+  Integración perfecta con cientos de LLMs propietarios / de código abierto de docenas de proveedores de inferencia y soluciones auto-alojadas, que cubren GPT, Mistral, Llama3 y cualquier modelo compatible con la API de OpenAI. Se puede encontrar una lista completa de proveedores de modelos admitidos [aquí](https://docs.dify.ai/getting-started/readme/model-providers).

 ![proveedores-v5](https://github.com/langgenius/dify/assets/13230914/5a17bdbe-097a-4100-8363-40255b70f6e3)

@@ -54,7 +54,7 @@ Dify est une plateforme de développement d'applications LLM open source. Son in


 **2. Prise en charge complète des modèles**: 
-  Intégration transparente avec des centaines de LLM propriétaires / open source provenant de dizaines de fournisseurs d'inférence et de solutions auto-hébergées, couvrant GPT, Mistral, Llama2, et tous les modèles compatibles avec l'API OpenAI. Une liste complète des fournisseurs de modèles pris en charge se trouve [ici](https://docs.dify.ai/getting-started/readme/model-providers).
+  Intégration transparente avec des centaines de LLM propriétaires / open source provenant de dizaines de fournisseurs d'inférence et de solutions auto-hébergées, couvrant GPT, Mistral, Llama3, et tous les modèles compatibles avec l'API OpenAI. Une liste complète des fournisseurs de modèles pris en charge se trouve [ici](https://docs.dify.ai/getting-started/readme/model-providers).

 ![providers-v5](https://github.com/langgenius/dify/assets/13230914/5a17bdbe-097a-4100-8363-40255b70f6e3)

@@ -55,9 +55,7 @@ DifyはオープンソースのLLMアプリケーション開発プラットフ


 **2. 網羅的なモデルサポート**: 
-  数百のプロプライエタリ/オープンソースのLLMと、数十の推論プロバイダーおよびセルフホスティングソリューションとのシームレスな統合を提供します。GPT、Mistral、Llama2、およびOpenAI API互換のモデルをカバーします。サポートされているモデルプロバイダーの完全なリストは[こちら](https://docs
-
-.dify.ai/getting-started/readme/model-providers)をご覧ください。
+  数百のプロプライエタリ/オープンソースのLLMと、数十の推論プロバイダーおよびセルフホスティングソリューションとのシームレスな統合を提供します。GPT、Mistral、Llama3、およびOpenAI API互換のモデルをカバーします。サポートされているモデルプロバイダーの完全なリストは[こちら](https://docs.dify.ai/getting-started/readme/model-providers)をご覧ください。

 ![providers-v5](https://github.com/langgenius/dify/assets/13230914/5a17bdbe-097a-4100-8363-40255b70f6e3)

@@ -155,9 +153,7 @@ DifyはオープンソースのLLMアプリケーション開発プラットフ
 さらなる参照や詳細な手順については、[ドキュメント](https://docs.dify.ai)をご覧ください。

 - **エンタープライズ/組織向けのDify</br>**
-追加のエンタープライズ向け機能を提供しています。[こちらからミーティ
-
-ングを予約](https://cal.com/guchenhe/30min)したり、[メールを送信](mailto:business@dify.ai?subject=[GitHub]Business%20License%20Inquiry)してエンタープライズのニーズについて相談してください。 </br>
+追加のエンタープライズ向け機能を提供しています。[こちらからミーティングを予約](https://cal.com/guchenhe/30min)したり、[メールを送信](mailto:business@dify.ai?subject=[GitHub]Business%20License%20Inquiry)してエンタープライズのニーズについて相談してください。 </br>
  > AWSを使用しているスタートアップや中小企業の場合は、[AWS Marketplace](https://aws.amazon.com/marketplace/pp/prodview-t22mebxzwjhu6)のDify Premiumをチェックして、ワンクリックで独自のAWS VPCにデプロイできます。カスタムロゴとブランディングでアプリを作成するオプションを備えた手頃な価格のAMIオファリングです。


@@ -54,7 +54,7 @@ Dify is an open-source LLM app development platform. Its intuitive interface com


 **2. Comprehensive model support**: 
-  Seamless integration with hundreds of proprietary / open-source LLMs from dozens of inference providers and self-hosted solutions, covering GPT, Mistral, Llama2, and any OpenAI API-compatible models. A full list of supported model providers can be found [here](https://docs.dify.ai/getting-started/readme/model-providers).
+  Seamless integration with hundreds of proprietary / open-source LLMs from dozens of inference providers and self-hosted solutions, covering GPT, Mistral, Llama3, and any OpenAI API-compatible models. A full list of supported model providers can be found [here](https://docs.dify.ai/getting-started/readme/model-providers).

 ![providers-v5](https://github.com/langgenius/dify/assets/13230914/5a17bdbe-097a-4100-8363-40255b70f6e3)

@@ -1,6 +1,3 @@
-# Server Edition
-EDITION=SELF_HOSTED
-
 # Your App secret key will be used for securely signing the session cookie
 # Make sure you are changing this key for your deployment with a strong key.
 # You can generate a strong key using `openssl rand -base64 42`.
@@ -52,12 +49,20 @@ AZURE_BLOB_ACCOUNT_NAME=your-account-name
 AZURE_BLOB_ACCOUNT_KEY=your-account-key
 AZURE_BLOB_CONTAINER_NAME=yout-container-name
 AZURE_BLOB_ACCOUNT_URL=https://<your_account_name>.blob.core.windows.net
+# Aliyun oss Storage configuration
+ALIYUN_OSS_BUCKET_NAME=your-bucket-name
+ALIYUN_OSS_ACCESS_KEY=your-access-key
+ALIYUN_OSS_SECRET_KEY=your-secret-key
+ALIYUN_OSS_ENDPOINT=your-endpoint
+# Google Storage configuration
+GOOGLE_STORAGE_BUCKET_NAME=yout-bucket-name
+GOOGLE_STORAGE_SERVICE_ACCOUNT_JSON=your-google-service-account-json-base64-string

 # CORS configuration
 WEB_API_CORS_ALLOW_ORIGINS=http://127.0.0.1:3000,*
 CONSOLE_CORS_ALLOW_ORIGINS=http://127.0.0.1:3000,*

-# Vector database configuration, support: weaviate, qdrant, milvus, relyt
+# Vector database configuration, support: weaviate, qdrant, milvus, relyt, pgvecto_rs
 VECTOR_STORE=weaviate

 # Weaviate configuration
@@ -70,6 +75,8 @@ WEAVIATE_BATCH_SIZE=100
 QDRANT_URL=http://localhost:6333
 QDRANT_API_KEY=difyai123456
 QDRANT_CLIENT_TIMEOUT=20
+QDRANT_GRPC_ENABLED=false
+QDRANT_GRPC_PORT=6334

 # Milvus configuration
 MILVUS_HOST=127.0.0.1
@@ -85,6 +92,13 @@ RELYT_USER=postgres
 RELYT_PASSWORD=postgres
 RELYT_DATABASE=postgres

+# PGVECTO_RS configuration
+PGVECTO_RS_HOST=localhost
+PGVECTO_RS_PORT=5431
+PGVECTO_RS_USER=postgres
+PGVECTO_RS_PASSWORD=difyai123456
+PGVECTO_RS_DATABASE=postgres
+
 # Upload configuration
 UPLOAD_FILE_SIZE_LIMIT=15
 UPLOAD_FILE_BATCH_LIMIT=5
@@ -118,25 +132,6 @@ NOTION_CLIENT_SECRET=you-client-secret
 NOTION_CLIENT_ID=you-client-id
 NOTION_INTERNAL_SECRET=you-internal-secret

-# Hosted Model Credentials
-HOSTED_OPENAI_API_KEY=
-HOSTED_OPENAI_API_BASE=
-HOSTED_OPENAI_API_ORGANIZATION=
-HOSTED_OPENAI_TRIAL_ENABLED=false
-HOSTED_OPENAI_QUOTA_LIMIT=200
-HOSTED_OPENAI_PAID_ENABLED=false
-
-HOSTED_AZURE_OPENAI_ENABLED=false
-HOSTED_AZURE_OPENAI_API_KEY=
-HOSTED_AZURE_OPENAI_API_BASE=
-HOSTED_AZURE_OPENAI_QUOTA_LIMIT=200
-
-HOSTED_ANTHROPIC_API_BASE=
-HOSTED_ANTHROPIC_API_KEY=
-HOSTED_ANTHROPIC_TRIAL_ENABLED=false
-HOSTED_ANTHROPIC_QUOTA_LIMIT=600000
-HOSTED_ANTHROPIC_PAID_ENABLED=false
-
 ETL_TYPE=dify
 UNSTRUCTURED_API_URL=

@@ -160,3 +155,11 @@ CODE_MAX_NUMBER_ARRAY_LENGTH=1000
 # API Tool configuration
 API_TOOL_DEFAULT_CONNECT_TIMEOUT=10
 API_TOOL_DEFAULT_READ_TIMEOUT=60
+
+# HTTP Node configuration
+HTTP_REQUEST_MAX_CONNECT_TIMEOUT=300
+HTTP_REQUEST_MAX_READ_TIMEOUT=600
+HTTP_REQUEST_MAX_WRITE_TIMEOUT=600
+
+# Log file path
+LOG_FILE=
@@ -1,28 +1,28 @@
 import os
-import sys
-from logging.handlers import RotatingFileHandler

 if not os.environ.get("DEBUG") or os.environ.get("DEBUG").lower() != 'true':
    from gevent import monkey

    monkey.patch_all()
-    # if os.environ.get("VECTOR_STORE") == 'milvus':
+
    import grpc.experimental.gevent

    grpc.experimental.gevent.init_gevent()

 import json
 import logging
+import sys
 import threading
 import time
 import warnings
+from logging.handlers import RotatingFileHandler

 from flask import Flask, Response, request
 from flask_cors import CORS
 from werkzeug.exceptions import Unauthorized

 from commands import register_commands
-from config import CloudEditionConfig, Config
+from config import Config

 # DO NOT REMOVE BELOW
 from events import event_handlers
@@ -75,16 +75,9 @@ config_type = os.getenv('EDITION', default='SELF_HOSTED')  # ce edition first
 # ----------------------------


-def create_app(test_config=None) -> Flask:
+def create_app() -> Flask:
    app = DifyApp(__name__)
-
-    if test_config:
-        app.config.from_object(test_config)
-    else:
-        if config_type == "CLOUD":
-            app.config.from_object(CloudEditionConfig())
-        else:
-            app.config.from_object(Config())
+    app.config.from_object(Config())

    app.secret_key = app.config['SECRET_KEY']

@@ -101,6 +94,7 @@ def create_app(test_config=None) -> Flask:
            ),
            logging.StreamHandler(sys.stdout)
        ]
+
    logging.basicConfig(
        level=app.config.get('LOG_LEVEL'),
        format=app.config.get('LOG_FORMAT'),
@@ -5,6 +5,7 @@ import dotenv
 dotenv.load_dotenv()

 DEFAULTS = {
+    'EDITION': 'SELF_HOSTED',
    'DB_USERNAME': 'postgres',
    'DB_PASSWORD': '',
    'DB_HOST': 'localhost',
@@ -36,6 +37,8 @@ DEFAULTS = {
    'WEAVIATE_GRPC_ENABLED': 'True',
    'WEAVIATE_BATCH_SIZE': 100,
    'QDRANT_CLIENT_TIMEOUT': 20,
+    'QDRANT_GRPC_ENABLED': 'False',
+    'QDRANT_GRPC_PORT': '6334',
    'CELERY_BACKEND': 'database',
    'LOG_LEVEL': 'INFO',
    'LOG_FILE': '',
@@ -104,9 +107,9 @@ class Config:
        # ------------------------
        # General Configurations.
        # ------------------------
-        self.CURRENT_VERSION = "0.6.4"
+        self.CURRENT_VERSION = "0.6.6"
        self.COMMIT_SHA = get_env('COMMIT_SHA')
-        self.EDITION = "SELF_HOSTED"
+        self.EDITION = get_env('EDITION')
        self.DEPLOY_ENV = get_env('DEPLOY_ENV')
        self.TESTING = False
        self.LOG_LEVEL = get_env('LOG_LEVEL')
@@ -208,6 +211,12 @@ class Config:
        self.AZURE_BLOB_ACCOUNT_KEY = get_env('AZURE_BLOB_ACCOUNT_KEY')
        self.AZURE_BLOB_CONTAINER_NAME = get_env('AZURE_BLOB_CONTAINER_NAME')
        self.AZURE_BLOB_ACCOUNT_URL = get_env('AZURE_BLOB_ACCOUNT_URL')
+        self.ALIYUN_OSS_BUCKET_NAME=get_env('ALIYUN_OSS_BUCKET_NAME')
+        self.ALIYUN_OSS_ACCESS_KEY=get_env('ALIYUN_OSS_ACCESS_KEY')
+        self.ALIYUN_OSS_SECRET_KEY=get_env('ALIYUN_OSS_SECRET_KEY')
+        self.ALIYUN_OSS_ENDPOINT=get_env('ALIYUN_OSS_ENDPOINT')
+        self.GOOGLE_STORAGE_BUCKET_NAME = get_env('GOOGLE_STORAGE_BUCKET_NAME')
+        self.GOOGLE_STORAGE_SERVICE_ACCOUNT_JSON_BASE64 = get_env('GOOGLE_STORAGE_SERVICE_ACCOUNT_JSON_BASE64')

        # ------------------------
        # Vector Store Configurations.
@@ -219,6 +228,8 @@ class Config:
        self.QDRANT_URL = get_env('QDRANT_URL')
        self.QDRANT_API_KEY = get_env('QDRANT_API_KEY')
        self.QDRANT_CLIENT_TIMEOUT = get_env('QDRANT_CLIENT_TIMEOUT')
+        self.QDRANT_GRPC_ENABLED = get_env('QDRANT_GRPC_ENABLED')
+        self.QDRANT_GRPC_PORT = get_env('QDRANT_GRPC_PORT')

        # milvus / zilliz setting
        self.MILVUS_HOST = get_env('MILVUS_HOST')
@@ -241,6 +252,13 @@ class Config:
        self.RELYT_PASSWORD = get_env('RELYT_PASSWORD')
        self.RELYT_DATABASE = get_env('RELYT_DATABASE')

+        # pgvecto rs settings
+        self.PGVECTO_RS_HOST = get_env('PGVECTO_RS_HOST')
+        self.PGVECTO_RS_PORT = get_env('PGVECTO_RS_PORT')
+        self.PGVECTO_RS_USER = get_env('PGVECTO_RS_USER')
+        self.PGVECTO_RS_PASSWORD = get_env('PGVECTO_RS_PASSWORD')
+        self.PGVECTO_RS_DATABASE = get_env('PGVECTO_RS_DATABASE')
+
        # ------------------------
        # Mail Configurations.
        # ------------------------
@@ -256,7 +274,7 @@ class Config:
        self.SMTP_USE_TLS = get_bool_env('SMTP_USE_TLS')
        
        # ------------------------
-        # Workpace Configurations.
+        # Workspace Configurations.
        # ------------------------
        self.INVITE_EXPIRY_HOURS = int(get_env('INVITE_EXPIRY_HOURS'))

@@ -295,6 +313,12 @@ class Config:
        # ------------------------
        # Platform Configurations.
        # ------------------------
+        self.GITHUB_CLIENT_ID = get_env('GITHUB_CLIENT_ID')
+        self.GITHUB_CLIENT_SECRET = get_env('GITHUB_CLIENT_SECRET')
+        self.GOOGLE_CLIENT_ID = get_env('GOOGLE_CLIENT_ID')
+        self.GOOGLE_CLIENT_SECRET = get_env('GOOGLE_CLIENT_SECRET')
+        self.OAUTH_REDIRECT_PATH = get_env('OAUTH_REDIRECT_PATH')
+
        self.HOSTED_OPENAI_API_KEY = get_env('HOSTED_OPENAI_API_KEY')
        self.HOSTED_OPENAI_API_BASE = get_env('HOSTED_OPENAI_API_BASE')
        self.HOSTED_OPENAI_API_ORGANIZATION = get_env('HOSTED_OPENAI_API_ORGANIZATION')
@@ -341,17 +365,3 @@ class Config:

        self.KEYWORD_DATA_SOURCE_TYPE = get_env('KEYWORD_DATA_SOURCE_TYPE')
        self.ENTERPRISE_ENABLED = get_bool_env('ENTERPRISE_ENABLED')
-
-
-class CloudEditionConfig(Config):
-
-    def __init__(self):
-        super().__init__()
-
-        self.EDITION = "CLOUD"
-
-        self.GITHUB_CLIENT_ID = get_env('GITHUB_CLIENT_ID')
-        self.GITHUB_CLIENT_SECRET = get_env('GITHUB_CLIENT_SECRET')
-        self.GOOGLE_CLIENT_ID = get_env('GOOGLE_CLIENT_ID')
-        self.GOOGLE_CLIENT_SECRET = get_env('GOOGLE_CLIENT_SECRET')
-        self.OAUTH_REDIRECT_PATH = get_env('OAUTH_REDIRECT_PATH')
@@ -1,10 +1,11 @@


-languages = ['en-US', 'zh-Hans', 'pt-BR', 'es-ES', 'fr-FR', 'de-DE', 'ja-JP', 'ko-KR', 'ru-RU', 'it-IT', 'uk-UA', 'vi-VN']
+languages = ['en-US', 'zh-Hans', 'zh-Hant', 'pt-BR', 'es-ES', 'fr-FR', 'de-DE', 'ja-JP', 'ko-KR', 'ru-RU', 'it-IT', 'uk-UA', 'vi-VN']

 language_timezone_mapping = {
    'en-US': 'America/New_York',
    'zh-Hans': 'Asia/Shanghai',
+    'zh-Hant': 'Asia/Taipei',
    'pt-BR': 'America/Sao_Paulo',
    'es-ES': 'Europe/Madrid',
    'fr-FR': 'Europe/Paris',
@@ -53,5 +53,8 @@ from .explore import (
    workflow,
 )

+# Import tag controllers
+from .tag import tags
+
 # Import workspace controllers
 from .workspace import account, members, model_providers, models, tool_providers, workspace
@@ -1,17 +1,16 @@
 import json
+import uuid

 from flask_login import current_user
-from flask_restful import Resource, inputs, marshal_with, reqparse
-from werkzeug.exceptions import BadRequest, Forbidden
+from flask_restful import Resource, inputs, marshal, marshal_with, reqparse
+from werkzeug.exceptions import BadRequest, Forbidden, abort

 from controllers.console import api
 from controllers.console.app.wraps import get_app_model
 from controllers.console.setup import setup_required
 from controllers.console.wraps import account_initialization_required, cloud_edition_billing_resource_check
-from core.agent.entities import AgentToolEntity
 from core.tools.tool_manager import ToolManager
 from core.tools.utils.configuration import ToolParameterConfigurationManager
-from extensions.ext_database import db
 from fields.app_fields import (
    app_detail_fields,
    app_detail_fields_with_site,
@@ -20,6 +19,7 @@ from fields.app_fields import (
 from libs.login import login_required
 from models.model import App, AppMode, AppModelConfig
 from services.app_service import AppService
+from services.tag_service import TagService

 ALLOW_CREATE_APP_MODES = ['chat', 'agent-chat', 'advanced-chat', 'workflow', 'completion']

@@ -29,21 +29,29 @@ class AppListApi(Resource):
    @setup_required
    @login_required
    @account_initialization_required
-    @marshal_with(app_pagination_fields)
    def get(self):
        """Get app list"""
+        def uuid_list(value):
+            try:
+                return [str(uuid.UUID(v)) for v in value.split(',')]
+            except ValueError:
+                abort(400, message="Invalid UUID format in tag_ids.")
        parser = reqparse.RequestParser()
        parser.add_argument('page', type=inputs.int_range(1, 99999), required=False, default=1, location='args')
        parser.add_argument('limit', type=inputs.int_range(1, 100), required=False, default=20, location='args')
        parser.add_argument('mode', type=str, choices=['chat', 'workflow', 'agent-chat', 'channel', 'all'], default='all', location='args', required=False)
        parser.add_argument('name', type=str, location='args', required=False)
+        parser.add_argument('tag_ids', type=uuid_list, location='args', required=False)
+
        args = parser.parse_args()

        # get app list
        app_service = AppService()
        app_pagination = app_service.get_paginate_apps(current_user.current_tenant_id, args)
+        if not app_pagination:
+            return {'data': [], 'total': 0, 'page': 1, 'limit': 20, 'has_more': False}

-        return app_pagination
+        return marshal(app_pagination, app_pagination_fields)

    @setup_required
    @login_required
@@ -108,43 +116,9 @@ class AppApi(Resource):
    @marshal_with(app_detail_fields_with_site)
    def get(self, app_model):
        """Get app detail"""
-        # get original app model config
-        if app_model.mode == AppMode.AGENT_CHAT.value or app_model.is_agent:
-            model_config: AppModelConfig = app_model.app_model_config
-            agent_mode = model_config.agent_mode_dict
-            # decrypt agent tool parameters if it's secret-input
-            for tool in agent_mode.get('tools') or []:
-                if not isinstance(tool, dict) or len(tool.keys()) <= 3:
-                    continue
-                agent_tool_entity = AgentToolEntity(**tool)
-                # get tool
-                try:
-                    tool_runtime = ToolManager.get_agent_tool_runtime(
-                        tenant_id=current_user.current_tenant_id,
-                        agent_tool=agent_tool_entity,
-                    )
-                    manager = ToolParameterConfigurationManager(
-                        tenant_id=current_user.current_tenant_id,
-                        tool_runtime=tool_runtime,
-                        provider_name=agent_tool_entity.provider_id,
-                        provider_type=agent_tool_entity.provider_type,
-                    )
+        app_service = AppService()

-                    # get decrypted parameters
-                    if agent_tool_entity.tool_parameters:
-                        parameters = manager.decrypt_tool_parameters(agent_tool_entity.tool_parameters or {})
-                        masked_parameter = manager.mask_tool_parameters(parameters or {})
-                    else:
-                        masked_parameter = {}
-
-                    # override tool parameters
-                    tool['tool_parameters'] = masked_parameter
-                except Exception as e:
-                    pass
-
-            # override agent mode
-            model_config.agent_mode = json.dumps(agent_mode)
-            db.session.commit()
+        app_model = app_service.get_app(app_model)

        return app_model

@@ -57,6 +57,7 @@ class ModelConfigResource(Resource):
                try:
                    tool_runtime = ToolManager.get_agent_tool_runtime(
                        tenant_id=current_user.current_tenant_id,
+                        app_id=app_model.id,
                        agent_tool=agent_tool_entity,
                    )
                    manager = ToolParameterConfigurationManager(
@@ -64,6 +65,7 @@ class ModelConfigResource(Resource):
                        tool_runtime=tool_runtime,
                        provider_name=agent_tool_entity.provider_id,
                        provider_type=agent_tool_entity.provider_type,
+                        identity_id=f'AGENT.{app_model.id}'
                    )
                except Exception as e:
                    continue
@@ -94,6 +96,7 @@ class ModelConfigResource(Resource):
                    try:
                        tool_runtime = ToolManager.get_agent_tool_runtime(
                            tenant_id=current_user.current_tenant_id,
+                            app_id=app_model.id,
                            agent_tool=agent_tool_entity,
                        )
                    except Exception as e:
@@ -104,6 +107,7 @@ class ModelConfigResource(Resource):
                    tool_runtime=tool_runtime,
                    provider_name=agent_tool_entity.provider_id,
                    provider_type=agent_tool_entity.provider_type,
+                    identity_id=f'AGENT.{app_model.id}'
                )
                manager.delete_tool_parameters_cache()

@@ -111,9 +115,11 @@ class ModelConfigResource(Resource):
                if agent_tool_entity.tool_parameters:
                    if key not in masked_parameter_map:
                        continue
-
-                    if agent_tool_entity.tool_parameters == masked_parameter_map[key]:
-                        agent_tool_entity.tool_parameters = parameter_map[key]
+                    
+                    for masked_key, masked_value in masked_parameter_map[key].items():
+                        if masked_key in agent_tool_entity.tool_parameters and \
+                                agent_tool_entity.tool_parameters[masked_key] == masked_value:
+                            agent_tool_entity.tool_parameters[masked_key] = parameter_map[key].get(masked_key)

                # encrypt parameters
                if agent_tool_entity.tool_parameters:
@@ -48,11 +48,14 @@ class DatasetListApi(Resource):
        limit = request.args.get('limit', default=20, type=int)
        ids = request.args.getlist('ids')
        provider = request.args.get('provider', default="vendor")
+        search = request.args.get('keyword', default=None, type=str)
+        tag_ids = request.args.getlist('tag_ids')
+
        if ids:
            datasets, total = DatasetService.get_datasets_by_ids(ids, current_user.current_tenant_id)
        else:
            datasets, total = DatasetService.get_datasets(page, limit, provider,
-                                                          current_user.current_tenant_id, current_user)
+                                                          current_user.current_tenant_id, current_user, search, tag_ids)

        # check embedding setting
        provider_manager = ProviderManager()
@@ -184,6 +187,10 @@ class DatasetApi(Resource):
                            help='Invalid indexing technique.')
        parser.add_argument('permission', type=str, location='json', choices=(
            'only_me', 'all_team_members'), help='Invalid permission.')
+        parser.add_argument('embedding_model', type=str,
+                            location='json', help='Invalid embedding model.')
+        parser.add_argument('embedding_model_provider', type=str,
+                            location='json', help='Invalid embedding model provider.')
        parser.add_argument('retrieval_model', type=dict, location='json', help='Invalid retrieval model.')
        args = parser.parse_args()

@@ -469,7 +476,7 @@ class DatasetRetrievalSettingApi(Resource):
    @account_initialization_required
    def get(self):
        vector_type = current_app.config['VECTOR_STORE']
-        if vector_type == 'milvus':
+        if vector_type == 'milvus' or vector_type == 'pgvecto_rs' or vector_type == 'relyt':
            return {
                'retrieval_method': [
                    'semantic_search'
@@ -491,7 +498,7 @@ class DatasetRetrievalSettingMockApi(Resource):
    @account_initialization_required
    def get(self, vector_type):

-        if vector_type == 'milvus':
+        if vector_type == 'milvus' or vector_type == 'relyt':
            return {
                'retrieval_method': [
                    'semantic_search'
@@ -506,10 +513,27 @@ class DatasetRetrievalSettingMockApi(Resource):
        else:
            raise ValueError("Unsupported vector db type.")

