Skip to content

armgit / dify

Public Imported from GitHub
0

Build Agentic workflows, RAG pipelines, with rich AI model and tool support on one collaborative workspace. Deploy on cloud, VPC, or self-hosted, so teams move from prototype to production without rebuilding the stack.

test: migrate media and execution service sessions and ORM models to SQLite (#40552)

Asuka Minatoe27eb51f8a1b
diff --git a/api/tests/unit_tests/controllers/console/app/test_audio.py b/api/tests/unit_tests/controllers/console/app/test_audio.py
index 347627809ff..e9f01b00c48 100644
--- a/api/tests/unit_tests/controllers/console/app/test_audio.py
+++ b/api/tests/unit_tests/controllers/console/app/test_audio.py
@@ -2,12 +2,12 @@ from __future__ import annotations
import io
from inspect import unwrap
-from types import SimpleNamespace
from unittest.mock import patch
from uuid import UUID
import pytest
from flask import Flask
+from sqlalchemy.orm import Session
from werkzeug.datastructures import FileStorage
from werkzeug.exceptions import Forbidden, InternalServerError
@@ -34,7 +34,7 @@ from controllers.console.app.error import (
)
from core.errors.error import ModelCurrentlyNotSupportError, ProviderTokenNotInitError, QuotaExceededError
from graphon.model_runtime.errors.invoke import InvokeError
-from models import AppMode
+from models import Account, App, AppMode
from models.agent import AgentConfigDraftType
from models.agent_config_entities import AgentSoulConfig
from services.agent.composer_service import AgentComposerService
@@ -56,11 +56,30 @@ def _file_data():
return FileStorage(stream=io.BytesIO(b"audio"), filename="audio.wav", content_type="audio/wav")
+def _app(*, app_id: str = "a1", tenant_id: str = "tenant-1") -> App:
+ return App(
+ id=app_id,
+ tenant_id=tenant_id,
+ name="Audio app",
+ description="",
+ mode=AppMode.CHAT,
+ enable_site=True,
+ enable_api=True,
+ max_active_requests=0,
+ )
+
+
+def _account(account_id: str = "account-1") -> Account:
+ account = Account(name="Audio account", email=f"{account_id}@example.com")
+ account.id = account_id
+ return account
+
+
def test_console_audio_api_success(app: Flask, monkeypatch: pytest.MonkeyPatch) -> None:
monkeypatch.setattr(AudioService, "transcript_asr", lambda **_kwargs: {"text": "ok"})
api = ChatMessageAudioApi()
handler = unwrap(api.post)
- app_model = SimpleNamespace(id="a1")
+ app_model = _app()
with app.test_request_context("/console/api/apps/app/audio-to-text", method="POST", data={"file": _file_data()}):
response = handler(api, app_model=app_model)
@@ -72,9 +91,11 @@ def test_console_audio_api_accepts_published_agent_apps() -> None:
assert AppMode.AGENT in audio_module._CONSOLE_AUDIO_TRANSCRIPT_APP_MODES
-def test_agent_console_audio_api_uses_agent_draft(app: Flask, monkeypatch: pytest.MonkeyPatch) -> None:
+def test_agent_console_audio_api_uses_agent_draft(
+ app: Flask, monkeypatch: pytest.MonkeyPatch, unbound_session: Session
+) -> None:
agent_id = UUID("019ef3d2-b24c-7803-b428-18b5ee8fb853")
- app_model = SimpleNamespace(id="backing-app-1")
+ app_model = _app(app_id="backing-app-1")
agent_soul = AgentSoulConfig.model_validate({"app_features": {"speech_to_text": {"enabled": True}}})
calls: dict[str, object] = {}
@@ -100,8 +121,8 @@ def test_agent_console_audio_api_uses_agent_draft(app: Flask, monkeypatch: pytes
api = AgentChatMessageAudioApi()
handler = unwrap(api.post)
- session = SimpleNamespace()
- current_user = SimpleNamespace(id="account-1")
+ session = unbound_session
+ current_user = _account()
with app.test_request_context(
f"/console/api/agent/{agent_id}/audio-to-text",
method="POST",
@@ -140,13 +161,15 @@ def test_agent_console_audio_api_uses_agent_draft(app: Flask, monkeypatch: pytes
}
-def test_agent_console_audio_api_defaults_to_normal_draft(app: Flask, monkeypatch: pytest.MonkeyPatch) -> None:
+def test_agent_console_audio_api_defaults_to_normal_draft(
+ app: Flask, monkeypatch: pytest.MonkeyPatch, unbound_session: Session
+) -> None:
agent_id = UUID("019ef3d2-b24c-7803-b428-18b5ee8fb853")
captured: dict[str, object] = {}
monkeypatch.setattr(
audio_module,
"resolve_agent_runtime_app_model",
- lambda **_kwargs: SimpleNamespace(id="backing-app-1"),
+ lambda **_kwargs: _app(app_id="backing-app-1"),
)
def load_agent_soul_for_debug(**kwargs):
@@ -165,9 +188,9 @@ def test_agent_console_audio_api_defaults_to_normal_draft(app: Flask, monkeypatc
):
response = handler(
api,
- session=SimpleNamespace(),
+ session=unbound_session,
current_tenant_id="tenant-1",
- current_user=SimpleNamespace(id="account-1"),
+ current_user=_account(),
agent_id=agent_id,
)
@@ -175,9 +198,11 @@ def test_agent_console_audio_api_defaults_to_normal_draft(app: Flask, monkeypatc
assert captured["draft_type"] == AgentConfigDraftType.DRAFT
-def test_agent_console_audio_api_checks_rbac_with_backing_app_id(app: Flask, monkeypatch: pytest.MonkeyPatch) -> None:
+def test_agent_console_audio_api_checks_rbac_with_backing_app_id(
+ app: Flask, monkeypatch: pytest.MonkeyPatch, unbound_session: Session
+) -> None:
agent_id = UUID("019ef3d2-b24c-7803-b428-18b5ee8fb853")
- app_model = SimpleNamespace(id="backing-app-1")
+ app_model = _app(app_id="backing-app-1")
soul_loaded = False
monkeypatch.setattr(audio_module, "resolve_agent_runtime_app_model", lambda **_kwargs: app_model)
@@ -204,21 +229,23 @@ def test_agent_console_audio_api_checks_rbac_with_backing_app_id(app: Flask, mon
with pytest.raises(Forbidden):
handler(
api,
- session=SimpleNamespace(),
+ session=unbound_session,
current_tenant_id="tenant-1",
- current_user=SimpleNamespace(id="account-1"),
+ current_user=_account(),
agent_id=agent_id,
)
assert soul_loaded is False
-def test_agent_console_audio_api_preserves_missing_build_draft_404(app: Flask, monkeypatch: pytest.MonkeyPatch) -> None:
+def test_agent_console_audio_api_preserves_missing_build_draft_404(
+ app: Flask, monkeypatch: pytest.MonkeyPatch, unbound_session: Session
+) -> None:
agent_id = UUID("019ef3d2-b24c-7803-b428-18b5ee8fb853")
monkeypatch.setattr(
audio_module,
"resolve_agent_runtime_app_model",
- lambda **_kwargs: SimpleNamespace(id="backing-app-1"),
+ lambda **_kwargs: _app(app_id="backing-app-1"),
)
monkeypatch.setattr(
AgentComposerService,
@@ -236,9 +263,9 @@ def test_agent_console_audio_api_preserves_missing_build_draft_404(app: Flask, m
with pytest.raises(AgentVersionNotFoundError):
handler(
api,
- session=SimpleNamespace(),
+ session=unbound_session,
current_tenant_id="tenant-1",
- current_user=SimpleNamespace(id="account-1"),
+ current_user=_account(),
agent_id=agent_id,
)
@@ -262,7 +289,7 @@ def test_console_audio_api_error_mapping(app: Flask, monkeypatch: pytest.MonkeyP
monkeypatch.setattr(AudioService, "transcript_asr", lambda **_kwargs: (_ for _ in ()).throw(exc))
api = ChatMessageAudioApi()
handler = unwrap(api.post)
