296 lines
8.7 KiB
Python
296 lines
8.7 KiB
Python
from __future__ import annotations
|
|
|
|
import asyncio
|
|
from datetime import UTC, datetime, timedelta
|
|
from types import SimpleNamespace
|
|
from typing import Any
|
|
|
|
import pytest
|
|
|
|
import app.gateway.llmwiki_index_scheduler as scheduler
|
|
from app.gateway.llmwiki_index_scheduler import (
|
|
_mapping_is_due,
|
|
schedule_local_wiki_mutation_sync,
|
|
)
|
|
|
|
|
|
class _RemoteClient:
|
|
def __init__(self, processing: list[Any]) -> None:
|
|
self.processing = list(processing)
|
|
self.status_reads = 0
|
|
|
|
async def get_knowledge_base(self, knowledge_base_id: str) -> dict[str, Any]:
|
|
assert knowledge_base_id == "remote-kb"
|
|
self.status_reads += 1
|
|
current = self.processing.pop(0) if self.processing else False
|
|
return {"id": knowledge_base_id, "is_processing": current}
|
|
|
|
|
|
class _SyncService:
|
|
def __init__(
|
|
self,
|
|
processing: list[bool],
|
|
results: list[dict[str, Any]] | None = None,
|
|
) -> None:
|
|
self.client = _RemoteClient(processing)
|
|
self.results = list(results or [])
|
|
self.sync_calls = 0
|
|
self.seen_mappings: list[dict[str, Any]] = []
|
|
|
|
async def sync_mapping(
|
|
self,
|
|
mapping: dict[str, Any],
|
|
*,
|
|
force: bool,
|
|
) -> dict[str, Any]:
|
|
assert mapping["id"] == "mapping-kb"
|
|
assert force is False
|
|
self.sync_calls += 1
|
|
self.seen_mappings.append(dict(mapping))
|
|
if self.results:
|
|
return self.results.pop(0)
|
|
return {"status": "completed"}
|
|
|
|
|
|
class _BlockingBackoffSyncService(_SyncService):
|
|
def __init__(self) -> None:
|
|
super().__init__([False, False])
|
|
self.started = asyncio.Event()
|
|
self.release = asyncio.Event()
|
|
|
|
async def sync_mapping(
|
|
self,
|
|
mapping: dict[str, Any],
|
|
*,
|
|
force: bool,
|
|
) -> dict[str, Any]:
|
|
assert force is False
|
|
self.sync_calls += 1
|
|
self.seen_mappings.append(dict(mapping))
|
|
if self.sync_calls == 1:
|
|
self.started.set()
|
|
await self.release.wait()
|
|
return {
|
|
"status": "provider_backoff",
|
|
"failure_scope": "embedding_provider",
|
|
"retry_after_seconds": 60,
|
|
}
|
|
|
|
|
|
def _app(service: _SyncService) -> SimpleNamespace:
|
|
return SimpleNamespace(state=SimpleNamespace(llmwiki_sync_service=service))
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_mutation_sync_reserves_scan_until_remote_processing_finishes() -> None:
|
|
service = _SyncService([True, True, False, False, False, False])
|
|
task = schedule_local_wiki_mutation_sync(
|
|
_app(service),
|
|
{"id": "mapping-kb", "weknora_id": "remote-kb"},
|
|
delays=(0.001,),
|
|
)
|
|
|
|
assert task is not None
|
|
await asyncio.wait_for(task, timeout=1)
|
|
# One scan immediately after processing clears plus one short quiet
|
|
# confirmation protects against providers that publish the Wiki directory
|
|
# a few seconds after flipping the processing flag.
|
|
assert service.client.status_reads == 6
|
|
assert service.sync_calls == 2
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_mutation_sync_timeout_starts_from_latest_processing_observation(
|
|
monkeypatch: pytest.MonkeyPatch,
|
|
) -> None:
|
|
service = _SyncService([True, False, False, False])
|
|
clock = [0.0, 9.0, 11.0, 15.0, 16.0, 17.0]
|
|
monkeypatch.setattr(scheduler, "_MUTATION_SYNC_MAX_SECONDS", 10.0)
|
|
monkeypatch.setattr(
|
|
scheduler,
|
|
"monotonic",
|
|
lambda: clock.pop(0) if clock else 17.0,
|
|
)
|
|
|
|
task = schedule_local_wiki_mutation_sync(
|
|
_app(service),
|
|
{"id": "mapping-kb", "weknora_id": "remote-kb"},
|
|
delays=(0.0,),
|
|
)
|
|
|
|
assert task is not None
|
|
await asyncio.wait_for(task, timeout=1)