+class DatasetErrorDocs(Resource):
+    @setup_required
+    @login_required
+    @account_initialization_required
+    def get(self, dataset_id):
+        dataset_id_str = str(dataset_id)
+        dataset = DatasetService.get_dataset(dataset_id_str)
+        if dataset is None:
+            raise NotFound("Dataset not found.")
+        results = DocumentService.get_error_documents_by_dataset_id(dataset_id_str)
+
+        return {
+            'data': [marshal(item, document_status_fields) for item in results],
+            'total': len(results)
+        }, 200
+

 api.add_resource(DatasetListApi, '/datasets')
 api.add_resource(DatasetApi, '/datasets/<uuid:dataset_id>')
 api.add_resource(DatasetQueryApi, '/datasets/<uuid:dataset_id>/queries')
+api.add_resource(DatasetErrorDocs, '/datasets/<uuid:dataset_id>/error-docs')
 api.add_resource(DatasetIndexingEstimateApi, '/datasets/indexing-estimate')
 api.add_resource(DatasetRelatedAppListApi, '/datasets/<uuid:dataset_id>/related-apps')
 api.add_resource(DatasetIndexingStatusApi, '/datasets/<uuid:dataset_id>/indexing-status')
@@ -1,3 +1,4 @@
+import logging
 from datetime import datetime, timezone

 from flask import request
@@ -233,7 +234,7 @@ class DatasetDocumentListApi(Resource):
                            location='json')
        parser.add_argument('data_source', type=dict, required=False, location='json')
        parser.add_argument('process_rule', type=dict, required=False, location='json')
-        parser.add_argument('duplicate', type=bool, nullable=False, location='json')
+        parser.add_argument('duplicate', type=bool, default=True, nullable=False, location='json')
        parser.add_argument('original_document_id', type=str, required=False, location='json')
        parser.add_argument('doc_form', type=str, default='text_model', required=False, nullable=False, location='json')
        parser.add_argument('doc_language', type=str, default='English', required=False, nullable=False,
@@ -393,9 +394,6 @@ class DocumentBatchIndexingEstimateApi(DocumentResource):
    def get(self, dataset_id, batch):
        dataset_id = str(dataset_id)
        batch = str(batch)
-        dataset = DatasetService.get_dataset(dataset_id)
-        if dataset is None:
-            raise NotFound("Dataset not found.")
        documents = self.get_batch_documents(dataset_id, batch)
        response = {
            "tokens": 0,
@@ -883,6 +881,49 @@ class DocumentRecoverApi(DocumentResource):
        return {'result': 'success'}, 204


+class DocumentRetryApi(DocumentResource):
+    @setup_required
+    @login_required
+    @account_initialization_required
+    def post(self, dataset_id):
+        """retry document."""
+
+        parser = reqparse.RequestParser()
+        parser.add_argument('document_ids', type=list, required=True, nullable=False,
+                            location='json')
+        args = parser.parse_args()
+        dataset_id = str(dataset_id)
+        dataset = DatasetService.get_dataset(dataset_id)
+        retry_documents = []
+        if not dataset:
+            raise NotFound('Dataset not found.')
+        for document_id in args['document_ids']:
+            try:
+                document_id = str(document_id)
+
+                document = DocumentService.get_document(dataset.id, document_id)
+
+                # 404 if document not found
+                if document is None:
+                    raise NotFound("Document Not Exists.")
+
+                # 403 if document is archived
+                if DocumentService.check_archived(document):
+                    raise ArchivedDocumentImmutableError()
+
+                # 400 if document is completed
+                if document.indexing_status == 'completed':
+                    raise DocumentAlreadyFinishedError()
+                retry_documents.append(document)
+            except Exception as e:
+                logging.error(f"Document {document_id} retry failed: {str(e)}")
+                continue
+        # retry document
+        DocumentService.retry_document(dataset_id, retry_documents)
+
+        return {'result': 'success'}, 204
+
+
 api.add_resource(GetProcessRuleApi, '/datasets/process-rule')
 api.add_resource(DatasetDocumentListApi,
                 '/datasets/<uuid:dataset_id>/documents')
@@ -908,3 +949,4 @@ api.add_resource(DocumentStatusApi,
                 '/datasets/<uuid:dataset_id>/documents/<uuid:document_id>/status/<string:action>')
 api.add_resource(DocumentPauseApi, '/datasets/<uuid:dataset_id>/documents/<uuid:document_id>/processing/pause')
 api.add_resource(DocumentRecoverApi, '/datasets/<uuid:dataset_id>/documents/<uuid:document_id>/processing/resume')
+api.add_resource(DocumentRetryApi, '/datasets/<uuid:dataset_id>/retry')
@@ -0,0 +1,159 @@
+from flask import request
+from flask_login import current_user
+from flask_restful import Resource, marshal_with, reqparse
+from werkzeug.exceptions import Forbidden
+
+from controllers.console import api
+from controllers.console.setup import setup_required
+from controllers.console.wraps import account_initialization_required
+from fields.tag_fields import tag_fields
+from libs.login import login_required
+from models.model import Tag
+from services.tag_service import TagService
+
+
+def _validate_name(name):
+    if not name or len(name) < 1 or len(name) > 40:
+        raise ValueError('Name must be between 1 to 50 characters.')
+    return name
+
+
+class TagListApi(Resource):
+
+    @setup_required
+    @login_required
+    @account_initialization_required
+    @marshal_with(tag_fields)
+    def get(self):
+        tag_type = request.args.get('type', type=str)
+        keyword = request.args.get('keyword', default=None, type=str)
+        tags = TagService.get_tags(tag_type, current_user.current_tenant_id, keyword)
+
+        return tags, 200
+
+    @setup_required
+    @login_required
+    @account_initialization_required
+    def post(self):
+        # The role of the current user in the ta table must be admin or owner
+        if not current_user.is_admin_or_owner:
+            raise Forbidden()
+
+        parser = reqparse.RequestParser()
+        parser.add_argument('name', nullable=False, required=True,
+                            help='Name must be between 1 to 50 characters.',
+                            type=_validate_name)
+        parser.add_argument('type', type=str, location='json',
+                            choices=Tag.TAG_TYPE_LIST,
+                            nullable=True,
+                            help='Invalid tag type.')
+        args = parser.parse_args()
+        tag = TagService.save_tags(args)
+
+        response = {
+            'id': tag.id,
+            'name': tag.name,
+            'type': tag.type,
+            'binding_count': 0
+        }
+
+        return response, 200
+
+
+class TagUpdateDeleteApi(Resource):
+
+    @setup_required
+    @login_required
+    @account_initialization_required
+    def patch(self, tag_id):
+        tag_id = str(tag_id)
+        # The role of the current user in the ta table must be admin or owner
+        if not current_user.is_admin_or_owner:
+            raise Forbidden()
+
+        parser = reqparse.RequestParser()
+        parser.add_argument('name', nullable=False, required=True,
+                            help='Name must be between 1 to 50 characters.',
+                            type=_validate_name)
+        args = parser.parse_args()
+        tag = TagService.update_tags(args, tag_id)
+
+        binding_count = TagService.get_tag_binding_count(tag_id)
+
+        response = {
+            'id': tag.id,
+            'name': tag.name,
+            'type': tag.type,
+            'binding_count': binding_count
+        }
+
+        return response, 200
+
+    @setup_required
+    @login_required
+    @account_initialization_required
+    def delete(self, tag_id):
+        tag_id = str(tag_id)
+        # The role of the current user in the ta table must be admin or owner
+        if not current_user.is_admin_or_owner:
+            raise Forbidden()
+
+        TagService.delete_tag(tag_id)
+
+        return 200
+
+
+class TagBindingCreateApi(Resource):
+
+    @setup_required
+    @login_required
+    @account_initialization_required
+    def post(self):
+        # The role of the current user in the ta table must be admin or owner
+        if not current_user.is_admin_or_owner:
+            raise Forbidden()
+
+        parser = reqparse.RequestParser()
+        parser.add_argument('tag_ids', type=list, nullable=False, required=True, location='json',
+                            help='Tag IDs is required.')
+        parser.add_argument('target_id', type=str, nullable=False, required=True, location='json',
+                            help='Target ID is required.')
+        parser.add_argument('type', type=str, location='json',
+                            choices=Tag.TAG_TYPE_LIST,
+                            nullable=True,
+                            help='Invalid tag type.')
+        args = parser.parse_args()
+        TagService.save_tag_binding(args)
+
+        return 200
+
+
+class TagBindingDeleteApi(Resource):
+
+    @setup_required
+    @login_required
+    @account_initialization_required
+    def post(self):
+        # The role of the current user in the ta table must be admin or owner
+        if not current_user.is_admin_or_owner:
+            raise Forbidden()
+
+        parser = reqparse.RequestParser()
+        parser.add_argument('tag_id', type=str, nullable=False, required=True,
+                            help='Tag ID is required.')
+        parser.add_argument('target_id', type=str, nullable=False, required=True,
+                            help='Target ID is required.')
+        parser.add_argument('type', type=str, location='json',
+                            choices=Tag.TAG_TYPE_LIST,
+                            nullable=True,
+                            help='Invalid tag type.')
+        args = parser.parse_args()
+        TagService.delete_tag_binding(args)
+
+        return 200
+
+
+api.add_resource(TagListApi, '/tags')
+api.add_resource(TagUpdateDeleteApi, '/tags/<uuid:tag_id>')
+api.add_resource(TagBindingCreateApi, '/tag-bindings/create')
+api.add_resource(TagBindingDeleteApi, '/tag-bindings/remove')
@@ -9,7 +9,7 @@ from controllers.console.wraps import account_initialization_required, cloud_edi
 from extensions.ext_database import db
 from fields.member_fields import account_with_role_list_fields
 from libs.login import login_required
-from models.account import Account
+from models.account import Account, TenantAccountRole
 from services.account_service import RegisterService, TenantService
 from services.errors.account import AccountAlreadyInTenantError

@@ -43,7 +43,7 @@ class MemberInviteEmailApi(Resource):
        invitee_emails = args['emails']
        invitee_role = args['role']
        interface_language = args['language']
-        if invitee_role not in ['admin', 'normal']:
+        if invitee_role not in [TenantAccountRole.ADMIN, TenantAccountRole.NORMAL]:
            return {'code': 'invalid-role', 'message': 'Invalid role'}, 400

        inviter = current_user
@@ -11,6 +11,7 @@ from core.model_runtime.entities.model_entities import ModelType
 from core.model_runtime.errors.validate import CredentialsValidateFailedError
 from core.model_runtime.utils.encoders import jsonable_encoder
 from libs.login import login_required
+from models.account import TenantAccountRole
 from services.model_provider_service import ModelProviderService


@@ -94,7 +95,7 @@ class ModelProviderModelApi(Resource):
    @login_required
    @account_initialization_required
    def post(self, provider: str):
-        if current_user.current_tenant.current_role not in ['admin', 'owner']:
+        if not TenantAccountRole.is_privileged_role(current_user.current_tenant.current_role):
            raise Forbidden()

        tenant_id = current_user.current_tenant_id
@@ -125,7 +126,7 @@ class ModelProviderModelApi(Resource):
    @login_required
    @account_initialization_required
    def delete(self, provider: str):
-        if current_user.current_tenant.current_role not in ['admin', 'owner']:
+        if not TenantAccountRole.is_privileged_role(current_user.current_tenant.current_role):
            raise Forbidden()

        tenant_id = current_user.current_tenant_id
@@ -26,8 +26,11 @@ class DatasetApi(DatasetApiResource):
        page = request.args.get('page', default=1, type=int)
        limit = request.args.get('limit', default=20, type=int)
        provider = request.args.get('provider', default="vendor")
+        search = request.args.get('keyword', default=None, type=str)
+        tag_ids = request.args.getlist('tag_ids')
+
        datasets, total = DatasetService.get_datasets(page, limit, provider,
-                                                      tenant_id, current_user)
+                                                      tenant_id, current_user, search, tag_ids)
        # check embedding setting
        provider_manager = ProviderManager()
        configurations = provider_manager.get_configurations(
@@ -163,6 +163,7 @@ class BaseAgentRunner(AppRunner):
        """
        tool_entity = ToolManager.get_agent_tool_runtime(
            tenant_id=self.tenant_id,
+            app_id=self.app_config.app_id,
            agent_tool=tool,
        )
        tool_entity.load_variables(self.variables_pool)
@@ -18,7 +18,7 @@ from core.workflow.entities.node_entities import SystemVariable
 from core.workflow.nodes.base_node import UserFrom
 from core.workflow.workflow_engine_manager import WorkflowEngineManager
 from extensions.ext_database import db
-from models.model import App, Conversation, Message
+from models.model import App, Conversation, EndUser, Message
 from models.workflow import Workflow

 logger = logging.getLogger(__name__)
@@ -56,6 +56,14 @@ class AdvancedChatAppRunner(AppRunner):
        query = application_generate_entity.query
        files = application_generate_entity.files

+        user_id = None
+        if application_generate_entity.invoke_from in [InvokeFrom.WEB_APP, InvokeFrom.SERVICE_API]:
+            end_user = db.session.query(EndUser).filter(EndUser.id == application_generate_entity.user_id).first()
+            if end_user:
+                user_id = end_user.session_id
+        else:
+            user_id = application_generate_entity.user_id
+
        # moderation
        if self.handle_input_moderation(
                queue_manager=queue_manager,
@@ -98,7 +106,8 @@ class AdvancedChatAppRunner(AppRunner):
            system_inputs={
                SystemVariable.QUERY: query,
                SystemVariable.FILES: files,
-                SystemVariable.CONVERSATION: conversation.id,
+                SystemVariable.CONVERSATION_ID: conversation.id,
+                SystemVariable.USER_ID: user_id
            },
            callbacks=workflow_callbacks
        )
@@ -28,9 +28,9 @@ from core.app.entities.task_entities import (
    AdvancedChatTaskState,
    ChatbotAppBlockingResponse,
    ChatbotAppStreamResponse,
+    ChatflowStreamGenerateRoute,
    ErrorStreamResponse,
    MessageEndStreamResponse,
-    StreamGenerateRoute,
    StreamResponse,
 )
 from core.app.task_pipeline.based_generate_task_pipeline import BasedGenerateTaskPipeline
@@ -84,13 +84,19 @@ class AdvancedChatAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCyc
        """
        super().__init__(application_generate_entity, queue_manager, user, stream)

+        if isinstance(self._user, EndUser):
+            user_id = self._user.session_id
+        else:
+            user_id = self._user.id
+
        self._workflow = workflow
        self._conversation = conversation
        self._message = message
        self._workflow_system_variables = {
            SystemVariable.QUERY: message.query,
            SystemVariable.FILES: application_generate_entity.files,
-            SystemVariable.CONVERSATION: conversation.id,
+            SystemVariable.CONVERSATION_ID: conversation.id,
+            SystemVariable.USER_ID: user_id
        }

        self._task_state = AdvancedChatTaskState(
@@ -337,7 +343,7 @@ class AdvancedChatAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCyc
            **extras
        )

-    def _get_stream_generate_routes(self) -> dict[str, StreamGenerateRoute]:
+    def _get_stream_generate_routes(self) -> dict[str, ChatflowStreamGenerateRoute]:
        """
        Get stream generate routes.
        :return:
@@ -360,7 +366,7 @@ class AdvancedChatAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCyc
                continue

            for start_node_id in start_node_ids:
-                stream_generate_routes[start_node_id] = StreamGenerateRoute(
+                stream_generate_routes[start_node_id] = ChatflowStreamGenerateRoute(
                    answer_node_id=answer_node_id,
                    generate_route=generate_route
                )
@@ -424,15 +430,14 @@ class AdvancedChatAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCyc
            for route_chunk in route_chunks:
                if route_chunk.type == 'text':
                    route_chunk = cast(TextGenerateRouteChunk, route_chunk)
-                    for token in route_chunk.text:
-                        # handle output moderation chunk
-                        should_direct_answer = self._handle_output_moderation_chunk(token)
-                        if should_direct_answer:
-                            continue

-                        self._task_state.answer += token
-                        yield self._message_to_stream_response(token, self._message.id)
-                        time.sleep(0.01)
+                    # handle output moderation chunk
+                    should_direct_answer = self._handle_output_moderation_chunk(route_chunk.text)
+                    if should_direct_answer:
+                        continue
+
+                    self._task_state.answer += route_chunk.text
+                    yield self._message_to_stream_response(route_chunk.text, self._message.id)
                else:
                    break

@@ -457,10 +462,8 @@ class AdvancedChatAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCyc
        for route_chunk in route_chunks:
            if route_chunk.type == 'text':
                route_chunk = cast(TextGenerateRouteChunk, route_chunk)
-                for token in route_chunk.text:
-                    self._task_state.answer += token
-                    yield self._message_to_stream_response(token, self._message.id)
-                    time.sleep(0.01)
+                self._task_state.answer += route_chunk.text
+                yield self._message_to_stream_response(route_chunk.text, self._message.id)
            else:
                route_chunk = cast(VarGenerateRouteChunk, route_chunk)
                value_selector = route_chunk.value_selector
@@ -23,20 +23,28 @@ class BaseAppGenerator:
            value = user_inputs[variable]

            if value:
-                if not isinstance(value, str):
+                if variable_config.type != VariableEntity.Type.NUMBER and not isinstance(value, str):
                    raise ValueError(f"{variable} in input form must be a string")
+                elif variable_config.type == VariableEntity.Type.NUMBER and isinstance(value, str):
+                    if '.' in value:
+                        value = float(value)
+                    else:
+                        value = int(value)

            if variable_config.type == VariableEntity.Type.SELECT:
                options = variable_config.options if variable_config.options is not None else []
                if value not in options:
                    raise ValueError(f"{variable} in input form must be one of the following: {options}")
-            else:
+            elif variable_config.type in [VariableEntity.Type.TEXT_INPUT, VariableEntity.Type.PARAGRAPH]:
                if variable_config.max_length is not None:
                    max_length = variable_config.max_length
                    if len(value) > max_length:
                        raise ValueError(f'{variable} in input form must be less than {max_length} characters')

-            filtered_inputs[variable] = value.replace('\x00', '') if value else None
+            if value and isinstance(value, str):
+                filtered_inputs[variable] = value.replace('\x00', '')
+            else:
+                filtered_inputs[variable] = value if value else None

        return filtered_inputs

@@ -14,7 +14,7 @@ from core.workflow.entities.node_entities import SystemVariable
 from core.workflow.nodes.base_node import UserFrom
 from core.workflow.workflow_engine_manager import WorkflowEngineManager
 from extensions.ext_database import db
-from models.model import App
+from models.model import App, EndUser
 from models.workflow import Workflow

 logger = logging.getLogger(__name__)
@@ -36,6 +36,14 @@ class WorkflowAppRunner:
        app_config = application_generate_entity.app_config
        app_config = cast(WorkflowAppConfig, app_config)

+        user_id = None
+        if application_generate_entity.invoke_from in [InvokeFrom.WEB_APP, InvokeFrom.SERVICE_API]:
+            end_user = db.session.query(EndUser).filter(EndUser.id == application_generate_entity.user_id).first()
+            if end_user:
+                user_id = end_user.session_id
+        else:
+            user_id = application_generate_entity.user_id
+
        app_record = db.session.query(App).filter(App.id == app_config.app_id).first()
        if not app_record:
            raise ValueError("App not found")
@@ -67,7 +75,8 @@ class WorkflowAppRunner:
            else UserFrom.END_USER,
            user_inputs=inputs,
            system_inputs={
-                SystemVariable.FILES: files
+                SystemVariable.FILES: files,
+                SystemVariable.USER_ID: user_id
            },
            callbacks=workflow_callbacks
        )
@@ -28,11 +28,13 @@ from core.app.entities.task_entities import (
    WorkflowAppBlockingResponse,
    WorkflowAppStreamResponse,
    WorkflowFinishStreamResponse,
+    WorkflowStreamGenerateNodes,
    WorkflowTaskState,
 )
 from core.app.task_pipeline.based_generate_task_pipeline import BasedGenerateTaskPipeline
 from core.app.task_pipeline.workflow_cycle_manage import WorkflowCycleManage
-from core.workflow.entities.node_entities import SystemVariable
+from core.workflow.entities.node_entities import NodeType, SystemVariable
+from core.workflow.nodes.end.end_node import EndNode
 from extensions.ext_database import db
 from models.account import Account
 from models.model import EndUser
@@ -40,6 +42,7 @@ from models.workflow import (
    Workflow,
    WorkflowAppLog,
    WorkflowAppLogCreatedFrom,
+    WorkflowNodeExecution,
    WorkflowRun,
 )

@@ -71,12 +74,19 @@ class WorkflowAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCycleMa
        """
        super().__init__(application_generate_entity, queue_manager, user, stream)

+        if isinstance(self._user, EndUser):
+            user_id = self._user.session_id
+        else:
+            user_id = self._user.id
+
        self._workflow = workflow
        self._workflow_system_variables = {
            SystemVariable.FILES: application_generate_entity.files,
+            SystemVariable.USER_ID: user_id
        }

        self._task_state = WorkflowTaskState()
+        self._stream_generate_nodes = self._get_stream_generate_nodes()

    def process(self) -> Union[WorkflowAppBlockingResponse, Generator[WorkflowAppStreamResponse, None, None]]:
        """
@@ -161,6 +171,14 @@ class WorkflowAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCycleMa
                )
            elif isinstance(event, QueueNodeStartedEvent):
                workflow_node_execution = self._handle_node_start(event)
+
+                # search stream_generate_routes if node id is answer start at node
+                if not self._task_state.current_stream_generate_state and event.node_id in self._stream_generate_nodes:
+                    self._task_state.current_stream_generate_state = self._stream_generate_nodes[event.node_id]
+
+                    # generate stream outputs when node started
+                    yield from self._generate_stream_outputs_when_node_started()
+
                yield self._workflow_node_start_to_stream_response(
                    event=event,
                    task_id=self._application_generate_entity.task_id,
@@ -168,6 +186,7 @@ class WorkflowAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCycleMa
                )
            elif isinstance(event, QueueNodeSucceededEvent | QueueNodeFailedEvent):
                workflow_node_execution = self._handle_node_finished(event)
+
                yield self._workflow_node_finish_to_stream_response(
                    task_id=self._application_generate_entity.task_id,
                    workflow_node_execution=workflow_node_execution
@@ -187,6 +206,11 @@ class WorkflowAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCycleMa
                if delta_text is None:
                    continue