- app_model = SimpleNamespace(id="a1")
+ app_model = _app()
with app.test_request_context("/console/api/apps/app/audio-to-text", method="POST", data={"file": _file_data()}):
with pytest.raises(expected):
@@ -273,7 +300,7 @@ def test_console_audio_api_unhandled_error(app: Flask, monkeypatch: pytest.Monke
monkeypatch.setattr(AudioService, "transcript_asr", lambda **_kwargs: (_ for _ in ()).throw(RuntimeError("boom")))
api = ChatMessageAudioApi()
handler = unwrap(api.post)
- app_model = SimpleNamespace(id="a1")
+ app_model = _app()
with app.test_request_context("/console/api/apps/app/audio-to-text", method="POST", data={"file": _file_data()}):
with pytest.raises(InternalServerError):
@@ -285,7 +312,7 @@ def test_console_text_api_success(app: Flask, monkeypatch: pytest.MonkeyPatch) -
api = ChatMessageTextApi()
handler = unwrap(api.post)
- app_model = SimpleNamespace(id="a1")
+ app_model = _app()
with app.test_request_context(
"/console/api/apps/app/text-to-audio",
@@ -300,7 +327,7 @@ def test_console_text_api_success(app: Flask, monkeypatch: pytest.MonkeyPatch) -
def test_console_text_api_builds_message_ref(app: Flask, monkeypatch: pytest.MonkeyPatch) -> None:
api = ChatMessageTextApi()
handler = unwrap(api.post)
- app_model = SimpleNamespace(id="app-1", tenant_id="tenant-1")
+ app_model = _app(app_id="app-1")
calls = {}
def fake_transcript_tts(**kwargs):
@@ -315,7 +342,7 @@ def test_console_text_api_builds_message_ref(app: Flask, monkeypatch: pytest.Mon
method="POST",
json={"text": "hello", "message_id": "message-1"},
),
- patch("controllers.console.app.audio.current_user", SimpleNamespace(id="account-1")),
+ patch("controllers.console.app.audio.current_user", _account()),
):
response = handler(api, TextToSpeechPayload(text="hello", message_id="message-1"), app_model=app_model)
@@ -328,7 +355,7 @@ def test_console_text_api_error_mapping(app: Flask, monkeypatch: pytest.MonkeyPa
api = ChatMessageTextApi()
handler = unwrap(api.post)
- app_model = SimpleNamespace(id="a1")
+ app_model = _app()
with app.test_request_context(
"/console/api/apps/app/text-to-audio",
@@ -345,7 +372,7 @@ def test_console_text_modes_success(app: Flask, monkeypatch: pytest.MonkeyPatch)
api = TextModesApi()
handler = unwrap(api.get)
- app_model = SimpleNamespace(tenant_id="t1")
+ app_model = _app(tenant_id="t1")
with app.test_request_context("/console/api/apps/app/text-to-audio/voices?language=en", method="GET"):
response = handler(api, TextToSpeechVoiceQuery(language="en-US"), app_model=app_model)
@@ -362,7 +389,7 @@ def test_console_text_modes_language_error(app: Flask, monkeypatch: pytest.Monke
api = TextModesApi()
handler = unwrap(api.get)
- app_model = SimpleNamespace(tenant_id="t1")
+ app_model = _app(tenant_id="t1")
with app.test_request_context("/console/api/apps/app/text-to-audio/voices?language=en", method="GET"):
with pytest.raises(AppUnavailableError):
@@ -376,7 +403,7 @@ def test_audio_to_text_success(app: Flask, monkeypatch: pytest.MonkeyPatch) -> N
response_payload = {"text": "hello"}
monkeypatch.setattr(AudioService, "transcript_asr", lambda **_kwargs: response_payload)
- app_model = SimpleNamespace(id="app-1")
+ app_model = _app(app_id="app-1")
data = {"file": (io.BytesIO(b"x"), "sample.wav")}
with app.test_request_context(
@@ -400,7 +427,7 @@ def test_audio_to_text_maps_audio_too_large(app: Flask, monkeypatch: pytest.Monk
lambda **_kwargs: (_ for _ in ()).throw(AudioTooLargeServiceError("too large")),
)
- app_model = SimpleNamespace(id="app-1")
+ app_model = _app(app_id="app-1")
data = {"file": (io.BytesIO(b"x"), "sample.wav")}
with app.test_request_context(
@@ -419,7 +446,7 @@ def test_text_to_audio_success(app: Flask, monkeypatch: pytest.MonkeyPatch) -> N
monkeypatch.setattr(AudioService, "transcript_tts", lambda **_kwargs: {"audio": "ok"})
- app_model = SimpleNamespace(id="app-1")
+ app_model = _app(app_id="app-1")
with app.test_request_context(
"/console/api/apps/app-1/text-to-audio",
@@ -438,7 +465,7 @@ def test_text_to_audio_voices_success(app: Flask, monkeypatch: pytest.MonkeyPatc
expected_voices = [{"name": "Voice 1", "value": "voice-1"}]
monkeypatch.setattr(AudioService, "transcript_tts_voices", lambda **_kwargs: expected_voices)
- app_model = SimpleNamespace(tenant_id="tenant-1")
+ app_model = _app()
with app.test_request_context(
"/console/api/apps/app-1/text-to-audio/voices",
@@ -456,7 +483,7 @@ def test_audio_to_text_with_invalid_file(app: Flask, monkeypatch: pytest.MonkeyP
monkeypatch.setattr(AudioService, "transcript_asr", lambda **_kwargs: {"text": "test"})
- app_model = SimpleNamespace(id="app-1")
+ app_model = _app(app_id="app-1")
data = {"file": (io.BytesIO(b"invalid"), "sample.xyz")}
with app.test_request_context(
@@ -476,7 +503,7 @@ def test_text_to_audio_with_language_param(app: Flask, monkeypatch: pytest.Monke
monkeypatch.setattr(AudioService, "transcript_tts", lambda **_kwargs: {"audio": "test"})
- app_model = SimpleNamespace(id="app-1")
+ app_model = _app(app_id="app-1")
with app.test_request_context(
"/console/api/apps/app-1/text-to-audio",
@@ -497,7 +524,7 @@ def test_text_to_audio_voices_with_language_filter(app: Flask, monkeypatch: pyte
lambda **_kwargs: [{"name": "Voice 1", "value": "voice-1"}],
)
- app_model = SimpleNamespace(tenant_id="tenant-1")
+ app_model = _app()
with app.test_request_context(
"/console/api/apps/app-1/text-to-audio/voices?language=en-US",
diff --git a/api/tests/unit_tests/controllers/service_api/app/test_audio.py b/api/tests/unit_tests/controllers/service_api/app/test_audio.py
index 4fb56e6ac8b..091c129c874 100644
--- a/api/tests/unit_tests/controllers/service_api/app/test_audio.py
+++ b/api/tests/unit_tests/controllers/service_api/app/test_audio.py
@@ -10,7 +10,6 @@ Tests coverage for:
import io
import uuid
from inspect import unwrap
-from types import SimpleNamespace
from unittest.mock import Mock, patch
import pytest
@@ -34,6 +33,8 @@ from controllers.service_api.app.error import (
)
from core.errors.error import ModelCurrentlyNotSupportError, ProviderTokenNotInitError, QuotaExceededError
from graphon.model_runtime.errors.invoke import InvokeError
+from models.enums import EndUserType
+from models.model import App, AppMode, EndUser
from services.app_ref_service import AppRef, MessageRef
from services.audio_service import AudioService
from services.errors.app_model_config import AppModelConfigBrokenError
@@ -50,6 +51,31 @@ def _file_data():
return FileStorage(stream=io.BytesIO(b"audio"), filename="audio.wav", content_type="audio/wav")
+def _app(*, app_id: str = "a1", tenant_id: str = "tenant-1") -> App:
+ return App(
+ id=app_id,
+ tenant_id=tenant_id,
+ name="Audio app",
+ description="",
+ mode=AppMode.CHAT,
+ enable_site=True,
+ enable_api=True,
+ max_active_requests=0,
+ )
+
+
