deerflow-code/offline-backend-20260512/backend/tests/test_llmwiki_wiki_sync.py
2026-09-07 18:24:55 +08:00

244 lines
8.9 KiB
Python

from __future__ import annotations
import asyncio
import json
from types import SimpleNamespace
import httpx
import pytest
from app.gateway import llmwiki_index_scheduler
from deerflow.config.llmwiki_config import LocalWikiEmbeddingConfig, LocalWikiIndexConfig
from deerflow.integrations.weknora.local_index.embedding import StrictWikiEmbeddingClient
from deerflow.integrations.weknora.local_index.sync import WikiSyncService
from deerflow.integrations.weknora.local_index.vector_cache import WikiVectorCache
from deerflow.persistence.llmwiki_index.memory import MemoryLlmWikiIndexStore
class FakeEmbedding:
dimensions = 2
fingerprint = "fp"
def __init__(self) -> None:
self.calls = 0
async def embed_texts(self, texts: list[str]) -> list[list[float]]:
self.calls += 1
return [[1.0, 0.0] for _ in texts]
class FailingEmbedding(FakeEmbedding):
async def embed_texts(self, texts: list[str]) -> list[list[float]]:
raise RuntimeError("embedding unavailable")
class NewFingerprintEmbedding(FakeEmbedding):
fingerprint = "fp-new"
class FakeWeKnora:
def __init__(self) -> None:
self.pages = {
"one": {
"id": "p1",
"slug": "one",
"title": "One",
"summary": "A long enough summary for indexing and retrieval.",
"content": "# Intro\nA long enough Wiki paragraph for indexing and semantic retrieval.",
"status": "published",
"version": 1,
},
}
async def get_knowledge_base(self, knowledge_base_id: str):
return {"id": knowledge_base_id, "indexing_strategy": {"wiki_enabled": True}}
async def list_wiki_pages(self, knowledge_base_id: str, *, page: int, page_size: int):
return {
"pages": [{"slug": item["slug"], "title": item["title"]} for item in self.pages.values()],
"total": len(self.pages),
"total_pages": 1,
}
async def get_wiki_page(self, knowledge_base_id: str, slug: str):
return dict(self.pages[slug])
@pytest.mark.asyncio
async def test_configured_dimensions_are_validated_but_not_sent_to_bge() -> None:
requests: list[dict] = []
def handler(request: httpx.Request) -> httpx.Response:
payload = json.loads(request.content)
requests.append(payload)
if "dimensions" in payload:
return httpx.Response(400, json={"detail": "dimensions is not supported"})
return httpx.Response(
200,
json={"data": [{"index": 0, "embedding": [1.0, 0.0, 0.0]}]},
)
config = LocalWikiEmbeddingConfig(
model="bge-embedding-m3",
base_url="http://embedding.test/v1",
dimensions=3,
max_retries=0,
)
async with httpx.AsyncClient(transport=httpx.MockTransport(handler)) as http_client:
client = StrictWikiEmbeddingClient(config, http_client=http_client)
vector = await client.embed_query("Wiki page")
assert vector == [1.0, 0.0, 0.0]
assert requests == [{"model": "bge-embedding-m3", "input": ["Wiki page"]}]
@pytest.mark.asyncio
async def test_embedding_error_includes_upstream_reason(caplog) -> None:
def handler(request: httpx.Request) -> httpx.Response:
return httpx.Response(400, json={"detail": "unsupported request parameter: dimensions"})
config = LocalWikiEmbeddingConfig(
model="bge-embedding-m3",
base_url="http://embedding.test/v1",
max_retries=0,
)
async with httpx.AsyncClient(transport=httpx.MockTransport(handler)) as http_client:
client = StrictWikiEmbeddingClient(config, http_client=http_client)
with pytest.raises(RuntimeError, match="unsupported request parameter: dimensions"):
await client.embed_query("Wiki page")
assert "unsupported request parameter: dimensions" in caplog.text
@pytest.mark.asyncio
async def test_omitted_dimensions_are_discovered_from_first_response() -> None:
def handler(request: httpx.Request) -> httpx.Response:
return httpx.Response(
200,
json={"data": [{"index": 0, "embedding": [0.0, 1.0]}]},
)
config = LocalWikiEmbeddingConfig(
model="bge-embedding-m3",
base_url="http://embedding.test/v1",
dimensions=None,
max_retries=0,
)
async with httpx.AsyncClient(transport=httpx.MockTransport(handler)) as http_client:
client = StrictWikiEmbeddingClient(config, http_client=http_client)
await client.embed_query("dimension probe")