+                if not self._is_stream_out_support(
+                        event=event
+                ):
+                    continue
+
                self._task_state.answer += delta_text
                yield self._text_chunk_to_stream_response(delta_text)
            elif isinstance(event, QueueMessageReplaceEvent):
@@ -248,3 +272,142 @@ class WorkflowAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCycleMa
            task_id=self._application_generate_entity.task_id,
            text=TextReplaceStreamResponse.Data(text=text)
        )
+
+    def _get_stream_generate_nodes(self) -> dict[str, WorkflowStreamGenerateNodes]:
+        """
+        Get stream generate nodes.
+        :return:
+        """
+        # find all answer nodes
+        graph = self._workflow.graph_dict
+        end_node_configs = [
+            node for node in graph['nodes']
+            if node.get('data', {}).get('type') == NodeType.END.value
+        ]
+
+        # parse stream output node value selectors of end nodes
+        stream_generate_routes = {}
+        for node_config in end_node_configs:
+            # get generate route for stream output
+            end_node_id = node_config['id']
+            generate_nodes = EndNode.extract_generate_nodes(graph, node_config)
+            start_node_ids = self._get_end_start_at_node_ids(graph, end_node_id)
+            if not start_node_ids:
+                continue
+
+            for start_node_id in start_node_ids:
+                stream_generate_routes[start_node_id] = WorkflowStreamGenerateNodes(
+                    end_node_id=end_node_id,
+                    stream_node_ids=generate_nodes
+                )
+
+        return stream_generate_routes
+
+    def _get_end_start_at_node_ids(self, graph: dict, target_node_id: str) \
+            -> list[str]:
+        """
+        Get end start at node id.
+        :param graph: graph
+        :param target_node_id: target node ID
+        :return:
+        """
+        nodes = graph.get('nodes')
+        edges = graph.get('edges')
+
+        # fetch all ingoing edges from source node
+        ingoing_edges = []
+        for edge in edges:
+            if edge.get('target') == target_node_id:
+                ingoing_edges.append(edge)
+
+        if not ingoing_edges:
+            return []
+
+        start_node_ids = []
+        for ingoing_edge in ingoing_edges:
+            source_node_id = ingoing_edge.get('source')
+            source_node = next((node for node in nodes if node.get('id') == source_node_id), None)
+            if not source_node:
+                continue
+
+            node_type = source_node.get('data', {}).get('type')
+            if node_type in [
+                NodeType.IF_ELSE.value,
+                NodeType.QUESTION_CLASSIFIER.value
+            ]:
+                start_node_id = target_node_id
+                start_node_ids.append(start_node_id)
+            elif node_type == NodeType.START.value:
+                start_node_id = source_node_id
+                start_node_ids.append(start_node_id)
+            else:
+                sub_start_node_ids = self._get_end_start_at_node_ids(graph, source_node_id)
+                if sub_start_node_ids:
+                    start_node_ids.extend(sub_start_node_ids)
+
+        return start_node_ids
+
+    def _generate_stream_outputs_when_node_started(self) -> Generator:
+        """
+        Generate stream outputs.
+        :return:
+        """
+        if self._task_state.current_stream_generate_state:
+            stream_node_ids = self._task_state.current_stream_generate_state.stream_node_ids
+
+            for node_id, node_execution_info in self._task_state.ran_node_execution_infos.items():
+                if node_id not in stream_node_ids:
+                    continue
+
+                node_execution_info = self._task_state.ran_node_execution_infos[node_id]
+
+                # get chunk node execution
+                route_chunk_node_execution = db.session.query(WorkflowNodeExecution).filter(
+                    WorkflowNodeExecution.id == node_execution_info.workflow_node_execution_id).first()
+
+                if not route_chunk_node_execution:
+                    continue
+
+                outputs = route_chunk_node_execution.outputs_dict
+
+                if not outputs:
+                    continue
+
+                # get value from outputs
+                text = outputs.get('text')
+
+                if text:
+                    self._task_state.answer += text
+                    yield self._text_chunk_to_stream_response(text)
+
+            db.session.close()
+
+    def _is_stream_out_support(self, event: QueueTextChunkEvent) -> bool:
+        """
+        Is stream out support
+        :param event: queue text chunk event
+        :return:
+        """
+        if not event.metadata:
+            return False
+
+        if 'node_id' not in event.metadata:
+            return False
+
+        node_id = event.metadata.get('node_id')
+        node_type = event.metadata.get('node_type')
+        stream_output_value_selector = event.metadata.get('value_selector')
+        if not stream_output_value_selector:
+            return False
+
+        if not self._task_state.current_stream_generate_state:
+            return False
+
+        if node_id not in self._task_state.current_stream_generate_state.stream_node_ids:
+            return False
+
+        if node_type != NodeType.LLM:
+            # only LLM support chunk stream output
+            return False
+
+        return True
@@ -6,6 +6,7 @@ from core.app.entities.queue_entities import (
    QueueNodeFailedEvent,
    QueueNodeStartedEvent,
    QueueNodeSucceededEvent,
+    QueueTextChunkEvent,
    QueueWorkflowFailedEvent,
    QueueWorkflowStartedEvent,
    QueueWorkflowSucceededEvent,
@@ -119,7 +120,15 @@ class WorkflowEventTriggerCallback(BaseWorkflowCallback):
        """
        Publish text chunk
        """
-        pass
+        self._queue_manager.publish(
+            QueueTextChunkEvent(
+                text=text,
+                metadata={
+                    "node_id": node_id,
+                    **metadata
+                }
+            ), PublishFrom.APPLICATION_MANAGER
+        )

    def on_event(self, event: AppQueueEvent) -> None:
        """
@@ -72,7 +72,7 @@ class AppGenerateEntity(BaseModel):
    # app config
    app_config: AppConfig

-    inputs: dict[str, str]
+    inputs: dict[str, Any]
    files: list[FileVar] = []
    user_id: str

@@ -9,9 +9,17 @@ from core.workflow.entities.node_entities import NodeType
 from core.workflow.nodes.answer.entities import GenerateRouteChunk


-class StreamGenerateRoute(BaseModel):
+class WorkflowStreamGenerateNodes(BaseModel):
    """
-    StreamGenerateRoute entity
+    WorkflowStreamGenerateNodes entity
+    """
+    end_node_id: str
+    stream_node_ids: list[str]
+
+
+class ChatflowStreamGenerateRoute(BaseModel):
+    """
+    ChatflowStreamGenerateRoute entity
    """
    answer_node_id: str
    generate_route: list[GenerateRouteChunk]
@@ -55,6 +63,8 @@ class WorkflowTaskState(TaskState):
    ran_node_execution_infos: dict[str, NodeExecutionInfo] = {}
    latest_node_execution_info: Optional[NodeExecutionInfo] = None

+    current_stream_generate_state: Optional[WorkflowStreamGenerateNodes] = None
+

 class AdvancedChatTaskState(WorkflowTaskState):
    """
@@ -62,7 +72,7 @@ class AdvancedChatTaskState(WorkflowTaskState):
    """
    usage: LLMUsage

-    current_stream_generate_state: Optional[StreamGenerateRoute] = None
+    current_stream_generate_state: Optional[ChatflowStreamGenerateRoute] = None


 class StreamEvent(Enum):
@@ -118,7 +118,8 @@ class MessageCycleManage:
        :param event: event
        :return:
        """
-        self._task_state.metadata['retriever_resources'] = event.retriever_resources
+        if self._application_generate_entity.app_config.additional_features.show_retrieve_source:
+            self._task_state.metadata['retriever_resources'] = event.retriever_resources

    def _get_response_metadata(self) -> dict:
        """
@@ -6,7 +6,7 @@ from yarl import URL

 from config import get_env
 from core.helper.code_executor.javascript_transformer import NodeJsTemplateTransformer
-from core.helper.code_executor.jina2_transformer import Jinja2TemplateTransformer
+from core.helper.code_executor.jinja2_transformer import Jinja2TemplateTransformer
 from core.helper.code_executor.python_transformer import PythonTemplateTransformer

 # Code Executor
@@ -1,10 +1,13 @@
 import json
 import re
+from base64 import b64encode

 from core.helper.code_executor.template_transformer import TemplateTransformer

 PYTHON_RUNNER = """
 import jinja2
+from json import loads
+from base64 import b64decode

 template = jinja2.Template('''{{code}}''')

@@ -12,7 +15,8 @@ def main(**inputs):
    return template.render(**inputs)

 # execute main function, and return the result
-output = main(**{{inputs}})
+inputs = b64decode('{{inputs}}').decode('utf-8')
+output = main(**loads(inputs))

 result = f'''<<RESULT>>{output}<<RESULT>>'''

@@ -39,6 +43,7 @@ JINJA2_PRELOAD_TEMPLATE = """{% set fruits = ['Apple'] %}

 JINJA2_PRELOAD = f"""
 import jinja2
+from base64 import b64decode

 def _jinja2_preload_():
    # prepare jinja2 environment, load template and render before to avoid sandbox issue
@@ -50,6 +55,7 @@ if __name__ == '__main__':

 """

+
 class Jinja2TemplateTransformer(TemplateTransformer):
    @classmethod
    def transform_caller(cls, code: str, inputs: dict) -> tuple[str, str]:
@@ -60,9 +66,11 @@ class Jinja2TemplateTransformer(TemplateTransformer):
        :return:
        """

+        inputs_str = b64encode(json.dumps(inputs, ensure_ascii=False).encode()).decode('utf-8')
+
        # transform jinja2 template to python code
        runner = PYTHON_RUNNER.replace('{{code}}', code)
-        runner = runner.replace('{{inputs}}', json.dumps(inputs, indent=4, ensure_ascii=False))
+        runner = runner.replace('{{inputs}}', inputs_str)

        return runner, JINJA2_PRELOAD

@@ -1,17 +1,22 @@
 import json
 import re
+from base64 import b64encode

 from core.helper.code_executor.template_transformer import TemplateTransformer

 PYTHON_RUNNER = """# declare main function here
 {{code}}

+from json import loads, dumps
+from base64 import b64decode
+
 # execute main function, and return the result
 # inputs is a dict, and it
-output = main(**{{inputs}})
+inputs = b64decode('{{inputs}}').decode('utf-8')
+output = main(**json.loads(inputs))

 # convert output to json and print
-output = json.dumps(output, indent=4)
+output = dumps(output, indent=4)

 result = f'''<<RESULT>>
 {output}
@@ -54,7 +59,7 @@ class PythonTemplateTransformer(TemplateTransformer):
        """
        
        # transform inputs to json string
-        inputs_str = json.dumps(inputs, indent=4, ensure_ascii=False)
+        inputs_str = b64encode(json.dumps(inputs, ensure_ascii=False).encode()).decode('utf-8')

        # replace code and inputs
        runner = PYTHON_RUNNER.replace('{{code}}', code)
@@ -11,12 +11,13 @@ class ToolParameterCacheType(Enum):

 class ToolParameterCache:
    def __init__(self, 
-                 tenant_id: str, 
-                 provider: str, 
-                 tool_name: str, 
-                 cache_type: ToolParameterCacheType
+            tenant_id: str, 
+            provider: str, 
+            tool_name: str, 
+            cache_type: ToolParameterCacheType,
+            identity_id: str
        ):
-        self.cache_key = f"{cache_type.value}_secret:tenant_id:{tenant_id}:provider:{provider}:tool_name:{tool_name}"
+        self.cache_key = f"{cache_type.value}_secret:tenant_id:{tenant_id}:provider:{provider}:tool_name:{tool_name}:identity_id:{identity_id}"

    def get(self) -> Optional[dict]:
        """
@@ -8,5 +8,10 @@
 - anthropic.claude-3-haiku-v1:0
 - cohere.command-light-text-v14
 - cohere.command-text-v14
+- meta.llama3-8b-instruct-v1:0
+- meta.llama3-70b-instruct-v1:0
 - meta.llama2-13b-chat-v1
 - meta.llama2-70b-chat-v1
+- mistral.mistral-large-2402-v1:0
+- mistral.mixtral-8x7b-instruct-v0:1
+- mistral.mistral-7b-instruct-v0:2
@@ -370,29 +370,14 @@ class BedrockLargeLanguageModel(LargeLanguageModel):
        :return:md = genai.GenerativeModel(model)
        """
        prefix = model.split('.')[0]
-
+        model_name = model.split('.')[1]
        if isinstance(messages, str):
            prompt = messages
        else:
-            prompt = self._convert_messages_to_prompt(messages, prefix)
+            prompt = self._convert_messages_to_prompt(messages, prefix, model_name)

        return self._get_num_tokens_by_gpt2(prompt)
    
-    def _convert_messages_to_prompt(self, model_prefix: str, messages: list[PromptMessage]) -> str:
-        """
-        Format a list of messages into a full prompt for the Google model
-
-        :param messages: List of PromptMessage to combine.
-        :return: Combined string with necessary human_prompt and ai_prompt tags.
-        """
-        messages = messages.copy()  # don't mutate the original list
-        
-        text = "".join(
-            self._convert_one_message_to_text(message, model_prefix)
-            for message in messages
-        )
-
-        return text.rstrip()

    def validate_credentials(self, model: str, credentials: dict) -> None:
        """
@@ -432,7 +417,7 @@ class BedrockLargeLanguageModel(LargeLanguageModel):
        except Exception as ex:
            raise CredentialsValidateFailedError(str(ex))

-    def _convert_one_message_to_text(self, message: PromptMessage, model_prefix: str) -> str:
+    def _convert_one_message_to_text(self, message: PromptMessage, model_prefix: str, model_name: Optional[str] = None) -> str:
        """
        Convert a single message to a string.

@@ -446,9 +431,21 @@ class BedrockLargeLanguageModel(LargeLanguageModel):
            ai_prompt = "\n\nAssistant:"

        elif model_prefix == "meta":
-            human_prompt_prefix = "\n[INST]"
+            # LLAMA3
+            if model_name.startswith("llama3"):
+                human_prompt_prefix = "<|eot_id|><|start_header_id|>user<|end_header_id|>\n\n"
+                human_prompt_postfix = "<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\n"
+                ai_prompt = "\n\nAssistant:"
+            else:
+                # LLAMA2
+                human_prompt_prefix = "\n[INST]"
+                human_prompt_postfix = "[\\INST]\n"
+                ai_prompt = ""
+
+        elif model_prefix == "mistral":
+            human_prompt_prefix = "<s>[INST]"
            human_prompt_postfix = "[\\INST]\n"
-            ai_prompt = ""
+            ai_prompt = "\n\nAssistant:"

        elif model_prefix == "amazon":
            human_prompt_prefix = "\n\nUser:"
@@ -473,11 +470,12 @@ class BedrockLargeLanguageModel(LargeLanguageModel):

        return message_text

-    def _convert_messages_to_prompt(self, messages: list[PromptMessage], model_prefix: str) -> str:
+    def _convert_messages_to_prompt(self, messages: list[PromptMessage], model_prefix: str, model_name: Optional[str] = None) -> str:
        """
        Format a list of messages into a full prompt for the Anthropic, Amazon and Llama models

        :param messages: List of PromptMessage to combine.
+        :param model_name: specific model name.Optional,just to distinguish llama2 and llama3
        :return: Combined string with necessary human_prompt and ai_prompt tags.
        """
        if not messages:
@@ -488,18 +486,20 @@ class BedrockLargeLanguageModel(LargeLanguageModel):
            messages.append(AssistantPromptMessage(content=""))

        text = "".join(
-            self._convert_one_message_to_text(message, model_prefix)
+            self._convert_one_message_to_text(message, model_prefix, model_name)
            for message in messages
        )

        # trim off the trailing ' ' that might come from the "Assistant: "
        return text.rstrip()

-    def _create_payload(self, model_prefix: str, prompt_messages: list[PromptMessage], model_parameters: dict, stop: Optional[list[str]] = None, stream: bool = True):
+    def _create_payload(self, model: str, prompt_messages: list[PromptMessage], model_parameters: dict, stop: Optional[list[str]] = None, stream: bool = True):
        """
        Create payload for bedrock api call depending on model provider
        """
        payload = dict()
+        model_prefix = model.split('.')[0]
+        model_name = model.split('.')[1]

        if model_prefix == "amazon":
            payload["textGenerationConfig"] = { **model_parameters }
@@ -519,6 +519,13 @@ class BedrockLargeLanguageModel(LargeLanguageModel):
                payload["frequencyPenalty"] = {model_parameters.get("frequencyPenalty")}
            if model_parameters.get("countPenalty"):
                payload["countPenalty"] = {model_parameters.get("countPenalty")}
+        
+        elif model_prefix == "mistral":
+            payload["temperature"] = model_parameters.get("temperature")
+            payload["top_p"] = model_parameters.get("top_p")
+            payload["max_tokens"] = model_parameters.get("max_tokens")
+            payload["prompt"] = self._convert_messages_to_prompt(prompt_messages, model_prefix)
+            payload["stop"] = stop[:10] if stop else []

        elif model_prefix == "anthropic":
            payload = { **model_parameters }
@@ -532,7 +539,7 @@ class BedrockLargeLanguageModel(LargeLanguageModel):
        
        elif model_prefix == "meta":
            payload = { **model_parameters }
-            payload["prompt"] = self._convert_messages_to_prompt(prompt_messages, model_prefix)
+            payload["prompt"] = self._convert_messages_to_prompt(prompt_messages, model_prefix, model_name)

        else:
            raise ValueError(f"Got unknown model prefix {model_prefix}")
@@ -567,7 +574,7 @@ class BedrockLargeLanguageModel(LargeLanguageModel):
        )

        model_prefix = model.split('.')[0]
-        payload = self._create_payload(model_prefix, prompt_messages, model_parameters, stop, stream)
+        payload = self._create_payload(model, prompt_messages, model_parameters, stop, stream)

        # need workaround for ai21 models which doesn't support streaming
        if stream and model_prefix != "ai21":
@@ -648,6 +655,11 @@ class BedrockLargeLanguageModel(LargeLanguageModel):
            output = response_body.get("generation").strip('\n')
            prompt_tokens = response_body.get("prompt_token_count")
            completion_tokens = response_body.get("generation_token_count")
+        
+        elif model_prefix == "mistral":
+            output = response_body.get("outputs")[0].get("text")
+            prompt_tokens = response.get('ResponseMetadata').get('HTTPHeaders').get('x-amzn-bedrock-input-token-count')
+            completion_tokens = response.get('ResponseMetadata').get('HTTPHeaders').get('x-amzn-bedrock-output-token-count')

        else:
            raise ValueError(f"Got unknown model prefix {model_prefix} when handling block response")
@@ -731,6 +743,10 @@ class BedrockLargeLanguageModel(LargeLanguageModel):
                content_delta = payload.get("text")
                finish_reason = payload.get("finish_reason")
            
+            elif model_prefix == "mistral":
+                content_delta = payload.get('outputs')[0].get("text")
+                finish_reason = payload.get('outputs')[0].get("stop_reason")
+
            elif model_prefix == "meta":
                content_delta = payload.get("generation").strip('\n')
                finish_reason = payload.get("stop_reason")
@@ -0,0 +1,23 @@
+model: meta.llama3-70b-instruct-v1:0
+label:
+  en_US: Llama 3 Instruct 70B
+model_type: llm
+model_properties:
+  mode: completion
+  context_size: 8192
+parameter_rules:
+  - name: temperature
+    use_template: temperature
+  - name: top_p
+    use_template: top_p
+  - name: max_gen_len
+    use_template: max_tokens
+    required: true
+    default: 512
+    min: 1
+    max: 2048
+pricing:
+  input: '0.00265'
+  output: '0.0035'
+  unit: '0.00001'
+  currency: USD
@@ -0,0 +1,23 @@
+model: meta.llama3-8b-instruct-v1:0
+label:
+  en_US: Llama 3 Instruct 8B
+model_type: llm
+model_properties:
+  mode: completion
+  context_size: 8192
+parameter_rules:
+  - name: temperature
+    use_template: temperature
+  - name: top_p
+    use_template: top_p
+  - name: max_gen_len
+    use_template: max_tokens
+    required: true
+    default: 512
+    min: 1
+    max: 2048
+pricing:
+  input: '0.0004'
+  output: '0.0006'
+  unit: '0.0001'
+  currency: USD
@@ -0,0 +1,39 @@
+model: mistral.mistral-7b-instruct-v0:2
+label:
+  en_US: Mistral 7B Instruct
+model_type: llm
+model_properties:
+  mode: completion
+  context_size: 32000
+parameter_rules:
+  - name: temperature
+    use_template: temperature
+    required: false
+    default: 0.5
+  - name: top_p
+    use_template: top_p
+    required: false
+    default: 0.9
+  - name: top_k
+    use_template: top_k
+    label:
+      zh_Hans: 取样数量
+      en_US: Top k
+    type: int
+    help:
+      zh_Hans: 仅从每个后续标记的前 K 个选项中采样。
+      en_US: Only sample from the top K options for each subsequent token.
+    required: false
+    default: 50
+    max: 200
+  - name: max_tokens
+    use_template: max_tokens
+    required: true
+    default: 512
+    min: 1
+    max: 8192
+pricing:
+  input: '0.00015'
+  output: '0.0002'
+  unit: '0.00001'
+  currency: USD
@@ -0,0 +1,27 @@
+model: mistral.mistral-large-2402-v1:0
+label:
+  en_US: Mistral Large
+model_type: llm
+model_properties:
+  mode: completion
+  context_size: 32000
+parameter_rules:
+  - name: temperature
+    use_template: temperature
+    required: false
+    default: 0.7
+  - name: top_p
+    use_template: top_p
+    required: false
+    default: 1
+  - name: max_tokens
+    use_template: max_tokens
+    required: true
+    default: 512
+    min: 1
+    max: 4096
+pricing:
+  input: '0.008'
+  output: '0.024'
+  unit: '0.001'
+  currency: USD
@@ -0,0 +1,39 @@
+model: mistral.mixtral-8x7b-instruct-v0:1
+label:
+  en_US: Mixtral 8X7B Instruct
+model_type: llm
+model_properties:
+  mode: completion
+  context_size: 32000
+parameter_rules:
+  - name: temperature
+    use_template: temperature
+    required: false
+    default: 0.5
+  - name: top_p
+    use_template: top_p
+    required: false
+    default: 0.9
+  - name: top_k
+    use_template: top_k
+    label:
+      zh_Hans: 取样数量
+      en_US: Top k
+    type: int
+    help:
+      zh_Hans: 仅从每个后续标记的前 K 个选项中采样。
+      en_US: Only sample from the top K options for each subsequent token.
+    required: false
+    default: 50
+    max: 200
+  - name: max_tokens
+    use_template: max_tokens
+    required: true
+    default: 512
+    min: 1
+    max: 8192
+pricing:
+  input: '0.00045'
+  output: '0.0007'
+  unit: '0.00001'
+  currency: USD
@@ -19,7 +19,7 @@ class GroqProvider(ModelProvider):
            model_instance = self.get_model_instance(ModelType.LLM)

            model_instance.validate_credentials(
-                model='llama2-70b-4096',
+                model='llama3-8b-8192',
                credentials=credentials
            )
        except CredentialsValidateFailedError as ex:
@@ -0,0 +1,25 @@
+model: llama3-70b-8192
+label:
+  zh_Hans: Llama-3-70B-8192
+  en_US: Llama-3-70B-8192
+model_type: llm
+features:
+  - agent-thought
+model_properties:
+  mode: chat
+  context_size: 8192
+parameter_rules:
+  - name: temperature
+    use_template: temperature
+  - name: top_p
+    use_template: top_p
+  - name: max_tokens
+    use_template: max_tokens
+    default: 512
+    min: 1
+    max: 8192
+pricing:
+  input: '0.05'
+  output: '0.1'
+  unit: '0.000001'
+  currency: USD
@@ -0,0 +1,25 @@
+model: llama3-8b-8192
+label:
+  zh_Hans: Llama-3-8B-8192
+  en_US: Llama-3-8B-8192
+model_type: llm
+features:
+  - agent-thought
+model_properties:
+  mode: chat
+  context_size: 8192
+parameter_rules:
+  - name: temperature
+    use_template: temperature
+  - name: top_p
+    use_template: top_p
+  - name: max_tokens
+    use_template: max_tokens
+    default: 512
+    min: 1
+    max: 8192
+pricing:
+  input: '0.59'
+  output: '0.79'
+  unit: '0.000001'
+  currency: USD
@@ -0,0 +1,37 @@
+model: abab6.5-chat
+label:
+  en_US: Abab6.5-Chat
+model_type: llm
+features:
+  - agent-thought
+  - tool-call
+  - stream-tool-call
+model_properties:
+  mode: chat
+  context_size: 8192
+parameter_rules:
+  - name: temperature
+    use_template: temperature
+    min: 0.01
+    max: 1
+    default: 0.1
+  - name: top_p
+    use_template: top_p
+    min: 0.01
+    max: 1
+    default: 0.95
+  - name: max_tokens
+    use_template: max_tokens
+    required: true
+    default: 2048
+    min: 1
+    max: 8192
+  - name: presence_penalty
+    use_template: presence_penalty
+  - name: frequency_penalty
+    use_template: frequency_penalty
+pricing:
+  input: '0.03'
+  output: '0.03'
+  unit: '0.001'
+  currency: RMB
@@ -0,0 +1,37 @@
+model: abab6.5s-chat
+label:
+  en_US: Abab6.5s-Chat
+model_type: llm
+features:
+  - agent-thought
+  - tool-call
+  - stream-tool-call
+model_properties:
+  mode: chat
+  context_size: 245760
+parameter_rules:
+  - name: temperature
+    use_template: temperature
+    min: 0.01
+    max: 1
+    default: 0.1
+  - name: top_p
+    use_template: top_p
+    min: 0.01
+    max: 1
+    default: 0.95
+  - name: max_tokens
+    use_template: max_tokens
+    required: true
+    default: 2048
+    min: 1
+    max: 245760
+  - name: presence_penalty
+    use_template: presence_penalty
+  - name: frequency_penalty
+    use_template: frequency_penalty
+pricing:
+  input: '0.01'
+  output: '0.01'
+  unit: '0.001'
+  currency: RMB
@@ -1,7 +1,7 @@
 - google/gemma-7b
 - google/codegemma-7b
 - meta/llama2-70b
- meta/llama3-8b
- meta/llama3-70b
+- meta/llama3-8b-instruct
+- meta/llama3-70b-instruct
 - mistralai/mixtral-8x7b-instruct-v0.1
 - fuyu-8b
@@ -1,7 +1,7 @@
-model: meta/llama3-70b
+model: meta/llama3-70b-instruct
 label:
-  zh_Hans: meta/llama3-70b
-  en_US: meta/llama3-70b
+  zh_Hans: meta/llama3-70b-instruct
+  en_US: meta/llama3-70b-instruct
 model_type: llm
 features:
  - agent-thought
@@ -1,7 +1,7 @@
-model: meta/llama3-8b
+model: meta/llama3-8b-instruct
 label:
-  zh_Hans: meta/llama3-8b
-  en_US: meta/llama3-8b
+  zh_Hans: meta/llama3-8b-instruct
+  en_US: meta/llama3-8b-instruct
 model_type: llm
 features:
  - agent-thought
@@ -26,8 +26,8 @@ class NVIDIALargeLanguageModel(OAIAPICompatLargeLanguageModel):
        'google/gemma-7b': '',
        'google/codegemma-7b': '',
        'meta/llama2-70b': '',
-        'meta/llama3-8b': '',
-        'meta/llama3-70b': ''
+        'meta/llama3-8b-instruct': '',
+        'meta/llama3-70b-instruct': ''
        
    }

@@ -33,11 +33,17 @@ class ReplicateLargeLanguageModel(_CommonReplicate, LargeLanguageModel):
                tools: Optional[list[PromptMessageTool]] = None, stop: Optional[list[str]] = None, stream: bool = True,
                user: Optional[str] = None) -> Union[LLMResult, Generator]:

-        version = credentials['model_version']
+        model_version = ''
+        if 'model_version' in credentials:
+            model_version = credentials['model_version']

        client = ReplicateClient(api_token=credentials['replicate_api_token'], timeout=30)
        model_info = client.models.get(model)
-        model_info_version = model_info.versions.get(version)
+
+        if model_version:
+            model_info_version = model_info.versions.get(model_version)
+        else:
+            model_info_version = model_info.latest_version

        inputs = {**model_parameters}

@@ -65,29 +71,35 @@ class ReplicateLargeLanguageModel(_CommonReplicate, LargeLanguageModel):
        if 'replicate_api_token' not in credentials:
            raise CredentialsValidateFailedError('Replicate Access Token must be provided.')