+def _end_user(*, end_user_id: str = "u1", external_user_id: str | None = None) -> EndUser:
+ return EndUser(
+ id=end_user_id,
+ tenant_id="tenant-1",
+ app_id="a1",
+ type=EndUserType.SERVICE_API,
+ external_user_id=external_user_id,
+ name="Audio user",
+ session_id=f"session-{end_user_id}",
+ )
+
+
# ---------------------------------------------------------------------------
# Pydantic Model Tests
# ---------------------------------------------------------------------------
@@ -197,8 +223,8 @@ class TestAudioApi:
monkeypatch.setattr(AudioService, "transcript_asr", lambda **_kwargs: {"text": "ok"})
api = AudioApi()
handler = unwrap(api.post)
- app_model = SimpleNamespace(id="a1")
- end_user = SimpleNamespace(id="u1")
+ app_model = _app()
+ end_user = _end_user()
with app.test_request_context("/audio-to-text", method="POST", data={"file": _file_data()}):
response = handler(api, app_model=app_model, end_user=end_user)
@@ -224,8 +250,8 @@ class TestAudioApi:
monkeypatch.setattr(AudioService, "transcript_asr", lambda **_kwargs: (_ for _ in ()).throw(exc))
api = AudioApi()
handler = unwrap(api.post)
- app_model = SimpleNamespace(id="a1")
- end_user = SimpleNamespace(id="u1")
+ app_model = _app()
+ end_user = _end_user()
with app.test_request_context("/audio-to-text", method="POST", data={"file": _file_data()}):
with pytest.raises(expected):
@@ -237,8 +263,8 @@ class TestAudioApi:
)
api = AudioApi()
handler = unwrap(api.post)
- app_model = SimpleNamespace(id="a1")
- end_user = SimpleNamespace(id="u1")
+ app_model = _app()
+ end_user = _end_user()
with app.test_request_context("/audio-to-text", method="POST", data={"file": _file_data()}):
with pytest.raises(InternalServerError):
@@ -251,8 +277,8 @@ class TestTextApi:
api = TextApi()
handler = unwrap(api.post)
- app_model = SimpleNamespace(id="a1")
- end_user = SimpleNamespace(id="end-user-1", external_user_id="ext")
+ app_model = _app()
+ end_user = _end_user(end_user_id="end-user-1", external_user_id="ext")
with app.test_request_context(
"/text-to-audio",
@@ -274,8 +300,8 @@ class TestTextApi:
api = TextApi()
handler = unwrap(api.post)
- app_model = SimpleNamespace(id="a1", tenant_id="tenant-1")
- end_user = SimpleNamespace(id="end-user-1", external_user_id="ext")
+ app_model = _app()
+ end_user = _end_user(end_user_id="end-user-1", external_user_id="ext")
with app.test_request_context(
"/text-to-audio",
@@ -294,8 +320,8 @@ class TestTextApi:
api = TextApi()
handler = unwrap(api.post)
- app_model = SimpleNamespace(id="a1")
- end_user = SimpleNamespace(id="end-user-1", external_user_id="ext")
+ app_model = _app()
+ end_user = _end_user(end_user_id="end-user-1", external_user_id="ext")
with app.test_request_context("/text-to-audio", method="POST", json={"text": "hello"}):
with pytest.raises(ProviderQuotaExceededError):
diff --git a/api/tests/unit_tests/services/test_api_token_service.py b/api/tests/unit_tests/services/test_api_token_service.py
index b3e5af87087..fd07797c0e0 100644
--- a/api/tests/unit_tests/services/test_api_token_service.py
+++ b/api/tests/unit_tests/services/test_api_token_service.py
@@ -14,10 +14,29 @@ from werkzeug.exceptions import Unauthorized
import services.api_token_service as api_token_service_module
from models.engine import db
+from models.enums import ApiTokenType
from models.model import ApiToken
from services.api_token_service import ApiTokenCache, CachedApiToken
+def _api_token(
+ *,
+ token_id: str = "id-123",
+ app_id: str = "app-123",
+ tenant_id: str = "tenant-123",
+ token: str = "token-123",
+) -> ApiToken:
+ """Create a mapped API token for cache and single-flight behavior tests."""
+ return ApiToken(
+ id=token_id,
+ app_id=app_id,
+ tenant_id=tenant_id,
+ type=ApiTokenType.APP,
+ token=token,
+ last_used_at=None,
+ )
+
+
@pytest.fixture
def api_token_db() -> Iterator[Session]:
"""Provide the production database extension with an isolated SQLite token table."""
@@ -124,7 +143,7 @@ class TestFetchTokenWithSingleFlight:
def test_should_query_db_when_lock_acquired_and_cache_missed(self):
auth_token = "token-123"
scope = "app"
- db_token = MagicMock()
+ db_token = _api_token()
lock = MagicMock()
lock.acquire.return_value = True
@@ -143,7 +162,7 @@ class TestFetchTokenWithSingleFlight:
def test_should_query_db_directly_when_lock_not_acquired(self):
auth_token = "token-123"
scope = "app"
- db_token = MagicMock()
+ db_token = _api_token()
lock = MagicMock()
lock.acquire.return_value = False
@@ -184,7 +203,7 @@ class TestFetchTokenWithSingleFlight:
def test_should_fallback_to_db_query_when_lock_raises_exception(self):
auth_token = "token-123"
scope = "app"
- db_token = MagicMock()
+ db_token = _api_token()
lock = MagicMock()
lock.acquire.side_effect = RuntimeError("redis lock error")
@@ -322,14 +341,7 @@ class TestApiTokenCacheCoreBranches:
@patch("services.api_token_service.redis_client")
def test_set_should_return_false_when_cache_write_raises_exception(self, mock_redis):
mock_redis.setex.side_effect = Exception("redis write failed")
- api_token = MagicMock()
- api_token.id = "id-123"
- api_token.app_id = "app-123"
- api_token.tenant_id = "tenant-123"
- api_token.type = "app"
- api_token.token = "token-123"
- api_token.last_used_at = None
- api_token.created_at = None
+ api_token = _api_token()
result = ApiTokenCache.set("token-123", "app", api_token)
assert result is False
diff --git a/api/tests/unit_tests/services/test_async_workflow_service.py b/api/tests/unit_tests/services/test_async_workflow_service.py
index 07285569b7d..71332dd5473 100644
--- a/api/tests/unit_tests/services/test_async_workflow_service.py
+++ b/api/tests/unit_tests/services/test_async_workflow_service.py
@@ -10,9 +10,11 @@ from sqlalchemy.engine import Engine
from sqlalchemy.orm import Session
import services.async_workflow_service as async_workflow_service_module
-from models.enums import AppTriggerType, CreatorUserRole, WorkflowRunTriggeredFrom, WorkflowTriggerStatus
-from models.model import App, AppMode
+from models.account import Account
+from models.enums import AppTriggerType, CreatorUserRole, EndUserType, WorkflowRunTriggeredFrom, WorkflowTriggerStatus
+from models.model import App, AppMode, EndUser
from models.trigger import WorkflowTriggerLog
+from models.workflow import Workflow, WorkflowType
from services.async_workflow_service import AsyncWorkflowService
from services.errors.app import QuotaExceededError, WorkflowNotFoundError
from services.workflow.entities import AsyncTriggerResponse, TriggerData
@@ -94,6 +96,41 @@ class AsyncWorkflowServiceTestDataFactory:
trigger_log.created_at = created_at
return trigger_log
+ @staticmethod
+ def create_workflow(
+ *, workflow_id: str = "workflow-123", app_id: str = "app-123", tenant_id: str = "tenant-123"
+ ) -> Workflow:
+ """Create a mapped workflow for service-return and trigger-log tests."""
