fix(api): scope segment batch deletes

2026-07-06 12:47:02 +08:00
2 changed files with 18 additions and 2 deletions
@@ -3872,7 +3872,14 @@ class SegmentService:
        session.add(document)

        # Delete database records
-        session.execute(delete(DocumentSegment).where(DocumentSegment.id.in_(segment_ids)))
+        session.execute(
+            delete(DocumentSegment).where(
+                DocumentSegment.id.in_(segment_db_ids),
+                DocumentSegment.dataset_id == dataset.id,
+                DocumentSegment.document_id == document.id,
+                DocumentSegment.tenant_id == current_user.current_tenant_id,
+            )
+        )
        session.commit()

    @classmethod
@@ -823,7 +823,9 @@ class TestSegmentServiceMutations:
            # scalars() for child_node_ids
            mock_db.session.scalars.return_value.all.return_value = ["child-1"]

-            SegmentService.delete_segments(["segment-1", "segment-2"], document, dataset, mock_db.session)
+            SegmentService.delete_segments(
+                ["segment-1", "segment-2", "foreign-segment"], document, dataset, mock_db.session
+            )

        assert document.word_count == 0
        mock_db.session.add.assert_called_once_with(document)
@@ -834,6 +836,13 @@ class TestSegmentServiceMutations:
            ["segment-1", "segment-2"],
            ["child-1"],
        )
+        delete_stmt = mock_db.session.execute.call_args_list[1].args[0]
+        delete_sql = str(delete_stmt.compile(compile_kwargs={"literal_binds": True}))
+        assert "document_segments.id IN ('segment-1', 'segment-2')" in delete_sql
+        assert "document_segments.dataset_id = 'dataset-1'" in delete_sql
+        assert "document_segments.document_id = 'doc-1'" in delete_sql
+        assert "document_segments.tenant_id = 'tenant-1'" in delete_sql
+        assert "foreign-segment" not in delete_sql
        mock_db.session.commit.assert_called_once()

    def test_update_segments_status_enables_only_segments_without_indexing_cache(self):