-        if 'model_version' not in credentials:
-            raise CredentialsValidateFailedError('Replicate Model Version must be provided.')
+        model_version = ''
+        if 'model_version' in credentials:
+            model_version = credentials['model_version']

        if model.count("/") != 1:
            raise CredentialsValidateFailedError('Replicate Model Name must be provided, '
                                                 'format: {user_name}/{model_name}')

-        version = credentials['model_version']
-
        try:
            client = ReplicateClient(api_token=credentials['replicate_api_token'], timeout=30)
            model_info = client.models.get(model)
-            model_info_version = model_info.versions.get(version)

-            self._check_text_generation_model(model_info_version, model, version)
+            if model_version:
+                model_info_version = model_info.versions.get(model_version)
+            else:
+                model_info_version = model_info.latest_version
+
+            self._check_text_generation_model(model_info_version, model, model_version, model_info.description)
        except ReplicateError as e:
            raise CredentialsValidateFailedError(
-                f"Model {model}:{version} not exists, cause: {e.__class__.__name__}:{str(e)}")
+                f"Model {model}:{model_version} not exists, cause: {e.__class__.__name__}:{str(e)}")
        except Exception as e:
            raise CredentialsValidateFailedError(str(e))

    @staticmethod
-    def _check_text_generation_model(model_info_version, model_name, version):
+    def _check_text_generation_model(model_info_version, model_name, version, description):
+        if 'language model' in description.lower():
+            return
+
        if 'temperature' not in model_info_version.openapi_schema['components']['schemas']['Input']['properties'] \
                or 'top_p' not in model_info_version.openapi_schema['components']['schemas']['Input']['properties'] \
                or 'top_k' not in model_info_version.openapi_schema['components']['schemas']['Input']['properties']:
@@ -113,11 +125,17 @@ class ReplicateLargeLanguageModel(_CommonReplicate, LargeLanguageModel):

    @classmethod
    def _get_customizable_model_parameter_rules(cls, model: str, credentials: dict) -> list[ParameterRule]:
-        version = credentials['model_version']
+        model_version = ''
+        if 'model_version' in credentials:
+            model_version = credentials['model_version']

        client = ReplicateClient(api_token=credentials['replicate_api_token'], timeout=30)
        model_info = client.models.get(model)
-        model_info_version = model_info.versions.get(version)
+
+        if model_version:
+            model_info_version = model_info.versions.get(model_version)
+        else:
+            model_info_version = model_info.latest_version

        parameter_rules = []

@@ -35,7 +35,7 @@ model_credential_schema:
      label:
        en_US: Model Version
      type: text-input
-      required: true
+      required: false
      placeholder:
-        zh_Hans: 在此输入您的模型版本
-        en_US: Enter your model version
+        zh_Hans: 在此输入您的模型版本，默认为最新版本
+        en_US: Enter your model version, default to the latest version
@@ -17,9 +17,16 @@ class ReplicateEmbeddingModel(_CommonReplicate, TextEmbeddingModel):
                user: Optional[str] = None) -> TextEmbeddingResult:

        client = ReplicateClient(api_token=credentials['replicate_api_token'], timeout=30)
-        replicate_model_version = f'{model}:{credentials["model_version"]}'

-        text_input_key = self._get_text_input_key(model, credentials['model_version'], client)
+        if 'model_version' in credentials:
+            model_version = credentials['model_version']
+        else:
+            model_info = client.models.get(model)
+            model_version = model_info.latest_version.id
+
+        replicate_model_version = f'{model}:{model_version}'
+
+        text_input_key = self._get_text_input_key(model, model_version, client)

        embeddings = self._generate_embeddings_by_text_input_key(client, replicate_model_version, text_input_key,
                                                                 texts)
@@ -43,14 +50,18 @@ class ReplicateEmbeddingModel(_CommonReplicate, TextEmbeddingModel):
        if 'replicate_api_token' not in credentials:
            raise CredentialsValidateFailedError('Replicate Access Token must be provided.')

-        if 'model_version' not in credentials:
-            raise CredentialsValidateFailedError('Replicate Model Version must be provided.')
-
        try:
            client = ReplicateClient(api_token=credentials['replicate_api_token'], timeout=30)
-            replicate_model_version = f'{model}:{credentials["model_version"]}'

-            text_input_key = self._get_text_input_key(model, credentials['model_version'], client)
+            if 'model_version' in credentials:
+                model_version = credentials['model_version']
+            else:
+                model_info = client.models.get(model)
+                model_version = model_info.latest_version.id
+
+            replicate_model_version = f'{model}:{model_version}'
+
+            text_input_key = self._get_text_input_key(model, model_version, client)

            self._generate_embeddings_by_text_input_key(client, replicate_model_version, text_input_key,
                                                        ['Hello worlds!'])
@@ -1,9 +1,23 @@
 from collections.abc import Generator
+from decimal import Decimal
 from typing import Optional, Union

-from core.model_runtime.entities.llm_entities import LLMResult
-from core.model_runtime.entities.message_entities import PromptMessage, PromptMessageTool
-from core.model_runtime.entities.model_entities import AIModelEntity
+from core.model_runtime.entities.common_entities import I18nObject
+from core.model_runtime.entities.llm_entities import LLMMode, LLMResult
+from core.model_runtime.entities.message_entities import (
+    PromptMessage,
+    PromptMessageTool,
+)
+from core.model_runtime.entities.model_entities import (
+    AIModelEntity,
+    DefaultParameterName,
+    FetchFrom,
+    ModelPropertyKey,
+    ModelType,
+    ParameterRule,
+    ParameterType,
+    PriceConfig,
+)
 from core.model_runtime.model_providers.openai_api_compatible.llm.llm import OAIAPICompatLargeLanguageModel


@@ -36,8 +50,98 @@ class TogetherAILargeLanguageModel(OAIAPICompatLargeLanguageModel):

    def get_customizable_model_schema(self, model: str, credentials: dict) -> AIModelEntity:
        cred_with_endpoint = self._update_endpoint_url(credentials=credentials)
+        REPETITION_PENALTY = "repetition_penalty"
+        TOP_K = "top_k"
+        features = []

-        return super().get_customizable_model_schema(model, cred_with_endpoint)
+        entity = AIModelEntity(
+            model=model,
+            label=I18nObject(en_US=model),
+            model_type=ModelType.LLM,
+            fetch_from=FetchFrom.CUSTOMIZABLE_MODEL,
+            features=features,
+            model_properties={
+                ModelPropertyKey.CONTEXT_SIZE: int(cred_with_endpoint.get('context_size', "4096")),
+                ModelPropertyKey.MODE: cred_with_endpoint.get('mode'),
+            },
+            parameter_rules=[
+                ParameterRule(
+                    name=DefaultParameterName.TEMPERATURE.value,
+                    label=I18nObject(en_US="Temperature"),
+                    type=ParameterType.FLOAT,
+                    default=float(cred_with_endpoint.get('temperature', 0.7)),
+                    min=0,
+                    max=2,
+                    precision=2
+                ),
+                ParameterRule(
+                    name=DefaultParameterName.TOP_P.value,
+                    label=I18nObject(en_US="Top P"),
+                    type=ParameterType.FLOAT,
+                    default=float(cred_with_endpoint.get('top_p', 1)),
+                    min=0,
+                    max=1,
+                    precision=2
+                ),
+                ParameterRule(
+                    name=TOP_K,
+                    label=I18nObject(en_US="Top K"),
+                    type=ParameterType.INT,
+                    default=int(cred_with_endpoint.get('top_k', 50)),
+                    min=-2147483647,
+                    max=2147483647,
+                    precision=0
+                ),
+                ParameterRule(
+                    name=REPETITION_PENALTY,
+                    label=I18nObject(en_US="Repetition Penalty"),
+                    type=ParameterType.FLOAT,
+                    default=float(cred_with_endpoint.get('repetition_penalty', 1)),
+                    min=-3.4,
+                    max=3.4,
+                    precision=1
+                ),
+                ParameterRule(
+                    name=DefaultParameterName.MAX_TOKENS.value,
+                    label=I18nObject(en_US="Max Tokens"),
+                    type=ParameterType.INT,
+                    default=512,
+                    min=1,
+                    max=int(cred_with_endpoint.get('max_tokens_to_sample', 4096)),
+                ),
+                ParameterRule(
+                    name=DefaultParameterName.FREQUENCY_PENALTY.value,
+                    label=I18nObject(en_US="Frequency Penalty"),
+                    type=ParameterType.FLOAT,
+                    default=float(credentials.get('frequency_penalty', 0)),
+                    min=-2,
+                    max=2
+                ),
+                ParameterRule(
+                    name=DefaultParameterName.PRESENCE_PENALTY.value,
+                    label=I18nObject(en_US="Presence Penalty"),
+                    type=ParameterType.FLOAT,
+                    default=float(credentials.get('presence_penalty', 0)),
+                    min=-2,
+                    max=2
+                )
+            ],
+            pricing=PriceConfig(
+                input=Decimal(cred_with_endpoint.get('input_price', 0)),
+                output=Decimal(cred_with_endpoint.get('output_price', 0)),
+                unit=Decimal(cred_with_endpoint.get('unit', 0)),
+                currency=cred_with_endpoint.get('currency', "USD")
+            ),
+        )
+
+        if cred_with_endpoint['mode'] == 'chat':
+            entity.model_properties[ModelPropertyKey.MODE] = LLMMode.CHAT.value
+        elif cred_with_endpoint['mode'] == 'completion':
+            entity.model_properties[ModelPropertyKey.MODE] = LLMMode.COMPLETION.value
+        else:
+            raise ValueError(f"Unknown completion type {cred_with_endpoint['completion_type']}")
+
+        return entity

    def get_num_tokens(self, model: str, credentials: dict, prompt_messages: list[PromptMessage],
                       tools: Optional[list[PromptMessageTool]] = None) -> int:
@@ -47,6 +47,20 @@ class XinferenceSpeech2TextModel(Speech2TextModel):
            if "/" in credentials['model_uid'] or "?" in credentials['model_uid'] or "#" in credentials['model_uid']:
                raise CredentialsValidateFailedError("model_uid should not contain /, ?, or #")

+            if credentials['server_url'].endswith('/'):
+                credentials['server_url'] = credentials['server_url'][:-1]
+
+            # initialize client
+            client = Client(
+                base_url=credentials['server_url']
+            )
+
+            xinference_client = client.get_model(model_uid=credentials['model_uid'])
+
+            if not isinstance(xinference_client, RESTfulAudioModelHandle):
+                raise InvokeBadRequestError(
+                    'please check model type, the model you want to invoke is not a audio model')
+
            audio_file_path = self._get_demo_file_path()

            with open(audio_file_path, 'rb') as audio_file:
@@ -110,17 +124,8 @@ class XinferenceSpeech2TextModel(Speech2TextModel):
        if credentials['server_url'].endswith('/'):
            credentials['server_url'] = credentials['server_url'][:-1]

-        # initialize client
-        client = Client(
-            base_url=credentials['server_url']
-        )
-
-        xinference_client = client.get_model(model_uid=credentials['model_uid'])
-
-        if not isinstance(xinference_client, RESTfulAudioModelHandle):
-            raise InvokeBadRequestError('please check model type, the model you want to invoke is not a audio model')
-        
-        response = xinference_client.transcriptions(
+        handle = RESTfulAudioModelHandle(credentials['model_uid'],credentials['server_url'],auth_headers={})
+        response = handle.transcriptions(
            audio=file,
            language = language,
            prompt = prompt,
@@ -32,3 +32,8 @@ parameter_rules:
      zh_Hans: SSE接口调用时，用于控制每次返回内容方式是增量还是全量，不提供此参数时默认为增量返回，true 为增量返回，false 为全量返回。
      en_US: When the SSE interface is called, it is used to control whether the content is returned incrementally or in full. If this parameter is not provided, the default is incremental return. true means incremental return, false means full return.
    required: false
+  - name: max_tokens
+    use_template: max_tokens
+    default: 1024
+    min: 1
+    max: 8192
@@ -31,7 +31,10 @@ class AdvancedPromptTransform(PromptTransform):
                   context: Optional[str],
                   memory_config: Optional[MemoryConfig],
                   memory: Optional[TokenBufferMemory],
-                   model_config: ModelConfigWithCredentialsEntity) -> list[PromptMessage]:
+                   model_config: ModelConfigWithCredentialsEntity,
+                   query_prompt_template: Optional[str] = None) -> list[PromptMessage]:
+        inputs = {key: str(value) for key, value in inputs.items()}
+
        prompt_messages = []

        model_mode = ModelMode.value_of(model_config.mode)
@@ -51,6 +54,7 @@ class AdvancedPromptTransform(PromptTransform):
                prompt_template=prompt_template,
                inputs=inputs,
                query=query,
+                query_prompt_template=query_prompt_template,
                files=files,
                context=context,
                memory_config=memory_config,
@@ -119,7 +123,8 @@ class AdvancedPromptTransform(PromptTransform):
                                        context: Optional[str],
                                        memory_config: Optional[MemoryConfig],
                                        memory: Optional[TokenBufferMemory],
-                                        model_config: ModelConfigWithCredentialsEntity) -> list[PromptMessage]:
+                                        model_config: ModelConfigWithCredentialsEntity,
+                                        query_prompt_template: Optional[str] = None) -> list[PromptMessage]:
        """
        Get chat model prompt messages.
        """
@@ -146,6 +151,20 @@ class AdvancedPromptTransform(PromptTransform):
            elif prompt_item.role == PromptMessageRole.ASSISTANT:
                prompt_messages.append(AssistantPromptMessage(content=prompt))

+        if query and query_prompt_template:
+            prompt_template = PromptTemplateParser(
+                template=query_prompt_template,
+                with_variable_tmpl=self.with_variable_tmpl
+            )
+            prompt_inputs = {k: inputs[k] for k in prompt_template.variable_keys if k in inputs}
+            prompt_inputs['#sys.query#'] = query
+
+            prompt_inputs = self._set_context_variable(context, prompt_template, prompt_inputs)
+
+            query = prompt_template.format(
+                prompt_inputs
+            )
+
        if memory and memory_config:
            prompt_messages = self._append_chat_histories(memory, memory_config, prompt_messages, model_config)

@@ -40,3 +40,4 @@ class MemoryConfig(BaseModel):

    role_prefix: Optional[RolePrefix] = None
    window: WindowConfig
+    query_prompt_template: Optional[str] = None
@@ -55,6 +55,8 @@ class SimplePromptTransform(PromptTransform):
                   memory: Optional[TokenBufferMemory],
                   model_config: ModelConfigWithCredentialsEntity) -> \
            tuple[list[PromptMessage], Optional[list[str]]]:
+        inputs = {key: str(value) for key, value in inputs.items()}
+
        model_mode = ModelMode.value_of(model_config.mode)
        if model_mode == ModelMode.CHAT:
            prompt_messages, stops = self._get_chat_model_prompt_messages(
@@ -110,19 +110,37 @@ class MilvusVector(BaseVector):
            return None

    def delete_by_metadata_field(self, key: str, value: str):
+        alias = uuid4().hex
+        if self._client_config.secure:
+            uri = "https://" + str(self._client_config.host) + ":" + str(self._client_config.port)
+        else:
+            uri = "http://" + str(self._client_config.host) + ":" + str(self._client_config.port)
+        connections.connect(alias=alias, uri=uri, user=self._client_config.user, password=self._client_config.password)

-        ids = self.get_ids_by_metadata_field(key, value)
-        if ids:
-            self._client.delete(collection_name=self._collection_name, pks=ids)
+        from pymilvus import utility
+        if utility.has_collection(self._collection_name, using=alias):

-    def delete_by_ids(self, doc_ids: list[str]) -> None:
+            ids = self.get_ids_by_metadata_field(key, value)
+            if ids:
+                self._client.delete(collection_name=self._collection_name, pks=ids)

-        result = self._client.query(collection_name=self._collection_name,
-                                    filter=f'metadata["doc_id"] in {doc_ids}',
-                                    output_fields=["id"])
-        if result:
-            ids = [item["id"] for item in result]
-            self._client.delete(collection_name=self._collection_name, pks=ids)
+    def delete_by_ids(self, ids: list[str]) -> None:
+        alias = uuid4().hex
+        if self._client_config.secure:
+            uri = "https://" + str(self._client_config.host) + ":" + str(self._client_config.port)
+        else:
+            uri = "http://" + str(self._client_config.host) + ":" + str(self._client_config.port)
+        connections.connect(alias=alias, uri=uri, user=self._client_config.user, password=self._client_config.password)
+
+        from pymilvus import utility
+        if utility.has_collection(self._collection_name, using=alias):
+
+            result = self._client.query(collection_name=self._collection_name,
+                                        filter=f'metadata["doc_id"] in {ids}',
+                                        output_fields=["id"])
+            if result:
+                ids = [item["id"] for item in result]
+                self._client.delete(collection_name=self._collection_name, pks=ids)

    def delete(self) -> None:
        alias = uuid4().hex
@@ -0,0 +1,12 @@
+from uuid import UUID
+
+from numpy import ndarray
+from sqlalchemy.orm import DeclarativeBase, Mapped
+
+
+class CollectionORM(DeclarativeBase):
+    __tablename__: str
+    id: Mapped[UUID]
+    text: Mapped[str]
+    meta: Mapped[dict]
+    vector: Mapped[ndarray]
@@ -0,0 +1,224 @@
+import logging
+from typing import Any
+from uuid import UUID, uuid4
+
+from numpy import ndarray
+from pgvecto_rs.sqlalchemy import Vector
+from pydantic import BaseModel, root_validator
+from sqlalchemy import Float, String, create_engine, insert, select, text
+from sqlalchemy import text as sql_text
+from sqlalchemy.dialects import postgresql
+from sqlalchemy.orm import Mapped, Session, mapped_column
+
+from core.rag.datasource.vdb.pgvecto_rs.collection import CollectionORM
+from core.rag.datasource.vdb.vector_base import BaseVector
+from core.rag.models.document import Document
+from extensions.ext_redis import redis_client
+
+logger = logging.getLogger(__name__)
+
+
+class PgvectoRSConfig(BaseModel):
+    host: str
+    port: int
+    user: str
+    password: str
+    database: str
+
+    @root_validator()
+    def validate_config(cls, values: dict) -> dict:
+        if not values['host']:
+            raise ValueError("config PGVECTO_RS_HOST is required")
+        if not values['port']:
+            raise ValueError("config PGVECTO_RS_PORT is required")
+        if not values['user']:
+            raise ValueError("config PGVECTO_RS_USER is required")
+        if not values['password']:
+            raise ValueError("config PGVECTO_RS_PASSWORD is required")
+        if not values['database']:
+            raise ValueError("config PGVECTO_RS_DATABASE is required")
+        return values
+
+
+class PGVectoRS(BaseVector):
+
+    def __init__(self, collection_name: str, config: PgvectoRSConfig, dim: int):
+        super().__init__(collection_name)
+        self._client_config = config
+        self._url = f"postgresql+psycopg2://{config.user}:{config.password}@{config.host}:{config.port}/{config.database}"
+        self._client = create_engine(self._url)
+        with Session(self._client) as session:
+            session.execute(text("CREATE EXTENSION IF NOT EXISTS vectors"))
+            session.commit()
+        self._fields = []
+
+        class _Table(CollectionORM):
+            __tablename__ = collection_name
+            __table_args__ = {"extend_existing": True}  # noqa: RUF012
+            id: Mapped[UUID] = mapped_column(
+                postgresql.UUID(as_uuid=True),
+                primary_key=True,
+            )
+            text: Mapped[str] = mapped_column(String)
+            meta: Mapped[dict] = mapped_column(postgresql.JSONB)
+            vector: Mapped[ndarray] = mapped_column(Vector(dim))
+
+        self._table = _Table
+        self._distance_op = "<=>"
+
+    def get_type(self) -> str:
+        return 'pgvecto-rs'
+
+    def create(self, texts: list[Document], embeddings: list[list[float]], **kwargs):
+        self.create_collection(len(embeddings[0]))
+        self.add_texts(texts, embeddings)
+
+    def create_collection(self, dimension: int):
+        lock_name = 'vector_indexing_lock_{}'.format(self._collection_name)
+        with redis_client.lock(lock_name, timeout=20):
+            collection_exist_cache_key = 'vector_indexing_{}'.format(self._collection_name)
+            if redis_client.get(collection_exist_cache_key):
+                return
+            index_name = f"{self._collection_name}_embedding_index"
+            with Session(self._client) as session:
+                create_statement = sql_text(f"""
+                    CREATE TABLE IF NOT EXISTS {self._collection_name} (
+                        id UUID PRIMARY KEY,
+                        text TEXT NOT NULL,
+                        meta JSONB NOT NULL,
+                        vector vector({dimension}) NOT NULL
+                    ) using heap; 
+                """)
+                session.execute(create_statement)
+                index_statement = sql_text(f"""
+                        CREATE INDEX IF NOT EXISTS {index_name}
+                        ON {self._collection_name} USING vectors(vector vector_l2_ops)
+                        WITH (options = $$
+                                optimizing.optimizing_threads = 30
+                                segment.max_growing_segment_size = 2000
+                                segment.max_sealed_segment_size = 30000000
+                                [indexing.hnsw]
+                                m=30
+                                ef_construction=500
+                                $$);
+                    """)
+                session.execute(index_statement)
+                session.commit()
+            redis_client.set(collection_exist_cache_key, 1, ex=3600)
+
+    def add_texts(self, documents: list[Document], embeddings: list[list[float]], **kwargs):
+        pks = []
+        with Session(self._client) as session:
+            for document, embedding in zip(documents, embeddings):
+                pk = uuid4()
+                session.execute(
+                    insert(self._table).values(
+                        id=pk,
+                        text=document.page_content,
+                        meta=document.metadata,
+                        vector=embedding,
+                    ),
+                )
+                pks.append(pk)
+            session.commit()
+
+        return pks
+
+    def delete_by_document_id(self, document_id: str):
+        ids = self.get_ids_by_metadata_field('document_id', document_id)
+        if ids:
+            with Session(self._client) as session:
+                select_statement = sql_text(f"DELETE FROM {self._collection_name} WHERE id = ANY(:ids)")
+                session.execute(select_statement, {'ids': ids})
+                session.commit()
+
+    def get_ids_by_metadata_field(self, key: str, value: str):
+        result = None
+        with Session(self._client) as session:
+            select_statement = sql_text(
+                f"SELECT id FROM {self._collection_name} WHERE meta->>'{key}' = '{value}'; "
+            )
+            result = session.execute(select_statement).fetchall()
+        if result:
+            return [item[0] for item in result]
+        else:
+            return None
+
+    def delete_by_metadata_field(self, key: str, value: str):
+
+        ids = self.get_ids_by_metadata_field(key, value)
+        if ids:
+            with Session(self._client) as session:
+                select_statement = sql_text(f"DELETE FROM {self._collection_name} WHERE id = ANY(:ids)")
+                session.execute(select_statement, {'ids': ids})
+                session.commit()
+
+    def delete_by_ids(self, ids: list[str]) -> None:
+        with Session(self._client) as session:
+            select_statement = sql_text(
+                f"SELECT id FROM {self._collection_name} WHERE meta->>'doc_id' = ANY (:doc_ids); "
+            )
+            result = session.execute(select_statement, {'doc_ids': ids}).fetchall()
+        if result:
+            ids = [item[0] for item in result]
+            if ids:
+                with Session(self._client) as session:
+                    select_statement = sql_text(f"DELETE FROM {self._collection_name} WHERE id = ANY(:ids)")
+                    session.execute(select_statement, {'ids': ids})
+                    session.commit()
+
+    def delete(self) -> None:
+        with Session(self._client) as session:
+            session.execute(sql_text(f"DROP TABLE IF EXISTS {self._collection_name}"))
+            session.commit()
+
+    def text_exists(self, id: str) -> bool:
+        with Session(self._client) as session:
+            select_statement = sql_text(
+                f"SELECT id FROM {self._collection_name} WHERE meta->>'doc_id' = '{id}' limit 1; "
+            )
+            result = session.execute(select_statement).fetchall()
+        return len(result) > 0
+
+    def search_by_vector(self, query_vector: list[float], **kwargs: Any) -> list[Document]:
+        with Session(self._client) as session:
+            stmt = (
+                select(
+                    self._table,
+                    self._table.vector.op(self._distance_op, return_type=Float)(
+                        query_vector,
+                    ).label("distance"),
+                )
+                .limit(kwargs.get('top_k', 2))
+                .order_by("distance")
+            )
+            res = session.execute(stmt)
+            results = [(row[0], row[1]) for row in res]
+
+        # Organize results.
+        docs = []
+        for record, dis in results:
+            metadata = record.meta
+            score = 1 - dis
+            metadata['score'] = score
+            score_threshold = kwargs.get('score_threshold') if kwargs.get('score_threshold') else 0.0
+            if score > score_threshold:
+                doc = Document(page_content=record.text,
+                               metadata=metadata)
+                docs.append(doc)
+        return docs
+
+    def search_by_full_text(self, query: str, **kwargs: Any) -> list[Document]:
+        # with Session(self._client) as session:
+        #     select_statement = sql_text(
+        #         f"SELECT text, meta FROM {self._collection_name} WHERE to_tsvector(text) @@ '{query}'::tsquery"
+        #     )
+        #     results = session.execute(select_statement).fetchall()
+        # if results:
+        #     docs = []
+        #     for result in results:
+        #         doc = Document(page_content=result[0],
+        #                        metadata=result[1])
+        #         docs.append(doc)
+        #     return docs
+        return []
@@ -36,6 +36,8 @@ class QdrantConfig(BaseModel):
    api_key: Optional[str]
    timeout: float = 20
    root_path: Optional[str]
+    grpc_port: int = 6334
+    prefer_grpc: bool = False