+ return Workflow(
+ id=workflow_id,
+ tenant_id=tenant_id,
+ app_id=app_id,
+ type=WorkflowType.WORKFLOW,
+ version="1",
+ graph="{}",
+ _features="{}",
+ created_by="account-123",
+ )
+
+ @staticmethod
+ def create_account(account_id: str = "account-123") -> Account:
+ """Create a mapped account for role-discrimination tests."""
+ account = Account(name="Account", email=f"{account_id}@example.com")
+ account.id = account_id
+ return account
+
+ @staticmethod
+ def create_end_user(end_user_id: str = "end-user-123") -> EndUser:
+ """Create a mapped end user for role-discrimination and retry tests."""
+ return EndUser(
+ id=end_user_id,
+ tenant_id="tenant-123",
+ app_id="app-123",
+ type=EndUserType.BROWSER,
+ name="End User",
+ session_id=f"session-{end_user_id}",
+ )
+
@pytest.mark.usefixtures("sqlite_session")
@pytest.mark.parametrize("sqlite_session", [(App, WorkflowTriggerLog)], indirect=True)
@@ -163,8 +200,7 @@ class TestAsyncWorkflowService:
sqlite_session.add(AsyncWorkflowServiceTestDataFactory.create_app())
sqlite_session.commit()
trigger_data = AsyncWorkflowServiceTestDataFactory.create_trigger_data()
- workflow = MagicMock()
- workflow.id = "workflow-123"
+ workflow = AsyncWorkflowServiceTestDataFactory.create_workflow()
mocks = async_workflow_trigger_mocks
mocks["dispatcher"].get_queue_name.return_value = queue_name
@@ -179,17 +215,12 @@ class TestAsyncWorkflowService:
quota_charge_mock = MagicMock()
mocks["quota_service"].reserve.return_value = quota_charge_mock
- class DummyAccount:
- def __init__(self, user_id: str):
- self.id = user_id
+ user = AsyncWorkflowServiceTestDataFactory.create_account()
- with patch.object(async_workflow_service_module, "Account", DummyAccount):
- user = DummyAccount("account-123")
-
- # Act
- result = AsyncWorkflowService.trigger_workflow_async(
- session=sqlite_session, user=user, trigger_data=trigger_data
- )
+ # Act
+ result = AsyncWorkflowService.trigger_workflow_async(
+ session=sqlite_session, user=user, trigger_data=trigger_data
+ )
# Assert
assert isinstance(result, AsyncTriggerResponse)
@@ -233,8 +264,7 @@ class TestAsyncWorkflowService:
sqlite_session.add(AsyncWorkflowServiceTestDataFactory.create_app())
sqlite_session.commit()
trigger_data = AsyncWorkflowServiceTestDataFactory.create_trigger_data()
- workflow = MagicMock()
- workflow.id = "workflow-123"
+ workflow = AsyncWorkflowServiceTestDataFactory.create_workflow()
mocks = async_workflow_trigger_mocks
mocks["dispatcher"].get_queue_name.return_value = QueuePriority.SANDBOX
@@ -243,7 +273,7 @@ class TestAsyncWorkflowService:
task_result = MagicMock(id="task-123")
mocks["sandbox_task"].delay.return_value = task_result
- user = SimpleNamespace(id="end-user-123")
+ user = AsyncWorkflowServiceTestDataFactory.create_end_user()
# Act
response = AsyncWorkflowService.trigger_workflow_async(
@@ -269,7 +299,7 @@ class TestAsyncWorkflowService:
with pytest.raises(WorkflowNotFoundError, match="App not found: missing-app"):
AsyncWorkflowService.trigger_workflow_async(
session=sqlite_session,
- user=SimpleNamespace(id="user-123"),
+ user=AsyncWorkflowServiceTestDataFactory.create_end_user("user-123"),
trigger_data=trigger_data,
)
@@ -284,8 +314,7 @@ class TestAsyncWorkflowService:
sqlite_session.add(AsyncWorkflowServiceTestDataFactory.create_app())
sqlite_session.commit()
trigger_data = AsyncWorkflowServiceTestDataFactory.create_trigger_data()
- workflow = MagicMock()
- workflow.id = "workflow-123"
+ workflow = AsyncWorkflowServiceTestDataFactory.create_workflow()
mocks = async_workflow_trigger_mocks
mocks["dispatcher"].get_queue_name.return_value = QueuePriority.TEAM
@@ -301,7 +330,7 @@ class TestAsyncWorkflowService:
with pytest.raises(QuotaExceededError) as exc_info:
AsyncWorkflowService.trigger_workflow_async(
session=sqlite_session,
- user=SimpleNamespace(id="user-123"),
+ user=AsyncWorkflowServiceTestDataFactory.create_end_user("user-123"),
trigger_data=trigger_data,
)
@@ -328,7 +357,7 @@ class TestAsyncWorkflowService:
with pytest.raises(ValueError, match="Trigger log not found: missing-log"):
AsyncWorkflowService.reinvoke_trigger(
session=sqlite_session,
- user=SimpleNamespace(id="user-123"),
+ user=AsyncWorkflowServiceTestDataFactory.create_end_user("user-123"),
workflow_trigger_log_id="missing-log",
)
@@ -354,7 +383,7 @@ class TestAsyncWorkflowService:
return_value=expected_response,
) as mock_trigger_workflow_async,
):
- user = SimpleNamespace(id="user-123")
+ user = AsyncWorkflowServiceTestDataFactory.create_end_user("user-123")
# Act
response = AsyncWorkflowService.reinvoke_trigger(
@@ -497,8 +526,8 @@ class TestAsyncWorkflowServiceGetWorkflow:
"""Test _get_workflow returns published workflow by id when provided."""