assert client.dimensions == 2
@pytest.mark.asyncio
async def test_sync_is_incremental_and_full_scan_deletes_missing_pages() -> None:
config = LocalWikiIndexConfig(chunk_min_chars=10, min_similarity=0)
store = MemoryLlmWikiIndexStore()
embedding = FakeEmbedding()
client = FakeWeKnora()
service = WikiSyncService(config, store, embedding, WikiVectorCache(store), client)
mapping = {"id": "kb", "weknora_id": "remote", "name": "Wiki"}
first = await service.sync_mapping(mapping)
assert first["changed_pages"] == 1
assert embedding.calls == 1
statuses = await store.list_page_index_status("kb")
assert statuses[0]["title"] == "One"
assert statuses[0]["vector_count"] > 0
assert store.states["kb"]["remote_page_count"] == 1
second = await service.sync_mapping(mapping)
assert second["changed_pages"] == 0
assert embedding.calls == 1
client.pages.clear()
deleted = await service.sync_mapping(mapping)
assert deleted["deleted_pages"] == 1
assert next(iter(store.pages.values()))["is_remote_deleted"] is True
@pytest.mark.asyncio
async def test_page_failure_keeps_previous_generation_and_marks_partial() -> None:
config = LocalWikiIndexConfig(chunk_min_chars=10, min_similarity=0)
store = MemoryLlmWikiIndexStore()
client = FakeWeKnora()
mapping = {"id": "kb", "weknora_id": "remote", "name": "Wiki"}
await WikiSyncService(config, store, FakeEmbedding(), WikiVectorCache(store), client).sync_mapping(mapping)
page = next(iter(store.pages.values()))
old_hash = page["content_hash"]
old_vectors = list(store.vectors[page["id"]])
client.pages["one"]["content"] += " Changed."
result = await WikiSyncService(config, store, FailingEmbedding(), WikiVectorCache(store), client).sync_mapping(mapping)
assert result["status"] == "partial"
assert store.states["kb"]["state"] == "partial"
assert store.states["kb"]["failed_page_count"] == 1
assert store.pages[page["id"]]["content_hash"] == old_hash
assert store.pages[page["id"]]["index_status"] == "ready"
assert store.vectors[page["id"]] == old_vectors
@pytest.mark.asyncio
async def test_fingerprint_change_automatically_rebuilds_entire_mapping() -> None:
config = LocalWikiIndexConfig(chunk_min_chars=10, min_similarity=0)
store = MemoryLlmWikiIndexStore()
client = FakeWeKnora()
mapping = {"id": "kb", "weknora_id": "remote", "name": "Wiki"}
await WikiSyncService(config, store, FakeEmbedding(), WikiVectorCache(store), client).sync_mapping(mapping)
embedding = NewFingerprintEmbedding()
result = await WikiSyncService(config, store, embedding, WikiVectorCache(store), client).sync_mapping(mapping)
assert result["status"] == "completed"
assert result["changed_pages"] == 1
assert embedding.calls == 1
assert store.states["kb"]["state"] == "idle"
assert store.states["kb"]["embedding_fingerprint"] == "fp-new"
assert {vector["embedding_fingerprint"] for vectors in store.vectors.values() for vector in vectors} == {"fp-new"}
@pytest.mark.asyncio
async def test_auto_sync_scheduler_scans_immediately_then_waits(monkeypatch) -> None:
calls: list[str] = []
async def fake_sleep(seconds: float) -> None:
calls.append(f"sleep:{seconds}")
raise asyncio.CancelledError
async def fake_sync_once(app) -> None:
calls.append("sync")
if calls.count("sync") > 1:
raise asyncio.CancelledError
monkeypatch.setattr(llmwiki_index_scheduler.asyncio, "sleep", fake_sleep)
monkeypatch.setattr(llmwiki_index_scheduler, "_sync_once", fake_sync_once)
app = SimpleNamespace(
state=SimpleNamespace(
config=SimpleNamespace(
llmwiki=SimpleNamespace(
local_wiki_index=SimpleNamespace(sync_interval_seconds=120),
)
)
)
)
with pytest.raises(asyncio.CancelledError):
await llmwiki_index_scheduler.local_wiki_index_scheduler_loop(app)
assert calls == ["sync", "sleep:120"]
def test_local_wiki_index_defaults_to_automatic_sync() -> None:
config = LocalWikiIndexConfig()
assert config.auto_sync is True
assert config.sync_interval_seconds == 30
def test_local_wiki_index_accepts_legacy_sync_flag() -> None:
config = LocalWikiIndexConfig.model_validate({"sync_enabled": True})
assert config.auto_sync is True
assert config.sync_enabled is True