    def to_qdrant_params(self):
        if self.endpoint and self.endpoint.startswith('path:'):
@@ -50,7 +52,10 @@ class QdrantConfig(BaseModel):
            return {
                'url': self.endpoint,
                'api_key': self.api_key,
-                'timeout': self.timeout
+                'timeout': self.timeout,
+                'verify': self.endpoint.startswith('https'),
+                'grpc_port': self.grpc_port,
+                'prefer_grpc': self.prefer_grpc
            }


@@ -112,8 +117,7 @@ class QdrantVector(BaseVector):

                # create payload index
                self._client.create_payload_index(collection_name, Field.GROUP_KEY.value,
-                                                  field_schema=PayloadSchemaType.KEYWORD,
-                                                  field_type=PayloadSchemaType.KEYWORD)
+                                                  field_schema=PayloadSchemaType.KEYWORD)
                # creat full text index
                text_index_params = TextIndexParams(
                    type=TextIndexType.TEXT,
@@ -217,29 +221,38 @@ class QdrantVector(BaseVector):
    def delete_by_metadata_field(self, key: str, value: str):

        from qdrant_client.http import models
+        from qdrant_client.http.exceptions import UnexpectedResponse

-        filter = models.Filter(
-            must=[
-                models.FieldCondition(
-                    key=f"metadata.{key}",
-                    match=models.MatchValue(value=value),
+        try:
+            filter = models.Filter(
+                must=[
+                    models.FieldCondition(
+                        key=f"metadata.{key}",
+                        match=models.MatchValue(value=value),
+                    ),
+                ],
+            )
+
+            self._reload_if_needed()
+
+            self._client.delete(
+                collection_name=self._collection_name,
+                points_selector=FilterSelector(
+                    filter=filter
                ),
-            ],
-        )
-
-        self._reload_if_needed()
-
-        self._client.delete(
-            collection_name=self._collection_name,
-            points_selector=FilterSelector(
-                filter=filter
-            ),
-        )
+            )
+        except UnexpectedResponse as e:
+            # Collection does not exist, so return
+            if e.status_code == 404:
+                return
+            # Some other error occurred, so re-raise the exception
+            else:
+                raise e

    def delete(self):
        from qdrant_client.http import models
        from qdrant_client.http.exceptions import UnexpectedResponse
-        
+
        try:
            filter = models.Filter(
                must=[
@@ -257,29 +270,40 @@ class QdrantVector(BaseVector):
            )
        except UnexpectedResponse as e:
            # Collection does not exist, so return
-            if e.status_code == 404:                
+            if e.status_code == 404:
                return
            # Some other error occurred, so re-raise the exception
            else:
                raise e
+
    def delete_by_ids(self, ids: list[str]) -> None:

        from qdrant_client.http import models
+        from qdrant_client.http.exceptions import UnexpectedResponse
+
        for node_id in ids:
-            filter = models.Filter(
-                must=[
-                    models.FieldCondition(
-                        key="metadata.doc_id",
-                        match=models.MatchValue(value=node_id),
+            try:
+                filter = models.Filter(
+                    must=[
+                        models.FieldCondition(
+                            key="metadata.doc_id",
+                            match=models.MatchValue(value=node_id),
+                        ),
+                    ],
+                )
+                self._client.delete(
+                    collection_name=self._collection_name,
+                    points_selector=FilterSelector(
+                        filter=filter
                    ),
-                ],
-            )
-            self._client.delete(
-                collection_name=self._collection_name,
-                points_selector=FilterSelector(
-                    filter=filter
-                ),
-            )
+                )
+            except UnexpectedResponse as e:
+                # Collection does not exist, so return
+                if e.status_code == 404:
+                    return
+                # Some other error occurred, so re-raise the exception
+                else:
+                    raise e

    def text_exists(self, id: str) -> bool:
        all_collection_name = []
@@ -1,16 +1,23 @@
-import logging
-from typing import Any
+import uuid
+from typing import Any, Optional

-from pgvecto_rs.sdk import PGVectoRs, Record
 from pydantic import BaseModel, root_validator
+from sqlalchemy import Column, Sequence, String, Table, create_engine, insert
 from sqlalchemy import text as sql_text
+from sqlalchemy.dialects.postgresql import JSON, TEXT
 from sqlalchemy.orm import Session

+try:
+    from sqlalchemy.orm import declarative_base
+except ImportError:
+    from sqlalchemy.ext.declarative import declarative_base
+
 from core.rag.datasource.vdb.vector_base import BaseVector
 from core.rag.models.document import Document
 from extensions.ext_redis import redis_client

-logger = logging.getLogger(__name__)
+Base = declarative_base()  # type: Any
+

 class RelytConfig(BaseModel):
    host: str
@@ -36,16 +43,14 @@ class RelytConfig(BaseModel):

 class RelytVector(BaseVector):

-    def __init__(self, collection_name: str, config: RelytConfig, dim: int):
+    def __init__(self, collection_name: str, config: RelytConfig, group_id: str):
        super().__init__(collection_name)
+        self.embedding_dimension = 1536
        self._client_config = config
        self._url = f"postgresql+psycopg2://{config.user}:{config.password}@{config.host}:{config.port}/{config.database}"
-        self._client = PGVectoRs(
-            db_url=self._url,
-            collection_name=self._collection_name,
-            dimension=dim
-        )
+        self.client = create_engine(self._url)
        self._fields = []
+        self._group_id = group_id

    def get_type(self) -> str:
        return 'relyt'
@@ -54,6 +59,7 @@ class RelytVector(BaseVector):
        index_params = {}
        metadatas = [d.metadata for d in texts]
        self.create_collection(len(embeddings[0]))
+        self.embedding_dimension = len(embeddings[0])
        self.add_texts(texts, embeddings)

    def create_collection(self, dimension: int):
@@ -63,21 +69,21 @@ class RelytVector(BaseVector):
            if redis_client.get(collection_exist_cache_key):
                return
            index_name = f"{self._collection_name}_embedding_index"
-            with Session(self._client._engine) as session:
-                drop_statement = sql_text(f"DROP TABLE IF EXISTS collection_{self._collection_name}")
+            with Session(self.client) as session:
+                drop_statement = sql_text(f"""DROP TABLE IF EXISTS "{self._collection_name}"; """)
                session.execute(drop_statement)
                create_statement = sql_text(f"""
-                    CREATE TABLE IF NOT EXISTS collection_{self._collection_name} (
-                        id UUID PRIMARY KEY,
-                        text TEXT NOT NULL,
-                        meta JSONB NOT NULL,
+                    CREATE TABLE IF NOT EXISTS "{self._collection_name}" (
+                        id TEXT PRIMARY KEY,
+                        document TEXT NOT NULL,
+                        metadata JSON NOT NULL,
                        embedding vector({dimension}) NOT NULL
                    ) using heap; 
                """)
                session.execute(create_statement)
                index_statement = sql_text(f"""
                        CREATE INDEX {index_name}
-                        ON collection_{self._collection_name} USING vectors(embedding vector_l2_ops)
+                        ON "{self._collection_name}" USING vectors(embedding vector_l2_ops)
                        WITH (options = $$
                                optimizing.optimizing_threads = 30
                                segment.max_growing_segment_size = 2000
@@ -92,21 +98,62 @@ class RelytVector(BaseVector):
            redis_client.set(collection_exist_cache_key, 1, ex=3600)

    def add_texts(self, documents: list[Document], embeddings: list[list[float]], **kwargs):
-        records = [Record.from_text(d.page_content, e, d.metadata) for d, e in zip(documents, embeddings)]
-        pks = [str(r.id) for r in records]
-        self._client.insert(records)
-        return pks
+        from pgvecto_rs.sqlalchemy import Vector
+
+        ids = [str(uuid.uuid1()) for _ in documents]
+        metadatas = [d.metadata for d in documents]
+        for metadata in metadatas:
+            metadata['group_id'] = self._group_id
+        texts = [d.page_content for d in documents]
+
+        # Define the table schema
+        chunks_table = Table(
+            self._collection_name,
+            Base.metadata,
+            Column("id", TEXT, primary_key=True),
+            Column("embedding", Vector(len(embeddings[0]))),
+            Column("document", String, nullable=True),
+            Column("metadata", JSON, nullable=True),
+            extend_existing=True,
+        )
+
+        chunks_table_data = []
+        with self.client.connect() as conn:
+            with conn.begin():
+                for document, metadata, chunk_id, embedding in zip(
+                        texts, metadatas, ids, embeddings
+                ):
+                    chunks_table_data.append(
+                        {
+                            "id": chunk_id,
+                            "embedding": embedding,
+                            "document": document,
+                            "metadata": metadata,
+                        }
+                    )
+
+                    # Execute the batch insert when the batch size is reached
+                    if len(chunks_table_data) == 500:
+                        conn.execute(insert(chunks_table).values(chunks_table_data))
+                        # Clear the chunks_table_data list for the next batch
+                        chunks_table_data.clear()
+
+                # Insert any remaining records that didn't make up a full batch
+                if chunks_table_data:
+                    conn.execute(insert(chunks_table).values(chunks_table_data))
+
+        return ids

    def delete_by_document_id(self, document_id: str):
        ids = self.get_ids_by_metadata_field('document_id', document_id)
        if ids:
-            self._client.delete_by_ids(ids)
+            self.delete_by_uuids(ids)

    def get_ids_by_metadata_field(self, key: str, value: str):
        result = None
-        with Session(self._client._engine) as session:
+        with Session(self.client) as session:
            select_statement = sql_text(
-                f"SELECT id FROM collection_{self._collection_name} WHERE meta->>'{key}' = '{value}'; "
+                f"""SELECT id FROM "{self._collection_name}" WHERE metadata->>'{key}' = '{value}'; """
            )
            result = session.execute(select_statement).fetchall()
        if result:
@@ -114,56 +161,140 @@ class RelytVector(BaseVector):
        else:
            return None

+    def delete_by_uuids(self, ids: list[str] = None):
+        """Delete by vector IDs.
+
+        Args:
+            ids: List of ids to delete.
+        """
+        from pgvecto_rs.sqlalchemy import Vector
+
+        if ids is None:
+            raise ValueError("No ids provided to delete.")
+
+        # Define the table schema
+        chunks_table = Table(
+            self._collection_name,
+            Base.metadata,
+            Column("id", TEXT, primary_key=True),
+            Column("embedding", Vector(self.embedding_dimension)),
+            Column("document", String, nullable=True),
+            Column("metadata", JSON, nullable=True),
+            extend_existing=True,
+        )
+
+        try:
+            with self.client.connect() as conn:
+                with conn.begin():
+                    delete_condition = chunks_table.c.id.in_(ids)
+                    conn.execute(chunks_table.delete().where(delete_condition))
+                    return True
+        except Exception as e:
+            print("Delete operation failed:", str(e))  # noqa: T201
+            return False
+
    def delete_by_metadata_field(self, key: str, value: str):

        ids = self.get_ids_by_metadata_field(key, value)
        if ids:
-            self._client.delete_by_ids(ids)
+            self.delete_by_uuids(ids)

-    def delete_by_ids(self, doc_ids: list[str]) -> None:
-        with Session(self._client._engine) as session:
+    def delete_by_ids(self, ids: list[str]) -> None:
+
+        with Session(self.client) as session:
+            ids_str = ','.join(f"'{doc_id}'" for doc_id in ids)
            select_statement = sql_text(
-                f"SELECT id FROM collection_{self._collection_name} WHERE meta->>'doc_id' in ('{doc_ids}'); "
+                f"""SELECT id FROM "{self._collection_name}" WHERE metadata->>'doc_id' in ({ids_str}); """
            )
            result = session.execute(select_statement).fetchall()
        if result:
            ids = [item[0] for item in result]
-            self._client.delete_by_ids(ids)
+            self.delete_by_uuids(ids)

    def delete(self) -> None:
-        with Session(self._client._engine) as session:
-            session.execute(sql_text(f"DROP TABLE IF EXISTS collection_{self._collection_name}"))
+        with Session(self.client) as session:
+            session.execute(sql_text(f"""DROP TABLE IF EXISTS "{self._collection_name}";"""))
            session.commit()

    def text_exists(self, id: str) -> bool:
-        with Session(self._client._engine) as session:
+        with Session(self.client) as session:
            select_statement = sql_text(
-                f"SELECT id FROM collection_{self._collection_name} WHERE meta->>'doc_id' = '{id}' limit 1; "
+                f"""SELECT id FROM "{self._collection_name}" WHERE metadata->>'doc_id' = '{id}' limit 1; """
            )
            result = session.execute(select_statement).fetchall()
        return len(result) > 0

    def search_by_vector(self, query_vector: list[float], **kwargs: Any) -> list[Document]:
-        from pgvecto_rs.sdk import filters
-        filter_condition = filters.meta_contains(kwargs.get('filter'))
-        results = self._client.search(
-            top_k=int(kwargs.get('top_k')),
+        results = self.similarity_search_with_score_by_vector(
+            k=int(kwargs.get('top_k')),
            embedding=query_vector,
-            filter=filter_condition
+            filter=kwargs.get('filter')
        )

        # Organize results.
        docs = []
-        for record, dis in results:
-            metadata = record.meta
-            metadata['score'] = dis
+        for document, score in results:
            score_threshold = kwargs.get('score_threshold') if kwargs.get('score_threshold') else 0.0
-            if dis > score_threshold:
-                doc = Document(page_content=record.text,
-                               metadata=metadata)
-                docs.append(doc)
+            if 1 - score > score_threshold:
+                docs.append(document)
        return docs

+    def similarity_search_with_score_by_vector(
+        self,
+        embedding: list[float],
+        k: int = 4,
+        filter: Optional[dict] = None,
+    ) -> list[tuple[Document, float]]:
+        # Add the filter if provided
+        try:
+            from sqlalchemy.engine import Row
+        except ImportError:
+            raise ImportError(
+                "Could not import Row from sqlalchemy.engine. "
+                "Please 'pip install sqlalchemy>=1.4'."
+            )
+
+        filter_condition = ""
+        if filter is not None:
+            conditions = [
+                f"metadata->>{key!r} in ({', '.join(map(repr, value))})" if len(value) > 1
+                else f"metadata->>{key!r} = {value[0]!r}"
+                for key, value in filter.items()
+            ]
+            filter_condition = f"WHERE {' AND '.join(conditions)}"
+
+        # Define the base query
+        sql_query = f"""
+            set vectors.enable_search_growing = on;
+            set vectors.enable_search_write = on;
+            SELECT document, metadata, embedding <-> :embedding as distance
+            FROM "{self._collection_name}"
+            {filter_condition}
+            ORDER BY embedding <-> :embedding
+            LIMIT :k
+        """
+
+        # Set up the query parameters
+        embedding_str = ", ".join(format(x) for x in embedding)
+        embedding_str = "[" + embedding_str + "]"
+        params = {"embedding": embedding_str, "k": k}
+
+        # Execute the query and fetch the results
+        with self.client.connect() as conn:
+            results: Sequence[Row] = conn.execute(sql_text(sql_query), params).fetchall()
+
+        documents_with_scores = [
+            (
+                Document(
+                    page_content=result.document,
+                    metadata=result.metadata,
+                ),
+                result.distance,
+            )
+            for result in results
+        ]
+        return documents_with_scores
+
    def search_by_full_text(self, query: str, **kwargs: Any) -> list[Document]:
        # milvus/zilliz/relyt doesn't support bm25 search
        return []
@@ -27,6 +27,12 @@ class BaseVector(ABC):
    def delete_by_ids(self, ids: list[str]) -> None:
        raise NotImplementedError

+    def delete_by_document_id(self, document_id: str):
+        raise NotImplementedError
+
+    def get_ids_by_metadata_field(self, key: str, value: str):
+        raise NotImplementedError
+
    @abstractmethod
    def delete_by_metadata_field(self, key: str, value: str) -> None:
        raise NotImplementedError
@@ -86,7 +86,9 @@ class Vector:
                    endpoint=config.get('QDRANT_URL'),
                    api_key=config.get('QDRANT_API_KEY'),
                    root_path=current_app.root_path,
-                    timeout=config.get('QDRANT_CLIENT_TIMEOUT')
+                    timeout=config.get('QDRANT_CLIENT_TIMEOUT'),
+                    grpc_port=config.get('QDRANT_GRPC_PORT'),
+                    prefer_grpc=config.get('QDRANT_GRPC_ENABLED')
                )
            )
        elif vector_type == "milvus":
@@ -126,7 +128,6 @@ class Vector:
                    "vector_store": {"class_prefix": collection_name}
                }
                self._dataset.index_struct = json.dumps(index_struct_dict)
-            dim = len(self._embeddings.embed_query("hello relyt"))
            return RelytVector(
                collection_name=collection_name,
                config=RelytConfig(
@@ -136,6 +137,31 @@ class Vector:
                    password=config.get('RELYT_PASSWORD'),
                    database=config.get('RELYT_DATABASE'),
                ),
+                group_id=self._dataset.id
+            )
+        elif vector_type == "pgvecto_rs":
+            from core.rag.datasource.vdb.pgvecto_rs.pgvecto_rs import PGVectoRS, PgvectoRSConfig
+            if self._dataset.index_struct_dict:
+                class_prefix: str = self._dataset.index_struct_dict['vector_store']['class_prefix']
+                collection_name = class_prefix.lower()
+            else:
+                dataset_id = self._dataset.id
+                collection_name = Dataset.gen_collection_name_by_id(dataset_id).lower()
+                index_struct_dict = {
+                    "type": 'pgvecto_rs',
+                    "vector_store": {"class_prefix": collection_name}
+                }
+                self._dataset.index_struct = json.dumps(index_struct_dict)
+            dim = len(self._embeddings.embed_query("pgvecto_rs"))
+            return PGVectoRS(
+                collection_name=collection_name,
+                config=PgvectoRSConfig(
+                    host=config.get('PGVECTO_RS_HOST'),
+                    port=config.get('PGVECTO_RS_PORT'),
+                    user=config.get('PGVECTO_RS_USER'),
+                    password=config.get('PGVECTO_RS_PASSWORD'),
+                    database=config.get('PGVECTO_RS_DATABASE'),
+                ),
                dim=dim
            )
        else:
@@ -121,18 +121,20 @@ class WeaviateVector(BaseVector):
        return ids

    def delete_by_metadata_field(self, key: str, value: str):
+        # check whether the index already exists
+        schema = self._default_schema(self._collection_name)
+        if self._client.schema.contains(schema):
+            where_filter = {
+                "operator": "Equal",
+                "path": [key],
+                "valueText": value
+            }