# Arrange
workflow_service = MagicMock()
- app_model = MagicMock()
- workflow = MagicMock()
+ app_model = AsyncWorkflowServiceTestDataFactory.create_app()
+ workflow = AsyncWorkflowServiceTestDataFactory.create_workflow()
workflow_service.get_published_workflow_by_id.return_value = workflow
# Act
@@ -517,7 +546,7 @@ class TestAsyncWorkflowServiceGetWorkflow:
"""Test _get_workflow raises WorkflowNotFoundError for unknown workflow id."""
# Arrange
workflow_service = MagicMock()
- app_model = MagicMock()
+ app_model = AsyncWorkflowServiceTestDataFactory.create_app()
workflow_service.get_published_workflow_by_id.return_value = None
# Act / Assert
@@ -530,9 +559,8 @@ class TestAsyncWorkflowServiceGetWorkflow:
"""Test _get_workflow returns default published workflow when no id is provided."""
# Arrange
workflow_service = MagicMock()
- app_model = MagicMock()
- app_model.id = "app-123"
- workflow = MagicMock()
+ app_model = AsyncWorkflowServiceTestDataFactory.create_app()
+ workflow = AsyncWorkflowServiceTestDataFactory.create_workflow()
workflow_service.get_published_workflow.return_value = workflow
# Act
@@ -547,8 +575,7 @@ class TestAsyncWorkflowServiceGetWorkflow:
"""Test _get_workflow raises WorkflowNotFoundError when app has no published workflow."""
# Arrange
workflow_service = MagicMock()
- app_model = MagicMock()
- app_model.id = "app-123"
+ app_model = AsyncWorkflowServiceTestDataFactory.create_app()
workflow_service.get_published_workflow.return_value = None
# Act / Assert
diff --git a/api/tests/unit_tests/services/test_audio_service.py b/api/tests/unit_tests/services/test_audio_service.py
index e4d6046a3ea..33612604fbe 100644
--- a/api/tests/unit_tests/services/test_audio_service.py
+++ b/api/tests/unit_tests/services/test_audio_service.py
@@ -53,9 +53,11 @@ Tests available voice retrieval:
- text_to_speech: Enables TTS functionality
"""
+import json
from decimal import Decimal
from typing import Any
-from unittest.mock import MagicMock, Mock, create_autospec, patch
+from unittest.mock import MagicMock, Mock, patch
+from uuid import uuid4
import pytest
from sqlalchemy.orm import Session
@@ -64,7 +66,7 @@ from werkzeug.datastructures import FileStorage
from models.agent_config_entities import AgentSoulConfig
from models.enums import ConversationFromSource, MessageStatus
from models.model import App, AppMode, AppModelConfig, Message
-from models.workflow import Workflow
+from models.workflow import Workflow, WorkflowType
from services.app_ref_service import AppRef, MessageRef
from services.audio_service import AudioService
from services.errors.audio import (
@@ -113,15 +115,18 @@ class AudioServiceTestDataFactory:
audio-related operations.
"""
- @staticmethod
+ def __init__(self, session: Session) -> None:
+ self.session = session
+
def create_app_mock(
- app_id: str = "app-123",
+ self,
+ app_id: str = APP_ID,
mode: AppMode = AppMode.CHAT,
- tenant_id: str = "tenant-123",
+ tenant_id: str = TENANT_ID,
**kwargs,
- ) -> Mock:
+ ) -> App:
"""
- Create a mock App object.
+ Create and persist an App model.
Args:
app_id: Unique identifier for the app
@@ -130,46 +135,65 @@ class AudioServiceTestDataFactory:
**kwargs: Additional attributes to set on the mock
Returns:
- Mock App object with specified attributes
+ Persisted App model with specified attributes
"""
- app = create_autospec(App, instance=True)
- app.id = app_id
- app.mode = mode
- app.tenant_id = tenant_id
- app.workflow = kwargs.get("workflow")
- app.app_model_config = kwargs.get("app_model_config")
- app.workflow_with_session.return_value = app.workflow
- app.app_model_config_with_session.return_value = app.app_model_config
+ workflow = kwargs.pop("workflow", None)
+ app_model_config = kwargs.pop("app_model_config", None)
+ app = App(
+ id=app_id,
+ tenant_id=tenant_id,
+ name="Audio test app",
+ description="",
+ mode=mode,
+ icon_type=None,
+ icon=None,
+ icon_background=None,
+ enable_site=False,
+ enable_api=False,
+ workflow_id=workflow.id if workflow else None,
+ app_model_config_id=app_model_config.id if app_model_config else None,
+ )
for key, value in kwargs.items():
setattr(app, key, value)
+ self.session.add(app)
+ self.session.commit()
return app
- @staticmethod
- def create_workflow_mock(features_dict: dict[str, Any] | None = None, **kwargs) -> Mock:
+ def create_workflow_mock(self, features_dict: dict[str, Any] | None = None, **kwargs) -> Workflow:
"""
- Create a mock Workflow object.
+ Create and persist a Workflow model.
Args:
features_dict: Dictionary of workflow features
**kwargs: Additional attributes to set on the mock
Returns:
- Mock Workflow object with specified attributes
+ Persisted Workflow model with specified attributes
"""
- workflow = create_autospec(Workflow, instance=True)
- workflow.features_dict = features_dict or {}
+ workflow = Workflow(
+ id=kwargs.pop("id", str(uuid4())),
+ tenant_id=kwargs.pop("tenant_id", TENANT_ID),
+ app_id=kwargs.pop("app_id", APP_ID),
+ type=kwargs.pop("type", WorkflowType.CHAT),
+ version=kwargs.pop("version", Workflow.VERSION_DRAFT),
+ graph=kwargs.pop("graph", "{}"),
+ _features=json.dumps(features_dict or {}),
+ created_by=kwargs.pop("created_by", ACCOUNT_ID),
+ )
for key, value in kwargs.items():
setattr(workflow, key, value)
+ self.session.add(workflow)
+ self.session.commit()
return workflow
- @staticmethod
def create_app_model_config_mock(
+ self,
speech_to_text_dict: dict[str, Any] | None = None,
text_to_speech_dict: dict[str, Any] | None = None,
**kwargs,
- ) -> Mock:
+ ) -> AppModelConfig:
"""
- Create a mock AppModelConfig object.
+ Create and persist an AppModelConfig model.
Args:
speech_to_text_dict: Speech-to-text configuration
@@ -177,13 +201,17 @@ class AudioServiceTestDataFactory:
**kwargs: Additional attributes to set on the mock
Returns:
- Mock AppModelConfig object with specified attributes
+ Persisted AppModelConfig model with specified attributes
"""
- config = create_autospec(AppModelConfig, instance=True)
- config.speech_to_text_dict = speech_to_text_dict or {"enabled": False}
- config.text_to_speech_dict = text_to_speech_dict or {"enabled": False}
+ config = AppModelConfig(
+ app_id=kwargs.pop("app_id", APP_ID),
+ speech_to_text=json.dumps(speech_to_text_dict or {"enabled": False}),
+ text_to_speech=json.dumps(text_to_speech_dict or {"enabled": False}),
+ )
for key, value in kwargs.items():
setattr(config, key, value)
+ self.session.add(config)
+ self.session.commit()
return config
@staticmethod
@@ -216,9 +244,9 @@ class AudioServiceTestDataFactory:
@pytest.fixture
-def factory():
+def factory(sqlite_session: Session) -> AudioServiceTestDataFactory:
"""Provide the test data factory to all tests."""