-        where_filter = {
-            "operator": "Equal",
-            "path": [key],
-            "valueText": value
-        }
-
-        self._client.batch.delete_objects(
-            class_name=self._collection_name,
-            where=where_filter,
-            output='minimal'
-        )
+            self._client.batch.delete_objects(
+                class_name=self._collection_name,
+                where=where_filter,
+                output='minimal'
+            )

    def delete(self):
        # check whether the index already exists
@@ -163,11 +165,14 @@ class WeaviateVector(BaseVector):
        return True

    def delete_by_ids(self, ids: list[str]) -> None:
-        for uuid in ids:
-            self._client.data_object.delete(
-                class_name=self._collection_name,
-                uuid=uuid,
-            )
+        # check whether the index already exists
+        schema = self._default_schema(self._collection_name)
+        if self._client.schema.contains(schema):
+            for uuid in ids:
+                self._client.data_object.delete(
+                    class_name=self._collection_name,
+                    uuid=uuid,
+                )

    def search_by_vector(self, query_vector: list[float], **kwargs: Any) -> list[Document]:
        """Look up similar documents by embedding vector in Weaviate."""
@@ -29,8 +29,7 @@ class WordExtractor(BaseExtractor):

            if r.status_code != 200:
                raise ValueError(
-                    "Check the url of your file; returned status code %s"
-                    % r.status_code
+                    f"Check the url of your file; returned status code {r.status_code}"
                )

            self.web_path = self.file_path
@@ -38,7 +37,7 @@ class WordExtractor(BaseExtractor):
            self.temp_file.write(r.content)
            self.file_path = self.temp_file.name
        elif not os.path.isfile(self.file_path):
-            raise ValueError("File path %s is not a valid file or url" % self.file_path)
+            raise ValueError(f"File path {self.file_path} is not a valid file or url")

    def __del__(self) -> None:
        if hasattr(self, "temp_file"):
@@ -44,27 +44,31 @@ class BingSearchTool(BuiltinTool):
            results = []
            if search_results:
                for result in search_results:
+                    url = f': {result["url"]}' if "url" in result else ""
                    results.append(self.create_text_message(
-                        text=f'{result["name"]}: {result["url"]}'
+                        text=f'{result["name"]}{url}'
                    ))


            if entities:
                for entity in entities:
+                    url = f': {entity["url"]}' if "url" in entity else ""
                    results.append(self.create_text_message(
-                        text=f'{entity["name"]}: {entity["url"]}'
+                        text=f'{entity.get("name", "")}{url}'
                    ))

            if news:
                for news_item in news:
+                    url = f': {news_item["url"]}' if "url" in news_item else ""
                    results.append(self.create_text_message(
-                        text=f'{news_item["name"]}: {news_item["url"]}'
+                        text=f'{news_item.get("name", "")}{url}'
                    ))

            if related_searches:
                for related in related_searches:
+                    url = f': {related["displayText"]}' if "displayText" in related else ""
                    results.append(self.create_text_message(
-                        text=f'{related["displayText"]}: {related["webSearchUrl"]}'
+                        text=f'{related.get("displayText", "")}{url}'
                    ))
                    
            return results
@@ -73,7 +77,7 @@ class BingSearchTool(BuiltinTool):
            text = ''
            if search_results:
                for i, result in enumerate(search_results):
-                    text += f'{i+1}: {result["name"]} - {result["snippet"]}\n'
+                    text += f'{i+1}: {result.get("name", "")} - {result.get("snippet", "")}\n'

            if computation and 'expression' in computation and 'value' in computation:
                text += '\nComputation:\n'
@@ -82,17 +86,20 @@ class BingSearchTool(BuiltinTool):
            if entities:
                text += '\nEntities:\n'
                for entity in entities:
-                    text += f'{entity["name"]} - {entity["url"]}\n'
+                    url = f'- {entity["url"]}' if "url" in entity else ""
+                    text += f'{entity.get("name", "")}{url}\n'

            if news:
                text += '\nNews:\n'
                for news_item in news:
-                    text += f'{news_item["name"]} - {news_item["url"]}\n'
+                    url = f'- {news_item["url"]}' if "url" in news_item else ""
+                    text += f'{news_item.get("name", "")}{url}\n'

            if related_searches:
                text += '\n\nRelated Searches:\n'
                for related in related_searches:
-                    text += f'{related["displayText"]} - {related["webSearchUrl"]}\n'
+                    url = f'- {related["webSearchUrl"]}' if "webSearchUrl" in related else ""
+                    text += f'{related.get("displayText", "")}{url}\n'

            return self.create_text_message(text=self.summary(user_id=user_id, content=text))
        
@@ -7,10 +7,10 @@ identity:
    pt_BR: Interpretador de Código
 description:
  human:
-    en_US: Run code and get the result back, when you're using a lower quality model, please make sure there are some tips help LLM to understand how to write the code.
-    zh_Hans: 运行一段代码并返回结果，当您使用较低质量的模型时，请确保有一些提示帮助LLM理解如何编写代码。
-    pt_BR: Execute um trecho de código e obtenha o resultado de volta, quando você estiver usando um modelo de qualidade inferior, certifique-se de que existam algumas dicas para ajudar o LLM a entender como escrever o código.
-  llm: A tool for running code and getting the result back, but only native packages are allowed, network/IO operations are disabled. and you must use print() or console.log() to output the result or result will be empty.
+    en_US: Run code and get the result back. When you're using a lower quality model, please make sure there are some tips help LLM to understand how to write the code.
+    zh_Hans: 运行一段代码并返回结果。当您使用较低质量的模型时，请确保有一些提示帮助LLM理解如何编写代码。
+    pt_BR: Execute um trecho de código e obtenha o resultado de volta. quando você estiver usando um modelo de qualidade inferior, certifique-se de que existam algumas dicas para ajudar o LLM a entender como escrever o código.
+  llm: A tool for running code and getting the result back. Only native packages are allowed, network/IO operations are disabled. and you must use print() or console.log() to output the result or result will be empty.
 parameters:
  - name: language
    type: string
@@ -0,0 +1,3 @@
+<svg xmlns="http://www.w3.org/2000/svg" width="111" height="111" viewBox="0 0 111 111" fill="none">
+    <text x="0" y="90" font-family="Verdana" font-size="85" fill="black">🔥</text>
+</svg>
@@ -0,0 +1,23 @@
+from core.tools.errors import ToolProviderCredentialValidationError
+from core.tools.provider.builtin.firecrawl.tools.crawl import CrawlTool
+from core.tools.provider.builtin_tool_provider import BuiltinToolProviderController
+
+
+class FirecrawlProvider(BuiltinToolProviderController):
+    def _validate_credentials(self, credentials: dict) -> None:
+        try:
+            # Example validation using the Crawl tool
+            CrawlTool().fork_tool_runtime(
+                meta={"credentials": credentials}
+            ).invoke(
+                user_id='',
+                tool_parameters={
+                    "url": "https://example.com",
+                    "includes": '', 
+                    "excludes": '', 
+                    "limit": 1,
+                    "onlyMainContent": True,
+                }
+            )
+        except Exception as e:
+            raise ToolProviderCredentialValidationError(str(e))
@@ -0,0 +1,24 @@
+identity:
+  author: Richards Tu
+  name: firecrawl
+  label:
+    en_US: Firecrawl
+    zh_CN: Firecrawl
+  description:
+    en_US: Firecrawl API integration for web crawling and scraping.
+    zh_CN: Firecrawl API 集成，用于网页爬取和数据抓取。
+  icon: icon.svg
+credentials_for_provider:
+  firecrawl_api_key:
+    type: secret-input
+    required: true
+    label:
+      en_US: Firecrawl API Key
+      zh_CN: Firecrawl API 密钥
+    placeholder:
+      en_US: Please input your Firecrawl API key
+      zh_CN: 请输入您的 Firecrawl API 密钥
+    help:
+      en_US: Get your Firecrawl API key from your Firecrawl account settings.
+      zh_CN: 从您的 Firecrawl 账户设置中获取 Firecrawl API 密钥。
+    url: https://www.firecrawl.dev/account
@@ -0,0 +1,50 @@
+from typing import Any, Union
+
+from firecrawl import FirecrawlApp
+
+from core.tools.entities.tool_entities import ToolInvokeMessage
+from core.tools.tool.builtin_tool import BuiltinTool
+
+
+class CrawlTool(BuiltinTool):
+    def _invoke(self, user_id: str, tool_parameters: dict[str, Any]) -> Union[ToolInvokeMessage, list[ToolInvokeMessage]]:
+        # initialize the app object with the api key
+        app = FirecrawlApp(api_key=self.runtime.credentials['firecrawl_api_key'])
+
+        options = {
+            'crawlerOptions': {
+                'excludes': tool_parameters.get('excludes', '').split(',') if tool_parameters.get('excludes') else [],
+                'includes': tool_parameters.get('includes', '').split(',') if tool_parameters.get('includes') else [],
+                'limit': tool_parameters.get('limit', 5)
+            },
+            'pageOptions': {
+                'onlyMainContent': tool_parameters.get('onlyMainContent', False)
+            }
+        }
+
+        # crawl the url
+        crawl_result = app.crawl_url(
+            url=tool_parameters['url'], 
+            params=options,
+            wait_until_done=True, 
+        )
+        
+        # reformat crawl result
+        crawl_output = "**Crawl Result**\n\n"
+        try:
+            for result in crawl_result:
+                crawl_output += f"**- Title:** {result.get('metadata', {}).get('title', '')}\n"
+                crawl_output += f"**- Description:** {result.get('metadata', {}).get('description', '')}\n"
+                crawl_output += f"**- URL:** {result.get('metadata', {}).get('ogUrl', '')}\n\n"
+                crawl_output += f"**- Web Content:**\n{result.get('markdown', '')}\n\n"
+                crawl_output += "---\n\n"
+        except Exception as e:
+            crawl_output += f"An error occurred: {str(e)}\n"
+            crawl_output += f"**- Title:** {result.get('metadata', {}).get('title', '')}\n"
+            crawl_output += f"**- Description:** {result.get('metadata', {}).get('description','')}\n"
+            crawl_output += f"**- URL:** {result.get('metadata', {}).get('ogUrl', '')}\n\n"
+            crawl_output += f"**- Web Content:**\n{result.get('markdown', '')}\n\n"
+            crawl_output += "---\n\n"
+
+
+        return self.create_text_message(crawl_output)
@@ -0,0 +1,78 @@
+identity:
+  name: crawl
+  author: Richards Tu
+  label:
+    en_US: Crawl
+    zh_Hans: 爬取
+description:
+  human:
+    en_US: Extract data from a website by crawling through a URL.
+    zh_Hans: 通过URL从网站中提取数据。
+  llm: This tool initiates a web crawl to extract data from a specified URL. It allows configuring crawler options such as including or excluding URL patterns, generating alt text for images using LLMs (paid plan required), limiting the maximum number of pages to crawl, and returning only the main content of the page. The tool can return either a list of crawled documents or a list of URLs based on the provided options.
+parameters:
+  - name: url
+    type: string
+    required: true
+    label:
+      en_US: URL to crawl
+      zh_Hans: 要爬取的URL
+    human_description:
+      en_US: The URL of the website to crawl and extract data from.
+      zh_Hans: 要爬取并提取数据的网站URL。
+    llm_description: The URL of the website that needs to be crawled. This is a required parameter.
+    form: llm
+  - name: includes
+    type: string
+    required: false
+    label:
+      en_US: URL patterns to include
+      zh_Hans: 要包含的URL模式
+    human_description:
+      en_US: Specify URL patterns to include during the crawl. Only pages matching these patterns will be crawled, you can use ',' to separate multiple patterns.
+      zh_Hans: 指定爬取过程中要包含的URL模式。只有与这些模式匹配的页面才会被爬取。
+    form: form
+    default: ''
+  - name: excludes
+    type: string
+    required: false
+    label:
+      en_US: URL patterns to exclude
+      zh_Hans: 要排除的URL模式
+    human_description:
+      en_US: Specify URL patterns to exclude during the crawl. Pages matching these patterns will be skipped, you can use ',' to separate multiple patterns.
+      zh_Hans: 指定爬取过程中要排除的URL模式。匹配这些模式的页面将被跳过。
+    form: form
+    default: 'blog/*'
+  - name: limit
+    type: number
+    required: false
+    label:
+      en_US: Maximum number of pages to crawl
+      zh_Hans: 最大爬取页面数
+    human_description:
+      en_US: Specify the maximum number of pages to crawl. The crawler will stop after reaching this limit.
+      zh_Hans: 指定要爬取的最大页面数。爬虫将在达到此限制后停止。
+    form: form
+    min: 1
+    max: 20
+    default: 5
+  - name: onlyMainContent
+    type: boolean
+    required: false
+    label:
+      en_US: Only return the main content of the page
+      zh_Hans: 仅返回页面的主要内容
+    human_description:
+      en_US: If enabled, the crawler will only return the main content of the page, excluding headers, navigation, footers, etc.
+      zh_Hans: 如果启用，爬虫将仅返回页面的主要内容，不包括标题、导航、页脚等。
+    form: form
+    options:
+      - value: true
+        label:
+          en_US: Yes
+          zh_Hans: 是
+      - value: false
+        label:
+          en_US: No
+          zh_Hans: 否
+    default: false
@@ -0,0 +1,21 @@
+<?xml version="1.0" standalone="no"?>
+<!DOCTYPE svg PUBLIC "-//W3C//DTD SVG 20010904//EN"
+ "http://www.w3.org/TR/2001/REC-SVG-20010904/DTD/svg10.dtd">
+<svg version="1.0" xmlns="http://www.w3.org/2000/svg"
+ width="128.000000pt" height="128.000000pt" viewBox="0 0 128.000000 128.000000"
+ preserveAspectRatio="xMidYMid meet">
+
+<g transform="translate(0.000000,128.000000) scale(0.100000,-0.100000)"
+fill="#000000" stroke="none">
+<path d="M0 975 l0 -305 33 1 c54 0 336 35 343 41 3 4 0 57 -7 118 -10 85 -17
+113 -29 120 -47 25 -45 104 2 133 13 8 118 26 246 41 208 26 225 26 248 11 14
+-9 30 -27 36 -41 10 -22 8 -33 -10 -68 l-23 -42 40 -316 40 -315 30 -31 c17
+-17 31 -38 31 -47 0 -25 -27 -72 -46 -79 -35 -13 -450 -59 -476 -53 -52 13
+-70 85 -32 127 10 13 10 33 -1 120 -8 58 -15 111 -15 118 0 16 -31 16 -237 -5
+l-173 -17 0 -243 0 -243 640 0 640 0 0 640 0 640 -640 0 -640 0 0 -305z"/>
+<path d="M578 977 c-128 -16 -168 -24 -168 -35 0 -10 8 -12 28 -8 15 3 90 12
+167 21 167 18 188 23 180 35 -7 12 -1 12 -207 -13z"/>
+<path d="M660 326 c-100 -13 -163 -25 -160 -31 3 -5 14 -9 25 -8 104 11 305
+35 323 39 12 2 22 9 22 14 0 13 -14 12 -210 -14z"/>
+</g>
+</svg>
@@ -0,0 +1,23 @@
+from typing import Any
+
+from core.tools.errors import ToolProviderCredentialValidationError
+from core.tools.provider.builtin.judge0ce.tools.executeCode import ExecuteCodeTool
+from core.tools.provider.builtin_tool_provider import BuiltinToolProviderController
+
+
+class Judge0CEProvider(BuiltinToolProviderController):
+    def _validate_credentials(self, credentials: dict[str, Any]) -> None:
+        try:
+            ExecuteCodeTool().fork_tool_runtime(
+                meta={
+                    "credentials": credentials,
+                }
+            ).invoke(
+                user_id='',
+                tool_parameters={
+                    "source_code": "print('hello world')",
+                    "language_id": 71,
+                },
+            )
+        except Exception as e:
+            raise ToolProviderCredentialValidationError(str(e))
@@ -0,0 +1,29 @@
+identity:
+  author: Richards Tu
+  name: judge0ce
+  label:
+    en_US: Judge0 CE
+    zh_Hans: Judge0 CE
+    pt_BR: Judge0 CE
+  description:
+    en_US: Judge0 CE is an open-source code execution system. Support various languages, including C, C++, Java, Python, Ruby, etc.
+    zh_Hans: Judge0 CE 是一个开源的代码执行系统。支持多种语言，包括 C、C++、Java、Python、Ruby 等。
+    pt_BR: Judge0 CE é um sistema de execução de código de código aberto. Suporta várias linguagens, incluindo C, C++, Java, Python, Ruby, etc.
+  icon: icon.svg
+credentials_for_provider:
+  X-RapidAPI-Key:
+    type: secret-input
+    required: true
+    label:
+      en_US: RapidAPI Key
+      zh_Hans: RapidAPI Key
+      pt_BR: RapidAPI Key
+    help:
+      en_US: RapidAPI Key is required to access the Judge0 CE API.
+      zh_Hans: RapidAPI Key 是访问 Judge0 CE API 所必需的。
+      pt_BR: RapidAPI Key é necessário para acessar a API do Judge0 CE.
+    placeholder:
+      en_US: Enter your RapidAPI Key
+      zh_Hans: 输入你的 RapidAPI Key
+      pt_BR: Insira sua RapidAPI Key
+    url: https://rapidapi.com/judge0-official/api/judge0-ce
@@ -0,0 +1,59 @@
+import json
+from typing import Any, Union
+
+import requests
+from httpx import post
+
+from core.tools.entities.tool_entities import ToolInvokeMessage
+from core.tools.tool.builtin_tool import BuiltinTool
+
+
+class ExecuteCodeTool(BuiltinTool):
+    def _invoke(self, user_id: str, tool_parameters: dict[str, Any]) -> Union[ToolInvokeMessage, list[ToolInvokeMessage]]:
+        """
+        invoke tools
+        """
+        api_key = self.runtime.credentials['X-RapidAPI-Key']
+
+        url = "https://judge0-ce.p.rapidapi.com/submissions"
+
+        querystring = {"base64_encoded": "false", "fields": "*"}
+
+        headers = {
+            "Content-Type": "application/json",
+            "X-RapidAPI-Key": api_key,
+            "X-RapidAPI-Host": "judge0-ce.p.rapidapi.com"
+        }
+
+        payload = {
+            "language_id": tool_parameters['language_id'],
+            "source_code": tool_parameters['source_code'],
+            "stdin": tool_parameters.get('stdin', ''),
+            "expected_output": tool_parameters.get('expected_output', ''),
+            "additional_files": tool_parameters.get('additional_files', ''),
+        }
+
+        response = post(url, data=json.dumps(payload), headers=headers, params=querystring)
+
+        if response.status_code != 201:
+            raise Exception(response.text)
+
+        token = response.json()['token']
+
+        url = f"https://judge0-ce.p.rapidapi.com/submissions/{token}"
+        headers = {
+            "X-RapidAPI-Key": api_key
+        }
+        
+        response = requests.get(url, headers=headers)
+        if response.status_code == 200:
+            result = response.json()
+            return self.create_text_message(text=f"stdout: {result.get('stdout', '')}\n"
+                             f"stderr: {result.get('stderr', '')}\n"
+                             f"compile_output: {result.get('compile_output', '')}\n"
+                             f"message: {result.get('message', '')}\n"
+                             f"status: {result['status']['description']}\n"
+                             f"time: {result.get('time', '')} seconds\n"
+                             f"memory: {result.get('memory', '')} bytes")
+        else:
+            return self.create_text_message(text=f"Error retrieving submission details: {response.text}")
@@ -0,0 +1,67 @@
+identity:
+  name: submitCodeExecutionTask
+  author: Richards Tu
+  label:
+    en_US: Submit Code Execution Task to Judge0 CE and get execution result.
+    zh_Hans: 提交代码执行任务到 Judge0 CE 并获取执行结果。
+description:
+  human:
+    en_US: A tool for executing code and getting the result.
+    zh_Hans: 一个用于执行代码并获取结果的工具。
+  llm: This tool is used for executing code and getting the result.
+parameters:
+  - name: source_code
+    type: string
+    required: true
+    label:
+      en_US: Source Code
+      zh_Hans: 源代码
+    human_description:
+      en_US: The source code to be executed.
+      zh_Hans: 要执行的源代码。
+    llm_description: The source code to be executed.
+    form: llm
+  - name: language_id
+    type: number
+    required: true
+    label:
+      en_US: Language ID
+      zh_Hans: 语言 ID
+    human_description:
+      en_US: The ID of the language in which the source code is written.
+      zh_Hans: 源代码所使用的语言的 ID。
+    llm_description: The ID of the language in which the source code is written. For example, 50 for C++, 71 for Python, etc.
+    form: llm
+  - name: stdin
+    type: string
+    required: false
+    label:
+      en_US: Standard Input
+      zh_Hans: 标准输入
+    human_description:
+      en_US: The standard input to be provided to the program.
+      zh_Hans: 提供给程序的标准输入。
+    llm_description: The standard input to be provided to the program. Optional.
+    form: llm
+  - name: expected_output
+    type: string
+    required: false
+    label:
+      en_US: Expected Output
+      zh_Hans: 期望输出
+    human_description:
+      en_US: The expected output of the program. Used for comparison in some scenarios.
+      zh_Hans: 程序的期望输出。在某些场景下用于比较。
+    llm_description: The expected output of the program. Used for comparison in some scenarios. Optional.
+    form: llm
+  - name: additional_files
+    type: string
+    required: false
+    label:
+      en_US: Additional Files
+      zh_Hans: 附加文件
+    human_description:
+      en_US: Base64 encoded additional files for the submission.
+      zh_Hans: 提交的 Base64 编码的附加文件。
+    llm_description: Base64 encoded additional files for the submission. Optional.
+    form: llm
@@ -222,7 +222,7 @@ class ToolManager:
        return parameter_value

    @classmethod
-    def get_agent_tool_runtime(cls, tenant_id: str, agent_tool: AgentToolEntity) -> Tool:
+    def get_agent_tool_runtime(cls, tenant_id: str, app_id: str, agent_tool: AgentToolEntity) -> Tool:
        """
            get the agent tool runtime
        """
@@ -245,6 +245,7 @@ class ToolManager:
            tool_runtime=tool_entity,
            provider_name=agent_tool.provider_id,
            provider_type=agent_tool.provider_type,
+            identity_id=f'AGENT.{app_id}'
        )
        runtime_parameters = encryption_manager.decrypt_tool_parameters(runtime_parameters)

@@ -252,7 +253,7 @@ class ToolManager:
        return tool_entity

    @classmethod
-    def get_workflow_tool_runtime(cls, tenant_id: str, workflow_tool: ToolEntity):
+    def get_workflow_tool_runtime(cls, tenant_id: str, app_id: str, node_id: str, workflow_tool: ToolEntity):
        """
            get the workflow tool runtime
        """
@@ -277,6 +278,7 @@ class ToolManager:
            tool_runtime=tool_entity,
            provider_name=workflow_tool.provider_id,
            provider_type=workflow_tool.provider_type,
+            identity_id=f'WORKFLOW.{app_id}.{node_id}'
        )

        if runtime_parameters:
@@ -113,12 +113,13 @@ class ToolParameterConfigurationManager(BaseModel):
    tool_runtime: Tool
    provider_name: str
    provider_type: str
+    identity_id: str

    def _deep_copy(self, parameters: dict[str, Any]) -> dict[str, Any]:
        """
        deep copy parameters
        """
-        return {key: value for key, value in parameters.items()}
+        return deepcopy(parameters)
    
    def _merge_parameters(self) -> list[ToolParameter]:
        """
@@ -176,6 +177,8 @@ class ToolParameterConfigurationManager(BaseModel):
        # override parameters
        current_parameters = self._merge_parameters()

+        parameters = self._deep_copy(parameters)
+
        for parameter in current_parameters:
            if parameter.form == ToolParameter.ToolParameterForm.FORM and parameter.type == ToolParameter.ToolParameterType.SECRET_INPUT:
                if parameter.name in parameters:
@@ -194,7 +197,8 @@ class ToolParameterConfigurationManager(BaseModel):
            tenant_id=self.tenant_id, 
            provider=f'{self.provider_type}.{self.provider_name}',
            tool_name=self.tool_runtime.identity.name,
-            cache_type=ToolParameterCacheType.PARAMETER
+            cache_type=ToolParameterCacheType.PARAMETER,
+            identity_id=self.identity_id
        )
        cached_parameters = cache.get()
        if cached_parameters:
@@ -223,7 +227,8 @@ class ToolParameterConfigurationManager(BaseModel):
            tenant_id=self.tenant_id, 
            provider=f'{self.provider_type}.{self.provider_name}',
            tool_name=self.tool_runtime.identity.name,
-            cache_type=ToolParameterCacheType.PARAMETER
+            cache_type=ToolParameterCacheType.PARAMETER,
+            identity_id=self.identity_id
        )
        cache.delete()

@@ -42,20 +42,19 @@ def get_url(url: str, user_agent: str = None) -> str:
    
    supported_content_types = extract_processor.SUPPORT_URL_CONTENT_TYPES + ["text/html"]

-    head_response = requests.head(url, headers=headers, allow_redirects=True, timeout=(5, 10))
+    response = requests.get(url, headers=headers, allow_redirects=True, timeout=(5, 10))

-    if head_response.status_code != 200:
-        return "URL returned status code {}.".format(head_response.status_code)
+    if response.status_code != 200:
+        return "URL returned status code {}.".format(response.status_code)

    # check content-type
-    main_content_type = head_response.headers.get('Content-Type').split(';')[0].strip()
+    main_content_type = response.headers.get('Content-Type').split(';')[0].strip()
    if main_content_type not in supported_content_types:
        return "Unsupported content-type [{}] of URL.".format(main_content_type)

    if main_content_type in extract_processor.SUPPORT_URL_CONTENT_TYPES:
        return ExtractProcessor.load_from_url(url, return_text=True)

-    response = requests.get(url, headers=headers, allow_redirects=True, timeout=(5, 30))
    a = extract_using_readabilipy(response.text)

    if not a['plain_text'] or not a['plain_text'].strip():
@@ -43,7 +43,8 @@ class SystemVariable(Enum):
    """
    QUERY = 'query'
    FILES = 'files'
-    CONVERSATION = 'conversation'
+    CONVERSATION_ID = 'conversation_id'
+    USER_ID = 'user_id'

    @classmethod
    def value_of(cls, value: str) -> 'SystemVariable':
@@ -141,10 +141,10 @@ class CodeNode(BaseNode):
        :return:
        """
        if not isinstance(value, str):
-            raise ValueError(f"{variable} in output form must be a string")
+            raise ValueError(f"Output variable `{variable}` must be a string")

        if len(value) > MAX_STRING_LENGTH:
-            raise ValueError(f'{variable} in output form must be less than {MAX_STRING_LENGTH} characters')
+            raise ValueError(f'The length of output variable `{variable}` must be less than {MAX_STRING_LENGTH} characters')

        return value.replace('\x00', '')

@@ -156,15 +156,15 @@ class CodeNode(BaseNode):
        :return:
        """
        if not isinstance(value, int | float):
-            raise ValueError(f"{variable} in output form must be a number")
+            raise ValueError(f"Output variable `{variable}` must be a number")

        if value > MAX_NUMBER or value < MIN_NUMBER:
-            raise ValueError(f'{variable} in input form is out of range.')
+            raise ValueError(f'Output variable `{variable}` is out of range, it must be between {MIN_NUMBER} and {MAX_NUMBER}.')

        if isinstance(value, float):
            # raise error if precision is too high
            if len(str(value).split('.')[1]) > MAX_PRECISION:
-                raise ValueError(f'{variable} in output form has too high precision.')
+                raise ValueError(f'Output variable `{variable}` has too high precision, it must be less than {MAX_PRECISION} digits.')

        return value

@@ -271,7 +271,7 @@ class CodeNode(BaseNode):

                if len(result[output_name]) > MAX_NUMBER_ARRAY_LENGTH:
                    raise ValueError(
-                        f'{prefix}{dot}{output_name} in output form must be less than {MAX_NUMBER_ARRAY_LENGTH} characters.'
+                        f'The length of output variable `{prefix}{dot}{output_name}` must be less than {MAX_NUMBER_ARRAY_LENGTH} elements.'
                    )

                transformed_result[output_name] = [
@@ -290,7 +290,7 @@ class CodeNode(BaseNode):

                if len(result[output_name]) > MAX_STRING_ARRAY_LENGTH:
                    raise ValueError(
-                        f'{prefix}{dot}{output_name} in output form must be less than {MAX_STRING_ARRAY_LENGTH} characters.'
+                        f'The length of output variable `{prefix}{dot}{output_name}` must be less than {MAX_STRING_ARRAY_LENGTH} elements.'
                    )

                transformed_result[output_name] = [
@@ -309,7 +309,7 @@ class CodeNode(BaseNode):

                if len(result[output_name]) > MAX_OBJECT_ARRAY_LENGTH:
                    raise ValueError(
-                        f'{prefix}{dot}{output_name} in output form must be less than {MAX_OBJECT_ARRAY_LENGTH} characters.'
+                        f'The length of output variable `{prefix}{dot}{output_name}` must be less than {MAX_OBJECT_ARRAY_LENGTH} elements.'
                    )
                
                for i, value in enumerate(result[output_name]):
@@ -36,6 +36,49 @@ class EndNode(BaseNode):
            outputs=outputs
        )

+    @classmethod
+    def extract_generate_nodes(cls, graph: dict, config: dict) -> list[str]:
+        """
+        Extract generate nodes
+        :param graph: graph
+        :param config: node config
+        :return:
+        """
+        node_data = cls._node_data_cls(**config.get("data", {}))
+        node_data = cast(cls._node_data_cls, node_data)
+
+        return cls.extract_generate_nodes_from_node_data(graph, node_data)
+
+    @classmethod
+    def extract_generate_nodes_from_node_data(cls, graph: dict, node_data: EndNodeData) -> list[str]:
+        """
+        Extract generate nodes from node data
+        :param graph: graph
+        :param node_data: node data object
+        :return:
+        """
+        nodes = graph.get('nodes')
+        node_mapping = {node.get('id'): node for node in nodes}
+
+        variable_selectors = node_data.outputs
+
+        generate_nodes = []
+        for variable_selector in variable_selectors:
+            if not variable_selector.value_selector:
+                continue
+
+            node_id = variable_selector.value_selector[0]
+            if node_id != 'sys' and node_id in node_mapping:
+                node = node_mapping[node_id]
+                node_type = node.get('data', {}).get('type')
+                if node_type == NodeType.LLM.value and variable_selector.value_selector[1] == 'text':
+                    generate_nodes.append(node_id)
+
+        # remove duplicates
+        generate_nodes = list(set(generate_nodes))
+
+        return generate_nodes
+
    @classmethod
    def _extract_variable_selector_to_variable_mapping(cls, node_data: BaseNodeData) -> dict[str, list[str]]:
        """
@@ -35,9 +35,15 @@ class HttpRequestNodeData(BaseNodeData):
        type: Literal['none', 'form-data', 'x-www-form-urlencoded', 'raw-text', 'json']
        data: Union[None, str]

+    class Timeout(BaseModel):
+        connect: int
+        read:  int
+        write:  int
+
    method: Literal['get', 'post', 'put', 'patch', 'delete', 'head']
    url: str
    authorization: Authorization
    headers: str
    params: str
-    body: Optional[Body]
+    body: Optional[Body]
+    timeout: Optional[Timeout]
@@ -13,7 +13,6 @@ from core.workflow.entities.variable_pool import ValueType, VariablePool
 from core.workflow.nodes.http_request.entities import HttpRequestNodeData
 from core.workflow.utils.variable_template_parser import VariableTemplateParser

-HTTP_REQUEST_DEFAULT_TIMEOUT = (10, 60)
 MAX_BINARY_SIZE = 1024 * 1024 * 10  # 10MB
 READABLE_MAX_BINARY_SIZE = '10MB'
 MAX_TEXT_SIZE = 1024 * 1024 // 10  # 0.1MB
@@ -137,14 +136,16 @@ class HttpExecutor:
    files: Union[None, dict[str, Any]]
    boundary: str
    variable_selectors: list[VariableSelector]
+    timeout: HttpRequestNodeData.Timeout

-    def __init__(self, node_data: HttpRequestNodeData, variable_pool: Optional[VariablePool] = None):
+    def __init__(self, node_data: HttpRequestNodeData, timeout: HttpRequestNodeData.Timeout, variable_pool: Optional[VariablePool] = None):
        """
        init
        """
        self.server_url = node_data.url
        self.method = node_data.method
        self.authorization = node_data.authorization
+        self.timeout = timeout
        self.params = {}
        self.headers = {}
        self.body = None
@@ -307,7 +308,7 @@ class HttpExecutor:
            'url': self.server_url,
            'headers': headers,
            'params': self.params,
-            'timeout': HTTP_REQUEST_DEFAULT_TIMEOUT,
+            'timeout': (self.timeout.connect, self.timeout.read, self.timeout.write),
            'follow_redirects': True
        }

@@ -1,4 +1,5 @@
 import logging
+import os
 from mimetypes import guess_extension
 from os import path
 from typing import cast
@@ -12,18 +13,49 @@ from core.workflow.nodes.http_request.entities import HttpRequestNodeData
 from core.workflow.nodes.http_request.http_executor import HttpExecutor, HttpExecutorResponse
 from models.workflow import WorkflowNodeExecutionStatus

+MAX_CONNECT_TIMEOUT = int(os.environ.get('HTTP_REQUEST_MAX_CONNECT_TIMEOUT', '300'))
+MAX_READ_TIMEOUT = int(os.environ.get('HTTP_REQUEST_MAX_READ_TIMEOUT', '600'))
+MAX_WRITE_TIMEOUT = int(os.environ.get('HTTP_REQUEST_MAX_WRITE_TIMEOUT', '600'))
+
+HTTP_REQUEST_DEFAULT_TIMEOUT = HttpRequestNodeData.Timeout(connect=min(10, MAX_CONNECT_TIMEOUT),
+                                                           read=min(60, MAX_READ_TIMEOUT),
+                                                           write=min(20, MAX_WRITE_TIMEOUT))
+

 class HttpRequestNode(BaseNode):
    _node_data_cls = HttpRequestNodeData
    node_type = NodeType.HTTP_REQUEST

+    @classmethod
+    def get_default_config(cls) -> dict:
+        return {
+            "type": "http-request",
+            "config": {
+                "method": "get",
+                "authorization": {
+                    "type": "no-auth",
+                },
+                "body": {
+                    "type": "none"
+                },
+                "timeout": {
+                    **HTTP_REQUEST_DEFAULT_TIMEOUT.dict(),
+                    "max_connect_timeout": MAX_CONNECT_TIMEOUT,
+                    "max_read_timeout": MAX_READ_TIMEOUT,
+                    "max_write_timeout": MAX_WRITE_TIMEOUT,
+                }
+            },
+        }
+
    def _run(self, variable_pool: VariablePool) -> NodeRunResult:
        node_data: HttpRequestNodeData = cast(self._node_data_cls, self.node_data)

        # init http executor
        http_executor = None
        try:
-            http_executor = HttpExecutor(node_data=node_data, variable_pool=variable_pool)
+            http_executor = HttpExecutor(node_data=node_data,
+                                         timeout=self._get_request_timeout(node_data),
+                                         variable_pool=variable_pool)

            # invoke http executor
            response = http_executor.invoke()
@@ -38,7 +70,7 @@ class HttpRequestNode(BaseNode):
                error=str(e),
                process_data=process_data
            )
-        
+
        files = self.extract_files(http_executor.server_url, response)

        return NodeRunResult(
@@ -54,6 +86,16 @@ class HttpRequestNode(BaseNode):
            }
        )

+    def _get_request_timeout(self, node_data: HttpRequestNodeData) -> HttpRequestNodeData.Timeout:
+        timeout = node_data.timeout
+        if timeout is None:
+            return HTTP_REQUEST_DEFAULT_TIMEOUT
+
+        timeout.connect = min(timeout.connect, MAX_CONNECT_TIMEOUT)
+        timeout.read = min(timeout.read, MAX_READ_TIMEOUT)
+        timeout.write = min(timeout.write, MAX_WRITE_TIMEOUT)
+        return timeout
+
    @classmethod
    def _extract_variable_selector_to_variable_mapping(cls, node_data: HttpRequestNodeData) -> dict[str, list[str]]:
        """
@@ -62,7 +104,7 @@ class HttpRequestNode(BaseNode):
        :return:
        """
        try:
-            http_executor = HttpExecutor(node_data=node_data)
+            http_executor = HttpExecutor(node_data=node_data, timeout=HTTP_REQUEST_DEFAULT_TIMEOUT)

            variable_selectors = http_executor.variable_selectors

@@ -84,7 +126,7 @@ class HttpRequestNode(BaseNode):
        # if not image, return directly
        if 'image' not in mimetype:
            return files
-        
+
        if mimetype:
            # extract filename from url
            filename = path.basename(url)
@@ -74,6 +74,7 @@ class LLMNode(BaseNode):
                node_data=node_data,
                query=variable_pool.get_variable_value(['sys', SystemVariable.QUERY.value])
                if node_data.memory else None,
+                query_prompt_template=node_data.memory.query_prompt_template if node_data.memory else None,
                inputs=inputs,
                files=files,
                context=context,
@@ -209,6 +210,17 @@ class LLMNode(BaseNode):

            inputs[variable_selector.variable] = variable_value

+        memory = node_data.memory
+        if memory and memory.query_prompt_template:
+            query_variable_selectors = (VariableTemplateParser(template=memory.query_prompt_template)
+                                        .extract_variable_selectors())
+            for variable_selector in query_variable_selectors:
+                variable_value = variable_pool.get_variable_value(variable_selector.value_selector)
+                if variable_value is None:
+                    raise ValueError(f'Variable {variable_selector.variable} not found')
+
+                inputs[variable_selector.variable] = variable_value
+
        return inputs

    def _fetch_files(self, node_data: LLMNodeData, variable_pool: VariablePool) -> list[FileVar]:
@@ -302,7 +314,8 @@ class LLMNode(BaseNode):

        return None

-    def _fetch_model_config(self, node_data_model: ModelConfig) -> tuple[ModelInstance, ModelConfigWithCredentialsEntity]:
+    def _fetch_model_config(self, node_data_model: ModelConfig) -> tuple[
+        ModelInstance, ModelConfigWithCredentialsEntity]:
        """
        Fetch model config
        :param node_data_model: node data model
@@ -385,7 +398,7 @@ class LLMNode(BaseNode):
            return None

        # get conversation id
-        conversation_id = variable_pool.get_variable_value(['sys', SystemVariable.CONVERSATION.value])
+        conversation_id = variable_pool.get_variable_value(['sys', SystemVariable.CONVERSATION_ID.value])
        if conversation_id is None:
            return None

@@ -407,6 +420,7 @@ class LLMNode(BaseNode):

    def _fetch_prompt_messages(self, node_data: LLMNodeData,
                               query: Optional[str],
+                               query_prompt_template: Optional[str],
                               inputs: dict[str, str],
                               files: list[FileVar],
                               context: Optional[str],
@@ -417,6 +431,7 @@ class LLMNode(BaseNode):
        Fetch prompt messages
        :param node_data: node data
        :param query: query
+        :param query_prompt_template: query prompt template
        :param inputs: inputs
        :param files: files
        :param context: context
@@ -433,7 +448,8 @@ class LLMNode(BaseNode):
            context=context,
            memory_config=node_data.memory,
            memory=memory,
-            model_config=model_config
+            model_config=model_config,
+            query_prompt_template=query_prompt_template,
        )
        stop = model_config.stop

@@ -539,12 +555,22 @@ class LLMNode(BaseNode):
        for variable_selector in variable_selectors:
            variable_mapping[variable_selector.variable] = variable_selector.value_selector

+        memory = node_data.memory
+        if memory and memory.query_prompt_template:
+            query_variable_selectors = (VariableTemplateParser(template=memory.query_prompt_template)
+                                        .extract_variable_selectors())
+            for variable_selector in query_variable_selectors:
+                variable_mapping[variable_selector.variable] = variable_selector.value_selector
+
        if node_data.context.enabled:
            variable_mapping['#context#'] = node_data.context.variable_selector

        if node_data.vision.enabled:
            variable_mapping['#files#'] = ['sys', SystemVariable.FILES.value]

+        if node_data.memory:
+            variable_mapping['#sys.query#'] = ['sys', SystemVariable.QUERY.value]
+
        return variable_mapping

    @classmethod
@@ -79,7 +79,6 @@ class QuestionClassifierNode(LLMNode):
                    prompt_messages=prompt_messages
                ),
                'usage': jsonable_encoder(usage),
-                'topics': categories[0] if categories else ''
            }
            outputs = {
                'class_name': categories[0] if categories else ''
@@ -1,8 +1,6 @@
-from typing import cast

-from core.app.app_config.entities import VariableEntity
 from core.workflow.entities.base_node_data_entities import BaseNodeData
-from core.workflow.entities.node_entities import NodeRunResult, NodeType, SystemVariable
+from core.workflow.entities.node_entities import NodeRunResult, NodeType
 from core.workflow.entities.variable_pool import VariablePool
 from core.workflow.nodes.base_node import BaseNode
 from core.workflow.nodes.start.entities import StartNodeData
@@ -19,17 +17,10 @@ class StartNode(BaseNode):
        :param variable_pool: variable pool
        :return:
        """
-        node_data = self.node_data
-        node_data = cast(self._node_data_cls, node_data)
-        variables = node_data.variables
-
        # Get cleaned inputs
-        cleaned_inputs = self._get_cleaned_inputs(variables, variable_pool.user_inputs)
+        cleaned_inputs = variable_pool.user_inputs

        for var in variable_pool.system_variables:
-            if var == SystemVariable.CONVERSATION:
-                continue
-
            cleaned_inputs['sys.' + var.value] = variable_pool.system_variables[var]

        return NodeRunResult(
@@ -38,42 +29,6 @@ class StartNode(BaseNode):
            outputs=cleaned_inputs
        )

-    def _get_cleaned_inputs(self, variables: list[VariableEntity], user_inputs: dict):
-        if user_inputs is None:
-            user_inputs = {}
-
-        filtered_inputs = {}
-
-        for variable_config in variables:
-            variable = variable_config.variable
-
-            if variable not in user_inputs or not user_inputs[variable]:
-                if variable_config.required:
-                    raise ValueError(f"Input form variable {variable} is required")
-                else:
-                    filtered_inputs[variable] = variable_config.default if variable_config.default is not None else ""
-                    continue
-
-            value = user_inputs[variable]
-
-            if value:
-                if not isinstance(value, str):
-                    raise ValueError(f"{variable} in input form must be a string")
-
-            if variable_config.type == VariableEntity.Type.SELECT:
-                options = variable_config.options if variable_config.options is not None else []
-                if value not in options:
-                    raise ValueError(f"{variable} in input form must be one of the following: {options}")
-            else:
-                if variable_config.max_length is not None:
-                    max_length = variable_config.max_length
-                    if len(value) > max_length:
-                        raise ValueError(f'{variable} in input form must be less than {max_length} characters')
-
-            filtered_inputs[variable] = value.replace('\x00', '') if value else None
-
-        return filtered_inputs
-
    @classmethod
    def _extract_variable_selector_to_variable_mapping(cls, node_data: BaseNodeData) -> dict[str, list[str]]:
        """
@@ -39,7 +39,8 @@ class ToolNode(BaseNode):
        parameters = self._generate_parameters(variable_pool, node_data)
        # get tool runtime
        try:
-            tool_runtime = ToolManager.get_workflow_tool_runtime(self.tenant_id, node_data)
+            self.app_id
+            tool_runtime = ToolManager.get_workflow_tool_runtime(self.tenant_id, self.app_id, self.node_id, node_data)
        except Exception as e:
            return NodeRunResult(
                status=WorkflowNodeExecutionStatus.FAILED,
@@ -22,5 +22,6 @@ def handle(sender, **kwargs):
                tool_runtime=tool_runtime,
                provider_name=tool_entity.provider_name,
                provider_type=tool_entity.provider_type,
+                identity_id=f'WORKFLOW.{app.id}.{node_data.get("id")}'
            )
            manager.delete_tool_parameters_cache()
@@ -1,70 +1,42 @@
-import os
-import shutil
 from collections.abc import Generator
-from contextlib import closing
-from datetime import datetime, timedelta, timezone
 from typing import Union

-import boto3
-from azure.storage.blob import AccountSasPermissions, BlobServiceClient, ResourceTypes, generate_account_sas
-from botocore.client import Config
-from botocore.exceptions import ClientError
 from flask import Flask

+from extensions.storage.aliyun_storage import AliyunStorage
+from extensions.storage.azure_storage import AzureStorage
+from extensions.storage.google_storage import GoogleStorage
+from extensions.storage.local_storage import LocalStorage
+from extensions.storage.s3_storage import S3Storage
+

 class Storage:
    def __init__(self):
-        self.storage_type = None
-        self.bucket_name = None
-        self.client = None
-        self.folder = None
+        self.storage_runner = None

    def init_app(self, app: Flask):
-        self.storage_type = app.config.get('STORAGE_TYPE')
-        if self.storage_type == 's3':
-            self.bucket_name = app.config.get('S3_BUCKET_NAME')
-            self.client = boto3.client(
-                's3',
-                aws_secret_access_key=app.config.get('S3_SECRET_KEY'),
-                aws_access_key_id=app.config.get('S3_ACCESS_KEY'),
-                endpoint_url=app.config.get('S3_ENDPOINT'),
-                region_name=app.config.get('S3_REGION'),
-                config=Config(s3={'addressing_style': app.config.get('S3_ADDRESS_STYLE')})
+        storage_type = app.config.get('STORAGE_TYPE')
+        if storage_type == 's3':
+            self.storage_runner = S3Storage(
+                app=app
            )
-        elif self.storage_type == 'azure-blob':
-            self.bucket_name = app.config.get('AZURE_BLOB_CONTAINER_NAME')
-            sas_token = generate_account_sas(
-                account_name=app.config.get('AZURE_BLOB_ACCOUNT_NAME'),
-                account_key=app.config.get('AZURE_BLOB_ACCOUNT_KEY'),
-                resource_types=ResourceTypes(service=True, container=True, object=True),
-                permission=AccountSasPermissions(read=True, write=True, delete=True, list=True, add=True, create=True),
-                expiry=datetime.now(timezone.utc).replace(tzinfo=None) + timedelta(hours=1)
+        elif storage_type == 'azure-blob':
+            self.storage_runner = AzureStorage(
+                app=app
+            )
+        elif storage_type == 'aliyun-oss':
+            self.storage_runner = AliyunStorage(
+                app=app
+            )
+        elif storage_type == 'google-storage':
+            self.storage_runner = GoogleStorage(
+                app=app
            )
-            self.client = BlobServiceClient(account_url=app.config.get('AZURE_BLOB_ACCOUNT_URL'),
-                                            credential=sas_token)
-
        else:
-            self.folder = app.config.get('STORAGE_LOCAL_PATH')
-            if not os.path.isabs(self.folder):
-                self.folder = os.path.join(app.root_path, self.folder)
+            self.storage_runner = LocalStorage(app=app)

    def save(self, filename, data):
-        if self.storage_type == 's3':
-            self.client.put_object(Bucket=self.bucket_name, Key=filename, Body=data)
-        elif self.storage_type == 'azure-blob':
-            blob_container = self.client.get_container_client(container=self.bucket_name)
-            blob_container.upload_blob(filename, data)
-        else:
-            if not self.folder or self.folder.endswith('/'):
-                filename = self.folder + filename
-            else:
-                filename = self.folder + '/' + filename
-
-            folder = os.path.dirname(filename)
-            os.makedirs(folder, exist_ok=True)
-
-            with open(os.path.join(os.getcwd(), filename), "wb") as f:
-                f.write(data)
+        self.storage_runner.save(filename, data)

    def load(self, filename: str, stream: bool = False) -> Union[bytes, Generator]:
        if stream:
@@ -73,118 +45,19 @@ class Storage:
            return self.load_once(filename)

    def load_once(self, filename: str) -> bytes:
-        if self.storage_type == 's3':
-            try:
-                with closing(self.client) as client:
-                    data = client.get_object(Bucket=self.bucket_name, Key=filename)['Body'].read()
-            except ClientError as ex:
-                if ex.response['Error']['Code'] == 'NoSuchKey':
-                    raise FileNotFoundError("File not found")
-                else:
-                    raise
-        elif self.storage_type == 'azure-blob':
-            blob = self.client.get_container_client(container=self.bucket_name)
-            blob = blob.get_blob_client(blob=filename)
-            data = blob.download_blob().readall()
-        else:
-            if not self.folder or self.folder.endswith('/'):
-                filename = self.folder + filename
-            else:
-                filename = self.folder + '/' + filename
-
-            if not os.path.exists(filename):
-                raise FileNotFoundError("File not found")
-
-            with open(filename, "rb") as f:
-                data = f.read()
-
-        return data
+        return self.storage_runner.load_once(filename)

    def load_stream(self, filename: str) -> Generator:
-        def generate(filename: str = filename) -> Generator:
-            if self.storage_type == 's3':
-                try:
-                    with closing(self.client) as client:
-                        response = client.get_object(Bucket=self.bucket_name, Key=filename)
-                        for chunk in response['Body'].iter_chunks():
-                            yield chunk
-                except ClientError as ex:
-                    if ex.response['Error']['Code'] == 'NoSuchKey':
-                        raise FileNotFoundError("File not found")
-                    else:
-                        raise
-            elif self.storage_type == 'azure-blob':
-                blob = self.client.get_blob_client(container=self.bucket_name, blob=filename)
-                with closing(blob.download_blob()) as blob_stream:
-                    while chunk := blob_stream.readall(4096):
-                        yield chunk
-            else:
-                if not self.folder or self.folder.endswith('/'):
-                    filename = self.folder + filename
-                else:
-                    filename = self.folder + '/' + filename
-
-                if not os.path.exists(filename):
-                    raise FileNotFoundError("File not found")
-
-                with open(filename, "rb") as f:
-                    while chunk := f.read(4096):  # Read in chunks of 4KB
-                        yield chunk
-
-        return generate()
+        return self.storage_runner.load_stream(filename)

    def download(self, filename, target_filepath):
-        if self.storage_type == 's3':
-            with closing(self.client) as client:
-                client.download_file(self.bucket_name, filename, target_filepath)
-        elif self.storage_type == 'azure-blob':
-            blob = self.client.get_blob_client(container=self.bucket_name, blob=filename)
-            with open(target_filepath, "wb") as my_blob:
-                blob_data = blob.download_blob()
-                blob_data.readinto(my_blob)
-        else:
-            if not self.folder or self.folder.endswith('/'):
-                filename = self.folder + filename
-            else:
-                filename = self.folder + '/' + filename
-
-            if not os.path.exists(filename):
-                raise FileNotFoundError("File not found")
-
-            shutil.copyfile(filename, target_filepath)
+        self.storage_runner.download(filename, target_filepath)