- return AudioServiceTestDataFactory
+ return AudioServiceTestDataFactory(sqlite_session)
class TestAudioServiceASR:
@@ -385,7 +413,6 @@ class TestAudioServiceASR:
self, mock_model_manager_class, factory: AudioServiceTestDataFactory
):
app_model_config = factory.create_app_model_config_mock(speech_to_text_dict={"enabled": True})
- app_model_config.to_dict.return_value = {"speech_to_text": {"enabled": True}}
app = factory.create_app_mock(mode=AppMode.AGENT, app_model_config=app_model_config)
file = factory.create_file_storage_mock()
mock_model_instance = MagicMock()
@@ -403,7 +430,6 @@ class TestAudioServiceASR:
def test_transcript_agent_asr_soul_disabled_overrides_legacy_feature(self, factory: AudioServiceTestDataFactory):
app_model_config = factory.create_app_model_config_mock(speech_to_text_dict={"enabled": True})
- app_model_config.to_dict.return_value = {"speech_to_text": {"enabled": True}}
app = factory.create_app_mock(mode=AppMode.AGENT, app_model_config=app_model_config)
file = factory.create_file_storage_mock()
agent_soul = AgentSoulConfig.model_validate({"app_features": {"speech_to_text": {"enabled": False}}})
diff --git a/api/tests/unit_tests/services/test_message_service.py b/api/tests/unit_tests/services/test_message_service.py
index e0a8402b84e..e6e5c76a9ca 100644
--- a/api/tests/unit_tests/services/test_message_service.py
+++ b/api/tests/unit_tests/services/test_message_service.py
@@ -38,6 +38,7 @@ from models.model import (
Message,
MessageFeedback,
)
+from models.workflow import Workflow, WorkflowType
from repositories.sqlalchemy_execution_extra_content_repository import SQLAlchemyExecutionExtraContentRepository
from services.errors.message import (
FirstMessageNotExistsError,
@@ -111,6 +112,19 @@ class MessageServiceTestDataFactory:
account.id = user_id
return account
+ @staticmethod
+ def create_workflow(*, features: dict[str, object] | None = None) -> Workflow:
+ return Workflow(
+ id="workflow-123",
+ tenant_id="tenant-123",
+ app_id="app-123",
+ type=WorkflowType.CHAT,
+ version="1",
+ graph="{}",
+ _features=json.dumps(features or {}),
+ created_by="account-123",
+ )
+
@staticmethod
def create_conversation(
conversation_id: str = "conv-001",
@@ -712,8 +726,7 @@ class TestMessageServiceSuggestedQuestions:
monkeypatch: pytest.MonkeyPatch,
conversation: Conversation,
) -> tuple[MagicMock, MagicMock, MagicMock]:
- message = MagicMock()
- message.conversation_id = conversation.id
+ message = MessageServiceTestDataFactory.create_message(message_id="msg-123", conversation_id=conversation.id)
monkeypatch.setattr(service_module.MessageService, "get_message", MagicMock(return_value=message))
monkeypatch.setattr(
service_module.ConversationService, "get_conversation", MagicMock(return_value=conversation)
@@ -747,8 +760,7 @@ class TestMessageServiceSuggestedQuestions:
) -> None:
conversation = factory.create_conversation()
_, _, llm_generator = self._chat_boundaries(monkeypatch, conversation)
- workflow = MagicMock()
- workflow.features_dict = {"suggested_questions_after_answer": {"enabled": True}}
+ workflow = factory.create_workflow(features={"suggested_questions_after_answer": {"enabled": True}})
workflow_service = MagicMock()
workflow_service.return_value.get_published_workflow.return_value = workflow
monkeypatch.setattr(service_module, "WorkflowService", workflow_service)
@@ -1103,7 +1115,7 @@ class TestMessageServiceSuggestedQuestions:
) -> None:
conversation = factory.create_conversation()
self._chat_boundaries(monkeypatch, conversation)
- workflow = MagicMock()
+ workflow = factory.create_workflow()
workflow_service = MagicMock()
workflow_service.return_value.get_published_workflow.return_value = workflow
monkeypatch.setattr(service_module, "WorkflowService", workflow_service)
diff --git a/api/tests/unit_tests/services/test_vector_service.py b/api/tests/unit_tests/services/test_vector_service.py
index 75d371c8fa9..50b0d2c58c4 100644
--- a/api/tests/unit_tests/services/test_vector_service.py
+++ b/api/tests/unit_tests/services/test_vector_service.py
@@ -16,7 +16,7 @@ import services.vector_service as vector_service_module
from core.rag.index_processor.constant.index_type import IndexStructureType, IndexTechniqueType
from extensions.storage.storage_type import StorageType
from models import UploadFile
-from models.dataset import ChildChunk, DatasetProcessRule, SegmentAttachmentBinding
+from models.dataset import ChildChunk, Dataset, DatasetProcessRule, DocumentSegment, SegmentAttachmentBinding
from models.dataset import Document as DatasetDocument
from models.enums import CreatorUserRole, DataSourceType, DocumentCreatedFrom, ProcessRuleMode
from services.vector_service import VectorService
@@ -42,16 +42,18 @@ def _make_dataset(
is_multimodal: bool = False,
embedding_model_provider: str | None = "openai",
embedding_model: str = "text-embedding",
-) -> MagicMock:
- dataset = MagicMock(name="dataset")
- dataset.id = dataset_id
- dataset.tenant_id = tenant_id
- dataset.doc_form = doc_form
- dataset.indexing_technique = indexing_technique
- dataset.is_multimodal = is_multimodal
- dataset.embedding_model_provider = embedding_model_provider
- dataset.embedding_model = embedding_model
- dataset.get_doc_form.return_value = doc_form
+) -> Dataset:
+ dataset = Dataset(
+ id=dataset_id,
+ tenant_id=tenant_id,
+ name="Dataset",
+ created_by="account-1",
+ indexing_technique=indexing_technique,
+ chunk_structure=doc_form,
+ is_multimodal=is_multimodal,
+ embedding_model_provider=embedding_model_provider,
+ embedding_model=embedding_model,
+ )
return dataset
@@ -64,24 +66,49 @@ def _make_segment(
content: str = "hello",
index_node_id: str = "node-1",
index_node_hash: str = "hash-1",
+ session: Session | None = None,
attachments: list[dict[str, str]] | None = None,
-) -> MagicMock:
- segment = MagicMock(name="segment")
+) -> DocumentSegment:
+ segment = DocumentSegment(
+ tenant_id=tenant_id,
+ dataset_id=dataset_id,
+ document_id=document_id,
+ position=1,
+ content=content,
+ word_count=len(content),
+ tokens=len(content),
+ created_by="account-1",
+ index_node_id=index_node_id,
+ index_node_hash=index_node_hash,
+ )
segment.id = segment_id
- segment.tenant_id = tenant_id
- segment.dataset_id = dataset_id
- segment.document_id = document_id
- segment.content = content
- segment.index_node_id = index_node_id
- segment.index_node_hash = index_node_hash
- segment.attachments = attachments or []
- segment.get_attachments.return_value = attachments or []
+ if attachments:
+ assert session is not None
+ for attachment in attachments:
+ upload_file = _upload_file(
+ file_id=attachment["id"],
+ name=attachment.get("name", f"{attachment['id']}.png"),
+ tenant_id=tenant_id,
+ )
+ session.add_all(
+ [
+ upload_file,
+ SegmentAttachmentBinding(
+ tenant_id=tenant_id,
+ dataset_id=dataset_id,
+ document_id=document_id,
+ segment_id=segment_id,
+ attachment_id=upload_file.id,
+ ),
+ ]
+ )
+ session.flush()
return segment
-def _upload_file(*, file_id: str = "file-1", name: str = "img.png") -> UploadFile:
+def _upload_file(*, file_id: str = "file-1", name: str = "img.png", tenant_id: str = "tenant-1") -> UploadFile:
upload_file = UploadFile(
- tenant_id="tenant-1",
+ tenant_id=tenant_id,
storage_type=StorageType.LOCAL,
key=f"uploads/{file_id}",
name=name,
@@ -97,6 +124,30 @@ def _upload_file(*, file_id: str = "file-1", name: str = "img.png") -> UploadFil
return upload_file
+def _make_child_chunk(
+ *,
+ index_node_id: str,
+ content: str = "child",
+ index_node_hash: str = "hash",
+ tenant_id: str = "tenant-1",
+ dataset_id: str = "dataset-1",
+ document_id: str = "doc-1",
+ segment_id: str = "seg-1",
+) -> ChildChunk:
+ return ChildChunk(
+ tenant_id=tenant_id,
+ dataset_id=dataset_id,
+ document_id=document_id,
+ segment_id=segment_id,
+ position=1,
+ content=content,
+ word_count=len(content),
+ created_by="account-1",
+ index_node_id=index_node_id,
+ index_node_hash=index_node_hash,
+ )
+
+
def test_create_segments_vector_regular_indexing_loads_documents_and_keywords(
monkeypatch: pytest.MonkeyPatch, sqlite_session: Session
) -> None:
@@ -126,10 +177,11 @@ def test_create_segments_vector_regular_indexing_loads_multimodal_documents(
) -> None:
dataset = _make_dataset(is_multimodal=True)
segment = _make_segment(
+ session=sqlite_session,
attachments=[
{"id": "img-1", "name": "a.png"},
{"id": "img-2", "name": "b.png"},
- ]
+ ],
)
index_processor = MagicMock(name="index_processor")
@@ -152,7 +204,7 @@ def test_create_segments_vector_regular_indexing_loads_multimodal_documents(
assert second_args[1] == []
assert len(second_args[2]) == 2
assert second_kwargs["with_keywords"] is False
- segment.get_attachments.assert_called_once_with(session=sqlite_session)
+ assert {document.page_content for document in second_args[2]} == {"a.png", "b.png"}
def test_create_segments_vector_with_no_segments_does_not_load(
@@ -171,7 +223,7 @@ def test_create_segments_vector_with_no_segments_does_not_load(
def _persist_parent_child_rows(
session: Session,
*,
- segment: MagicMock,
+ segment: DocumentSegment,
include_document: bool = True,
include_rule: bool = True,
) -> tuple[DatasetDocument | None, DatasetProcessRule | None]:
@@ -415,13 +467,24 @@ def test_generate_child_chunks_regenerate_cleans_then_saves_children(
dataset = _make_dataset(doc_form=IndexStructureType.PARAGRAPH_INDEX, tenant_id="tenant-1", dataset_id="dataset-1")
segment = _make_segment(segment_id="seg-1")
- dataset_document = MagicMock()
- dataset_document.id = segment.document_id
- dataset_document.doc_language = "en"
- dataset_document.created_by = "user-1"
-
- processing_rule = MagicMock()
- processing_rule.to_dict.return_value = {"rules": {}}
+ dataset_document = DatasetDocument(
+ id=segment.document_id,
+ tenant_id=segment.tenant_id,
+ dataset_id=segment.dataset_id,
+ position=1,
+ data_source_type=DataSourceType.UPLOAD_FILE,
+ batch="batch-1",
+ name="Document",
+ created_from=DocumentCreatedFrom.API,
+ created_by="user-1",
+ doc_language="en",
+ )
+ processing_rule = DatasetProcessRule(
+ dataset_id=segment.dataset_id,
+ mode=ProcessRuleMode.HIERARCHICAL,
+ rules="{}",
+ created_by="user-1",
+ )
child1 = _ChildDocStub(page_content="c1", metadata={"doc_id": "c1-id", "doc_hash": "c1-h"})
child2 = _ChildDocStub(page_content="c2", metadata={"doc_id": "c2-id", "doc_hash": "c2-h"})
@@ -456,12 +519,24 @@ def test_generate_child_chunks_flushes_even_when_no_children(
) -> None:
dataset = _make_dataset(doc_form=IndexStructureType.PARAGRAPH_INDEX)
segment = _make_segment()
- dataset_document = MagicMock()
- dataset_document.doc_language = "en"
- dataset_document.created_by = "user-1"
-
- processing_rule = MagicMock()
- processing_rule.to_dict.return_value = {"rules": {}}
+ dataset_document = DatasetDocument(
+ id=segment.document_id,
+ tenant_id=segment.tenant_id,
+ dataset_id=segment.dataset_id,
+ position=1,
+ data_source_type=DataSourceType.UPLOAD_FILE,
+ batch="batch-1",
+ name="Document",
+ created_from=DocumentCreatedFrom.API,
+ created_by="user-1",
+ doc_language="en",
+ )
+ processing_rule = DatasetProcessRule(
+ dataset_id=segment.dataset_id,
+ mode=ProcessRuleMode.HIERARCHICAL,
+ rules="{}",
+ created_by="user-1",
+ )
index_processor = MagicMock()
index_processor.transform.return_value = [_ParentDocStub(children=[])]
@@ -487,12 +562,7 @@ def test_create_child_chunk_vector_high_quality_adds_texts(
monkeypatch: pytest.MonkeyPatch, sqlite_session: Session
) -> None:
dataset = _make_dataset(indexing_technique=IndexTechniqueType.HIGH_QUALITY)
- child_chunk = MagicMock()
- child_chunk.content = "child"
- child_chunk.index_node_id = "id"
- child_chunk.index_node_hash = "h"
- child_chunk.document_id = "doc-1"
- child_chunk.dataset_id = "dataset-1"
+ child_chunk = _make_child_chunk(index_node_id="id", index_node_hash="h")
vector_instance = MagicMock()
vector_cls = MagicMock(return_value=vector_instance)
@@ -508,12 +578,7 @@ def test_create_child_chunk_vector_economy_noop(monkeypatch: pytest.MonkeyPatch,
vector_cls = MagicMock()
monkeypatch.setattr(vector_service_module, "Vector", vector_cls)
- child_chunk = MagicMock()
- child_chunk.content = "child"
- child_chunk.index_node_id = "id"
- child_chunk.index_node_hash = "h"
- child_chunk.document_id = "doc-1"
- child_chunk.dataset_id = "dataset-1"
+ child_chunk = _make_child_chunk(index_node_id="id", index_node_hash="h")
VectorService.create_child_chunk_vector(child_chunk, dataset, session=sqlite_session)
vector_cls.assert_not_called()
@@ -524,22 +589,13 @@ def test_update_child_chunk_vector_high_quality_updates_vector(
) -> None:
dataset = _make_dataset(indexing_technique=IndexTechniqueType.HIGH_QUALITY)
- new_chunk = MagicMock()
- new_chunk.content = "n"
- new_chunk.index_node_id = "nid"
- new_chunk.index_node_hash = "nh"
- new_chunk.document_id = "d"