    def exists(self, filename):
-        if self.storage_type == 's3':
-            with closing(self.client) as client:
-                try:
-                    client.head_object(Bucket=self.bucket_name, Key=filename)
-                    return True
-                except:
-                    return False
-        elif self.storage_type == 'azure-blob':
-            blob = self.client.get_blob_client(container=self.bucket_name, blob=filename)
-            return blob.exists()
-        else:
-            if not self.folder or self.folder.endswith('/'):
-                filename = self.folder + filename
-            else:
-                filename = self.folder + '/' + filename
-
-            return os.path.exists(filename)
+        return self.storage_runner.exists(filename)

    def delete(self, filename):
-        if self.storage_type == 's3':
-            self.client.delete_object(Bucket=self.bucket_name, Key=filename)
-        elif self.storage_type == 'azure-blob':
-            blob_container = self.client.get_container_client(container=self.bucket_name)
-            blob_container.delete_blob(filename)
-        else:
-            if not self.folder or self.folder.endswith('/'):
-                filename = self.folder + filename
-            else:
-                filename = self.folder + '/' + filename
-            if os.path.exists(filename):
-                os.remove(filename)
+        return self.storage_runner.delete(filename)


 storage = Storage()
@@ -0,0 +1,48 @@
+from collections.abc import Generator
+from contextlib import closing
+
+import oss2 as aliyun_s3
+from flask import Flask
+
+from extensions.storage.base_storage import BaseStorage
+
+
+class AliyunStorage(BaseStorage):
+    """Implementation for aliyun storage.
+    """
+
+    def __init__(self, app: Flask):
+        super().__init__(app)
+        app_config = self.app.config
+        self.bucket_name = app_config.get('ALIYUN_OSS_BUCKET_NAME')
+        self.client = aliyun_s3.Bucket(
+            aliyun_s3.Auth(app_config.get('ALIYUN_OSS_ACCESS_KEY'), app_config.get('ALIYUN_OSS_SECRET_KEY')),
+            app_config.get('ALIYUN_OSS_ENDPOINT'),
+            self.bucket_name,
+            connect_timeout=30
+        )
+
+    def save(self, filename, data):
+        self.client.put_object(filename, data)
+
+    def load_once(self, filename: str) -> bytes:
+        with closing(self.client.get_object(filename)) as obj:
+            data = obj.read()
+        return data
+
+    def load_stream(self, filename: str) -> Generator:
+        def generate(filename: str = filename) -> Generator:
+            with closing(self.client.get_object(filename)) as obj:
+                while chunk := obj.read(4096):
+                    yield chunk
+
+        return generate()
+
+    def download(self, filename, target_filepath):
+        self.client.get_object_to_file(filename, target_filepath)
+
+    def exists(self, filename):
+        return self.client.object_exists(filename)
+
+    def delete(self, filename):
+        self.client.delete_object(filename)
@@ -0,0 +1,58 @@
+from collections.abc import Generator
+from contextlib import closing
+from datetime import datetime, timedelta, timezone
+
+from azure.storage.blob import AccountSasPermissions, BlobServiceClient, ResourceTypes, generate_account_sas
+from flask import Flask
+
+from extensions.storage.base_storage import BaseStorage
+
+
+class AzureStorage(BaseStorage):
+    """Implementation for azure storage.
+    """
+    def __init__(self, app: Flask):
+        super().__init__(app)
+        app_config = self.app.config
+        self.bucket_name = app_config.get('AZURE_STORAGE_CONTAINER_NAME')
+        sas_token = generate_account_sas(
+            account_name=app_config.get('AZURE_BLOB_ACCOUNT_NAME'),
+            account_key=app_config.get('AZURE_BLOB_ACCOUNT_KEY'),
+            resource_types=ResourceTypes(service=True, container=True, object=True),
+            permission=AccountSasPermissions(read=True, write=True, delete=True, list=True, add=True, create=True),
+            expiry=datetime.now(timezone.utc).replace(tzinfo=None) + timedelta(hours=1)
+        )
+        self.client = BlobServiceClient(account_url=app_config.get('AZURE_BLOB_ACCOUNT_URL'),
+                                        credential=sas_token)
+    def save(self, filename, data):
+        blob_container = self.client.get_container_client(container=self.bucket_name)
+        blob_container.upload_blob(filename, data)
+
+    def load_once(self, filename: str) -> bytes:
+        blob = self.client.get_container_client(container=self.bucket_name)
+        blob = blob.get_blob_client(blob=filename)
+        data = blob.download_blob().readall()
+        return data
+
+    def load_stream(self, filename: str) -> Generator:
+        def generate(filename: str = filename) -> Generator:
+            blob = self.client.get_blob_client(container=self.bucket_name, blob=filename)
+            with closing(blob.download_blob()) as blob_stream:
+                while chunk := blob_stream.readall(4096):
+                    yield chunk
+
+        return generate()
+
+    def download(self, filename, target_filepath):
+        blob = self.client.get_blob_client(container=self.bucket_name, blob=filename)
+        with open(target_filepath, "wb") as my_blob:
+            blob_data = blob.download_blob()
+            blob_data.readinto(my_blob)
+
+    def exists(self, filename):
+        blob = self.client.get_blob_client(container=self.bucket_name, blob=filename)
+        return blob.exists()
+
+    def delete(self, filename):
+        blob_container = self.client.get_container_client(container=self.bucket_name)
+        blob_container.delete_blob(filename)
--- a/Show More
+++ b/Show More
Author	SHA1	Message	Date
takatostandGitHub	93393e005e	version to 0.6.6 (#4050 )	2024-05-02 16:06:40 +08:00
Bowen LiangandGitHub	4ea2755fce	test: remove explicit env settings for CI pytests (#4041 )	2024-05-02 00:49:39 +08:00
dependabot[bot]GitHubdependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>	ecb51a83d4	chore(deps): bump semver from 5.7.1 to 5.7.2 in /web (#4022 ) Signed-off-by: dependabot[bot] <support@github.com> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>	2024-04-30 18:47:05 +08:00
Bowen LiangandGitHub	093b5c0e63	fix: typo of jinja2 (#4019 )	2024-04-30 18:39:02 +08:00
Joel GitHub nite-knite	bf42b0ae44	fix: lodash version has warning (#4020 ) Co-authored-by: nite-knite <nkCoding@gmail.com>	2024-04-30 18:11:49 +08:00
dependabot[bot]	342b4fd19d	chore(deps): bump word-wrap from 1.2.3 to 1.2.5 in /web Bumps [word-wrap](https://github.com/jonschlinkert/word-wrap) from 1.2.3 to 1.2.5. - [Release notes](https://github.com/jonschlinkert/word-wrap/releases) - [Commits](https://github.com/jonschlinkert/word-wrap/compare/1.2.3...1.2.5) --- updated-dependencies: - dependency-name: word-wrap dependency-type: indirect ... Signed-off-by: dependabot[bot] <support@github.com>	2024-04-30 09:39:10 +00:00
orangeclk GitHub 陈力坤	cbdb861ee4	add glm-3-turbo max_tokens parameter setting (#4017 ) Co-authored-by: 陈力坤 <likunchen@caixin.com>	2024-04-30 17:08:04 +08:00
sinoandGitHub	da5a8b9a59	feat: support question classifier node output (#4000 )	2024-04-30 17:07:29 +08:00
WeaxsandGitHub	1e6e8b446d	feat: support minimax abab6.5, abab6.5s (#4012 )	2024-04-30 17:02:01 +08:00
JoelandGitHub	c1fdaa6ae0	fix: prompt undefined caused match problem (#4010 )	2024-04-30 16:31:36 +08:00
Bowen LiangandGitHub	142814d451	chore: skip deprecated field_schema param in creating payload index on Qdrant (#3903 )	2024-04-30 16:16:10 +08:00
crazywoolaandGitHub	704755d005	fix: submitCodeExecutionTask (#4006 )	2024-04-30 16:01:03 +08:00
Richards Tu GitHub crazywoola	d1263700c0	Update the description and labels in Judge0ce tool (#3990 ) Co-authored-by: crazywoola <427733928@qq.com>	2024-04-30 14:58:29 +08:00
Rhon Joe GitHub rongjun.qiu	0704fe9695	fix(web): copy button visible at chat page normally (#4005 ) Co-authored-by: rongjun.qiu <qiurj@hengtonggroup.com.cn>	2024-04-30 14:55:57 +08:00
Kei YAMAZAKIandGitHub	1d3f1d88ef	Enabled Notion integration setup in Docker Compose Deployment (#3919 )	2024-04-30 14:48:39 +08:00
zxhlyhandGitHub	8b3edac091	fix: prompt editor insert quickly (#4004 )	2024-04-30 14:25:21 +08:00
zxhlyhandGitHub	05cab85579	fix: workflow disable shortcuts when feature panel occured (#4001 )	2024-04-30 13:35:49 +08:00
YeuolyandGitHub	b72fbe200d	chore: add sandbox tag (#3997 )	2024-04-30 12:35:19 +08:00
crazywoolaandGitHub	b1194da6a5	fix: ci (#3983 )	2024-04-29 18:59:37 +08:00
JyongandGitHub	338e4669e5	add storage factory (#3922 )	2024-04-29 18:22:03 +08:00
crazywoolaandGitHub	c5e2659771	Feat/install process refinement (#3982 )	2024-04-29 17:55:52 +08:00
JyongandGitHub	1d432728ac	add default value for QDRANT_GRPC_PORT (#3976 )	2024-04-29 15:28:34 +08:00
KVOJJJinandGitHub	2fd702a319	Fix: password check in page of install (#3978 )	2024-04-29 15:27:45 +08:00
Richards Tu GitHub crazywoola Yeuoly	f26ad16af7	Add new tool: Firecrawl (#3819 ) Co-authored-by: crazywoola <427733928@qq.com> Co-authored-by: Yeuoly <admin@srmxy.cn>	2024-04-29 14:20:36 +08:00
majian GitHub Yeuoly	8f2ae51fe5	feat: add support for request timeout settings in the HTTP request node. (#3854 ) Co-authored-by: Yeuoly <admin@srmxy.cn>	2024-04-29 13:59:07 +08:00
JoshuaandGitHub	2f84d00300	fix-nvidia-llama3 (#3973 )	2024-04-29 13:41:15 +08:00
takatostandGitHub	b82a2d97ef	fix: db connections not being released during workflow execution (#3971 )	2024-04-29 12:42:09 +08:00
JyongandGitHub	3e9dbe3e0a	add pgvecto_rs support and upgrade SQLAlchemy (#3833 )	2024-04-29 11:58:17 +08:00
羊羽 GitHub baxiang	975b2fb79e	delete duplicate check get_dataset (#3966 ) Co-authored-by: baxiang <baxiang@lixiang.com>	2024-04-29 11:57:26 +08:00
JoelandGitHub	fa509ce64e	feat: rename var name sync to used jinjia code (#3964 )	2024-04-29 11:34:30 +08:00
TinsFoxandGitHub	99292edd46	chore: update `@types/react` (#3939 )	2024-04-28 19:01:09 +08:00
Joel GitHub StyleZhang	3e992cb23c	feat: code transform node editor support insert var by add slash or left brace (#3946 ) Co-authored-by: StyleZhang <jasonapring2015@outlook.com>	2024-04-28 17:51:58 +08:00
YeuolyandGitHub	e7b4d024ee	optimize: code node has a bad error message (#3949 )	2024-04-28 17:40:29 +08:00
takatost GitHub JzoNg	ff67a6d338	feat: llm text stream support for workflow app (#3798 ) Co-authored-by: JzoNg <jzongcode@gmail.com>	2024-04-28 17:37:00 +08:00
zxhlyhandGitHub	8e4989ed03	feat: workflow remove preview mode (#3941 )	2024-04-28 17:09:56 +08:00
呆萌闷油瓶andGitHub	0940f01634	enhancement:support Qdrant gRPC mode (#3929 )	2024-04-28 15:33:32 +08:00
majianandGitHub	9d1cb1bc92	improvement: Optimizing the experience of the app list page (#3885 )	2024-04-28 13:52:45 +08:00
Pascal MandGitHub	0ca4e30b19	feat: add start commands to devcontainer (#3902 )	2024-04-28 12:30:56 +08:00
JoelandGitHub	ba88f8a6f0	fix: code full screen in web app cause error (#3935 )	2024-04-28 11:59:57 +08:00
studyingloverandGitHub	aefe0cbf51	fix: api doc example error (#3925 )	2024-04-28 10:18:07 +08:00
miendinh GitHub miendinh	9ad489d133	feat: Add google storage support (#3887 ) Co-authored-by: miendinh <miendinh@users.noreply.github.com>	2024-04-27 18:26:52 +08:00
Bowen LiangandGitHub	661b30784e	chore: skip warning messages when pytest auto-collecting the vdb test class by removing `Test` prefix (#3906 )	2024-04-27 16:36:09 +08:00
longzhihunandGitHub	43a5ba9415	feat: add support for Bedrock LLAMA3 (#3890 )	2024-04-27 13:13:09 +08:00
TinsFoxandGitHub	08a65d74d5	fix: hydration warning (#3897 )	2024-04-26 21:34:29 +08:00
Garfield DaiandGitHub	cefe156811	feat: replicate supports default version. (#3884 )	2024-04-26 21:16:22 +08:00
yongjer GitHub crazywoola crazywoola	3b5b4d628b	Add support for Traditional Chinese language (#3899 ) Co-authored-by: crazywoola <100913391+crazywoola@users.noreply.github.com> Co-authored-by: crazywoola <427733928@qq.com>	2024-04-26 21:10:23 +08:00
TinsFoxandGitHub	8746e48df0	chore: integrate code-inspector-plugin (#3900 )	2024-04-26 21:00:29 +08:00
JyongandGitHub	0ec8b57825	add together ai model setting (#3895 )	2024-04-26 20:43:17 +08:00
Bowen LiangandGitHub	045827043d	test: improve vector store tests (#3855 )	2024-04-26 19:18:42 +08:00
yaleiandGitHub	4d66a86579	fix: fetch page name of notion wiki (#3847 )	2024-04-26 18:04:37 +08:00
miendinh GitHub miendinh	2a8881d0e8	fix: tool webscraper - too many redirects in case target url does not… (#3831 ) Co-authored-by: miendinh <miendinh@users.noreply.github.com>	2024-04-26 17:58:46 +08:00
akouandGitHub	ffc60bb917	add the comment in entrypoint.sh (#3882 )	2024-04-26 17:19:49 +08:00
EverandGitHub	2e454c770b	fix: copy invite link for HTTPS has deplicate origin (#3877 )	2024-04-26 15:19:30 +08:00
JoelandGitHub	7d711135bc	fix: full screen editor not follow panel width (#3876 )	2024-04-26 14:23:13 +08:00
Charlie.Wei GitHubluowei <glpat-EjySCyNjWiLqAED-YmwM>crazywoola crazywoola	f62b2b5b45	optimize the knowledge failed documents query (#3870 ) Co-authored-by: luowei <glpat-EjySCyNjWiLqAED-YmwM> Co-authored-by: crazywoola <427733928@qq.com> Co-authored-by: crazywoola <100913391+crazywoola@users.noreply.github.com>	2024-04-26 11:47:23 +08:00
Bowen LiangandGitHub	7919596a21	fix: UP031 style rule violation (#3866 )	2024-04-26 11:24:08 +08:00
JoelandGitHub	9b4898efeb	fix: chat api doc not show title in english vision (#3864 )	2024-04-26 10:32:45 +08:00
Bowen LiangandGitHub	45dd1683fd	test: add tests covering all methods of vector store (#3849 )	2024-04-25 22:27:30 +08:00
takatostandGitHub	8bca908f15	refactor: config file (#3852 )	2024-04-25 22:26:45 +08:00
Garfield DaiandGitHub	9cbb8ddd7f	fix: billing tenant account role. (#3850 )	2024-04-25 21:55:08 +08:00
Jingpan Xiong GitHub jingsi	1be222af2e	fix: using api can not execute relyt vector database (#3766 ) Co-authored-by: jingsi <jingsi@leadincloud.com>	2024-04-25 19:46:20 +08:00
Richards Tu GitHub crazywoola	bf9fc8fef4	Reduce tool redundancy for [Judge0 CE] (#3837 ) Co-authored-by: crazywoola <427733928@qq.com>	2024-04-25 19:20:54 +08:00
Bowen LiangandGitHub	86e7330fa2	test: refactor vdb tests by visitor design pattern (#3838 )	2024-04-25 18:55:49 +08:00
takatostandGitHub	34bfb715e1	fix: citations always appear in the chatflow app (#3844 )	2024-04-25 18:31:38 +08:00
JoelandGitHub	019d7069f8	fix: debug run not show total right tokens (#3843 )	2024-04-25 18:22:30 +08:00
Bowen LiangandGitHub	c54fcfb45d	extract enum type for tenant account role (#3788 )	2024-04-25 18:20:08 +08:00
zxhlyhandGitHub	cde87cb225	fix: model parameter default value (#3841 )	2024-04-25 18:04:37 +08:00
takatost GitHub Joel	12435774ca	feat: query prompt template support in chatflow (#3791 ) Co-authored-by: Joel <iamjoel007@gmail.com>	2024-04-25 18:01:53 +08:00
Henrybit GitHub henrybit	80b9507e7a	feat: add aliyun oss storage (#3690 ) Co-authored-by: henrybit <qipenghui3056@sina.com>	2024-04-25 16:57:19 +08:00
takatostandGitHub	0ac0f0ffd0	version to 0.6.5 (#3834 )	2024-04-25 16:50:37 +08:00
KVOJJJinandGitHub	3d14aba4b4	Fix: event of click away in message-log-modal (#3828 )	2024-04-25 15:58:03 +08:00
Leon capandGitHub	64f694865c	Update EN,KL,JA,FR,ES documentation Llma2 to Llama3 model support (#3827 )	2024-04-25 15:52:00 +08:00
zxhlyhandGitHub	d36b728088	fix: workflow sync data (#3824 )	2024-04-25 14:02:06 +08:00
KVOJJJinandGitHub	1a7b4c42ab	fix: event of keyboard "enter" in text generator app (#3823 )	2024-04-25 13:58:06 +08:00
JoelandGitHub	2a64ce740e	chore: remove anthropic pay entrance (#3822 )	2024-04-25 13:18:59 +08:00
呆萌闷油瓶andGitHub	78988ed60e	fix:still enable SSL verification when using qdrant based on HTTP protocol (#3805 )	2024-04-25 13:04:31 +08:00
YeuolyandGitHub	2832adda88	fix: missing url field when searching special keywords (#3820 )	2024-04-25 12:33:58 +08:00
takatostandGitHub	a4e4fb4094	fix: credentials validate failed for groqcloud model provider (#3817 )	2024-04-25 12:09:44 +08:00
YidaHuandGitHub	777ec64635	feat: add log_file environment variable (#3793 )	2024-04-24 21:55:14 +08:00
Bowen LiangandGitHub	9cec8c1750	test: add unit tests for vector stores of Milvus, Qdrant and Weaviate (#3688 )	2024-04-24 21:52:42 +08:00
Bowen LiangandGitHub	8ca5aa1190	use pymilvus 2.3.7 (#3790 )	2024-04-24 18:37:08 +08:00
takatost GitHub Joel	4d8f1b9ca4	feat: test all unit tests (#3787 ) Co-authored-by: Joel <iamjoel007@gmail.com>	2024-04-24 17:33:01 +08:00
takatost GitHub Joel	3da179f77b	feat: add conversation_id and user_id in chatflow/workflow system vars (#3771 ) Co-authored-by: Joel <iamjoel007@gmail.com>	2024-04-24 17:20:01 +08:00
Bowen LiangandGitHub	a34e8cb0bd	test: add test for PKCS1OAEP_Cipher with gmpy2 (#3760 )	2024-04-24 17:15:31 +08:00
KVOJJJinandGitHub	b249767c5c	Fix: redirection of app remove (#3770 )	2024-04-24 17:11:51 +08:00
JoelandGitHub	89a7434565	fix: handle inputs show the focus ui together in tools node (#3763 )	2024-04-24 15:53:07 +08:00
ugyujiandGitHub	3b537cbdeb	fix: endpoint for 'Update a document from a file' (#3751 )	2024-04-24 15:25:53 +08:00
zxhlyhandGitHub	731464f5b8	fix: workflow sync (#3756 )	2024-04-24 15:19:19 +08:00
JoelandGitHub	1ad70f8721	feat: support prompt messages sorting (#3757 )	2024-04-24 15:09:01 +08:00
takatostandGitHub	2ea8c73cd8	fix: type num of variable converted to str (#3758 )	2024-04-24 15:07:56 +08:00
Jyong GitHub crazywoola JzoNg	f257f2c396	Knowledge optimization (#3755 ) Co-authored-by: crazywoola <427733928@qq.com> Co-authored-by: JzoNg <jzongcode@gmail.com>	2024-04-24 15:02:29 +08:00
JoelandGitHub	3cd8e6f5c6	fix: llm editor readonly cover error (#3752 )	2024-04-24 13:28:22 +08:00
KVOJJJinandGitHub	0715db7681	chore: add selector for use app store (#3746 )	2024-04-24 13:07:20 +08:00
zxhlyhandGitHub	a39de8a686	fix: workflow restore (#3750 )	2024-04-24 13:05:33 +08:00
Bowen LiangandGitHub	ccaf335466	fix: rollback gmpy2 to 2.1.5 (#3745 )	2024-04-24 12:53:23 +08:00
legaoandGitHub	40e36e9b52	fix: toggling AppDetailNav causes unnecessary component rerenders (#3718 )	2024-04-24 12:07:28 +08:00
zxhlyhandGitHub	9eebe9d54e	fix: workflow node variable (#3743 )	2024-04-24 11:41:12 +08:00
crazywoolaandGitHub	a23a191615	feat: add copy button to code (#3719 )	2024-04-24 09:34:51 +08:00
Leo QandGitHub	7d9c5586f9	Update "@formatjs/intl-localematcher" to version 0.5.4 in package.json (#3726 )	2024-04-24 09:06:23 +08:00
Ikko Eltociear AshimineandGitHub	f07c89bba4	Update README_JA.md (#3727 )	2024-04-24 09:04:27 +08:00
1102 GitHub heshunchang crazywoola	59cba930e5	bedrock llm Model file name change (#3714 ) Co-authored-by: heshunchang <shuncanghe@clouditera.com> Co-authored-by: crazywoola <427733928@qq.com>	2024-04-23 18:57:34 +08:00
zxhlyhandGitHub	39ae56e136	fix: workflow connection (#3713 )	2024-04-23 18:02:15 +08:00
JoelandGitHub	f92130338b	feat: prompt editor support auto height by content height and fix some bugs (#3712 )	2024-04-23 17:46:59 +08:00
Bowen LiangandGitHub	2867d29021	fix: milvus usage with create_collection (#3683 )	2024-04-23 17:37:40 +08:00
呆萌闷油瓶andGitHub	f76ac8bdee	enhance:speedup xinference audio transcription (#3636 )	2024-04-23 17:09:30 +08:00
zxhlyhandGitHub	83caffe000	fix: workflow restore (#3711 )	2024-04-23 17:02:23 +08:00
Luvian77andGitHub	96160837d2	fix: cannot change file uploader method (#3710 )	2024-04-23 17:02:12 +08:00
YeuolyandGitHub	3480f1c59e	refactor: tool parameter cache (#3703 )	2024-04-23 15:22:42 +08:00
zxhlyhandGitHub	65ac4f69af	fix: workflow shortcuts (#3701 )	2024-04-23 14:45:57 +08:00
YeuolyandGitHub	2c50fab3dd	fix: skip dataset icon (#3696 )	2024-04-23 12:41:41 +08:00
Carson KahnandGitHub	9525ccac4f	Localize links to localized READMEs (#3689 )	2024-04-23 09:30:32 +08:00
Richards Tu GitHub crazywoola	ff76c4bd5d	Add new tool: Judge0 CE (#3684 ) Co-authored-by: crazywoola <427733928@qq.com>	2024-04-23 09:07:21 +08:00
Luvian77 GitHub crazywoola	5dacf77627	fix: Added prevention of click event propagation for overlay layer (#3666 ) Co-authored-by: crazywoola <427733928@qq.com>	2024-04-22 19:53:20 +08:00
YeuolyandGitHub	2a213c6af7	fix: incorrect type parser (#3682 )	2024-04-22 19:32:41 +08:00
Bowen Liang GitHub crazywoola	b2535e7db6	chore: update description of code interpreter tool (#3679 ) Co-authored-by: crazywoola <100913391+crazywoola@users.noreply.github.com>	2024-04-22 19:19:16 +08:00
longzhihun GitHub Chenhe Gu	28236147ee	feat: add support for bedrock Mistral AI model (#3676 ) Co-authored-by: Chenhe Gu <guchenhe@gmail.com>	2024-04-22 17:24:02 +08:00
Chenhe GuandGitHub	4969783383	add groq llama3 (#3673 )	2024-04-22 15:21:09 +08:00