- new_chunk.dataset_id = "ds"
-
- upd_chunk = MagicMock()
- upd_chunk.content = "u"
- upd_chunk.index_node_id = "uid"
- upd_chunk.index_node_hash = "uh"
- upd_chunk.document_id = "d"
- upd_chunk.dataset_id = "ds"
-
- del_chunk = MagicMock()
- del_chunk.index_node_id = "did"
+ new_chunk = _make_child_chunk(
+ content="n", index_node_id="nid", index_node_hash="nh", document_id="d", dataset_id="ds"
+ )
+ upd_chunk = _make_child_chunk(
+ content="u", index_node_id="uid", index_node_hash="uh", document_id="d", dataset_id="ds"
+ )
+ del_chunk = _make_child_chunk(index_node_id="did")
vector_instance = MagicMock()
vector_cls = MagicMock(return_value=vector_instance)
@@ -564,8 +620,7 @@ def test_update_child_chunk_vector_economy_noop(monkeypatch: pytest.MonkeyPatch,
def test_delete_child_chunk_vector_deletes_by_id(monkeypatch: pytest.MonkeyPatch, sqlite_session: Session) -> None:
dataset = _make_dataset()
- child_chunk = MagicMock()
- child_chunk.index_node_id = "cid"
+ child_chunk = _make_child_chunk(index_node_id="cid")
vector_instance = MagicMock()
vector_cls = MagicMock(return_value=vector_instance)
@@ -585,7 +640,7 @@ def test_update_multimodel_vector_returns_when_not_high_quality(
monkeypatch: pytest.MonkeyPatch, sqlite_session: Session
) -> None:
dataset = _make_dataset(indexing_technique=IndexTechniqueType.ECONOMY, is_multimodal=True)
- segment = _make_segment(tenant_id="t", attachments=[{"id": "a"}])
+ segment = _make_segment(tenant_id="t")
vector_cls = MagicMock()
monkeypatch.setattr(vector_service_module, "Vector", vector_cls)
@@ -601,7 +656,7 @@ def test_update_multimodel_vector_returns_when_no_actual_change(
monkeypatch: pytest.MonkeyPatch, sqlite_session: Session
) -> None:
dataset = _make_dataset(indexing_technique=IndexTechniqueType.HIGH_QUALITY, is_multimodal=True)
- segment = _make_segment(tenant_id="t", attachments=[{"id": "a"}, {"id": "b"}])
+ segment = _make_segment(tenant_id="t", session=sqlite_session, attachments=[{"id": "a"}, {"id": "b"}])
vector_cls = MagicMock()
monkeypatch.setattr(vector_service_module, "Vector", vector_cls)
@@ -610,7 +665,7 @@ def test_update_multimodel_vector_returns_when_no_actual_change(
session=sqlite_session, segment=segment, attachment_ids=["b", "a"], dataset=dataset
)
vector_cls.assert_not_called()
- assert not sqlite_session.in_transaction()
+ assert sqlite_session.in_transaction()
def test_update_multimodel_vector_deletes_bindings_and_commits_on_empty_new_ids(
@@ -618,23 +673,14 @@ def test_update_multimodel_vector_deletes_bindings_and_commits_on_empty_new_ids(
sqlite_session: Session,
) -> None:
dataset = _make_dataset(indexing_technique=IndexTechniqueType.HIGH_QUALITY, is_multimodal=True)
- segment = _make_segment(tenant_id="tenant-1", attachments=[{"id": "old-1"}, {"id": "old-2"}])
+ segment = _make_segment(
+ tenant_id="tenant-1",
+ session=sqlite_session,
+ attachments=[{"id": "old-1"}, {"id": "old-2"}],
+ )
vector_instance = MagicMock(name="vector_instance")
vector_cls = MagicMock(return_value=vector_instance)
- sqlite_session.add_all(
- [
- SegmentAttachmentBinding(
- tenant_id="tenant-1",
- dataset_id="dataset-1",
- document_id="doc-1",
- segment_id="seg-1",
- attachment_id=attachment_id,
- )
- for attachment_id in ("old-1", "old-2")
- ]
- )
- sqlite_session.flush()
monkeypatch.setattr(vector_service_module, "Vector", vector_cls)
@@ -650,7 +696,7 @@ def test_update_multimodel_vector_flushes_when_no_upload_files_found(
monkeypatch: pytest.MonkeyPatch, sqlite_session: Session
) -> None:
dataset = _make_dataset(indexing_technique=IndexTechniqueType.HIGH_QUALITY, is_multimodal=True)
- segment = _make_segment(tenant_id="tenant-1", attachments=[{"id": "old-1"}])
+ segment = _make_segment(tenant_id="tenant-1", session=sqlite_session, attachments=[{"id": "old-1"}])
vector_instance = MagicMock()
monkeypatch.setattr(vector_service_module, "Vector", MagicMock(return_value=vector_instance))
@@ -668,7 +714,12 @@ def test_update_multimodel_vector_adds_bindings_and_vectors_and_skips_missing_up
sqlite_session: Session,
) -> None:
dataset = _make_dataset(indexing_technique=IndexTechniqueType.HIGH_QUALITY, is_multimodal=True)
- segment = _make_segment(segment_id="seg-1", tenant_id="tenant-1", attachments=[{"id": "old-1"}])
+ segment = _make_segment(
+ segment_id="seg-1",
+ tenant_id="tenant-1",
+ session=sqlite_session,
+ attachments=[{"id": "old-1"}],
+ )
vector_instance = MagicMock()
monkeypatch.setattr(vector_service_module, "Vector", MagicMock(return_value=vector_instance))
@@ -696,7 +747,7 @@ def test_update_multimodel_vector_updates_bindings_without_multimodal_vector_ops
sqlite_session: Session,
) -> None:
dataset = _make_dataset(indexing_technique=IndexTechniqueType.HIGH_QUALITY, is_multimodal=False)
- segment = _make_segment(tenant_id="tenant-1", attachments=[{"id": "old-1"}])
+ segment = _make_segment(tenant_id="tenant-1", session=sqlite_session, attachments=[{"id": "old-1"}])
vector_instance = MagicMock()
monkeypatch.setattr(vector_service_module, "Vector", MagicMock(return_value=vector_instance))
@@ -719,7 +770,12 @@ def test_update_multimodel_vector_rolls_back_and_reraises_on_error(
sqlite_session: Session,
) -> None:
dataset = _make_dataset(indexing_technique=IndexTechniqueType.HIGH_QUALITY, is_multimodal=True)
- segment = _make_segment(segment_id="seg-1", tenant_id="tenant-1", attachments=[{"id": "old-1"}])
+ segment = _make_segment(
+ segment_id="seg-1",
+ tenant_id="tenant-1",
+ session=sqlite_session,
+ attachments=[{"id": "old-1"}],
+ )
vector_instance = MagicMock()
monkeypatch.setattr(vector_service_module, "Vector", MagicMock(return_value=vector_instance))
@@ -730,10 +786,11 @@ def test_update_multimodel_vector_rolls_back_and_reraises_on_error(
monkeypatch.setattr(sqlite_session, "flush", MagicMock(side_effect=RuntimeError("boom")))
with caplog.at_level(logging.ERROR, logger="services.vector_service"):
- with pytest.raises(RuntimeError, match="boom"):
- VectorService.update_multimodel_vector(
- session=sqlite_session, segment=segment, attachment_ids=["file-1"], dataset=dataset
- )
+ with sqlite_session.no_autoflush:
+ with pytest.raises(RuntimeError, match="boom"):
+ VectorService.update_multimodel_vector(
+ session=sqlite_session, segment=segment, attachment_ids=["file-1"], dataset=dataset
+ )
assert any(r.levelno >= logging.ERROR for r in caplog.records)
assert rollback_events == ["rollback"]