|
|
# The second poll happens after the original upload-based deadline. A
|
|
# confirmed in-progress observation must refresh that deadline so the
|
|
# completed snapshot is still synchronized.
|
|
assert service.sync_calls == 2
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_mutation_sync_stops_after_a_quiet_scan_observes_delta() -> None:
|
|
service = _SyncService(
|
|
[False, False],
|
|
results=[{"status": "completed", "added_pages": 1}],
|
|
)
|
|
task = schedule_local_wiki_mutation_sync(
|
|
_app(service),
|
|
{"id": "mapping-kb", "weknora_id": "remote-kb"},
|
|
delays=(0.001, 0.001, 0.001),
|
|
)
|
|
|
|
assert task is not None
|
|
await asyncio.wait_for(task, timeout=1)
|
|
assert service.sync_calls == 1
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_mutation_sync_retries_quiet_old_snapshot_with_bounded_delays() -> None:
|
|
service = _SyncService([False] * 6)
|
|
task = schedule_local_wiki_mutation_sync(
|
|
_app(service),
|
|
{"id": "mapping-kb", "weknora_id": "remote-kb"},
|
|
delays=(0.001, 0.001, 0.001),
|
|
)
|
|
|
|
assert task is not None
|
|
await asyncio.wait_for(task, timeout=1)
|
|
assert service.sync_calls == 3
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_mutation_sync_retries_service_processing_fence_then_confirms() -> None:
|
|
service = _SyncService(
|
|
[False] * 6,
|
|
results=[
|
|
{
|
|
"status": "partial",
|
|
"added_pages": 1,
|
|
"deferred_deletion": True,
|
|
},
|
|
{"status": "completed"},
|
|
{"status": "completed"},
|
|
],
|
|
)
|
|
task = schedule_local_wiki_mutation_sync(
|
|
_app(service),
|
|
{"id": "mapping-kb", "weknora_id": "remote-kb"},
|
|
delays=(0.001, 0.001, 0.001),
|
|
)
|
|
|
|
assert task is not None
|
|
await asyncio.wait_for(task, timeout=1)
|
|
assert service.sync_calls == 3
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_mutation_sync_stops_when_embedding_provider_breaker_is_open() -> None:
|
|
service = _SyncService(
|
|
[False],
|
|
results=[
|
|
{
|
|
"status": "provider_backoff",
|
|
"failure_scope": "embedding_provider",
|
|
"retry_after_seconds": 60,
|
|
}
|
|
],
|
|
)
|
|
task = schedule_local_wiki_mutation_sync(
|
|
_app(service),
|
|
{"id": "mapping-kb", "weknora_id": "remote-kb"},
|
|
delays=(0.001, 0.001, 0.001),
|
|
)
|
|
|
|
assert task is not None
|
|
await asyncio.wait_for(task, timeout=1)
|
|
assert service.sync_calls == 1
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_new_mutation_during_provider_backoff_is_consumed_and_cleaned_up() -> None:
|
|
service = _BlockingBackoffSyncService()
|
|
app = _app(service)
|
|
first = schedule_local_wiki_mutation_sync(
|
|
app,
|
|
{"id": "mapping-kb", "weknora_id": "remote-kb", "name": "old"},
|
|
delays=(0,),
|
|
)
|
|
assert first is not None
|
|
await asyncio.wait_for(service.started.wait(), timeout=1)
|
|
|
|
second = schedule_local_wiki_mutation_sync(
|
|
app,
|
|
{"id": "mapping-kb", "weknora_id": "remote-kb", "name": "new"},
|
|
delays=(0,),
|
|
)
|
|
assert second is first
|
|
service.release.set()
|
|
|
|
await asyncio.wait_for(first, timeout=1)
|
|
assert service.sync_calls == 2
|
|
assert [mapping["name"] for mapping in service.seen_mappings] == ["old", "new"]
|
|
assert app.state.llmwiki_mutation_sync_tasks == {}
|
|
assert app.state.llmwiki_mutation_sync_plans == {}
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_new_generation_wakes_backoff_and_uses_latest_mapping() -> None:
|
|
service = _SyncService(
|
|
[False, False],
|
|
results=[{"status": "completed", "added_pages": 1}],
|
|
)
|
|
app = _app(service)
|
|
first = schedule_local_wiki_mutation_sync(
|
|
app,
|
|
{"id": "mapping-kb", "weknora_id": "remote-kb", "name": "old"},
|
|
delays=(0.2,),
|
|
)
|
|
assert first is not None
|
|
await asyncio.sleep(0.05)
|
|
second = schedule_local_wiki_mutation_sync(
|
|
app,
|
|
{"id": "mapping-kb", "weknora_id": "remote-kb", "name": "new"},
|
|
delays=(0.2,),
|
|
)
|
|
|
|
assert second is first
|
|
# The original deadline has passed, but the newer mutation reset the
|
|
# debounce window; a lost wakeup would already have launched the scan.
|
|
await asyncio.sleep(0.17)
|
|
assert service.sync_calls == 0
|
|
await asyncio.wait_for(first, timeout=1)
|
|
assert service.seen_mappings[0]["name"] == "new"
|
|
|
|
|
|
def test_mapping_due_treats_string_false_as_not_processing() -> None:
|
|
now = datetime.now(UTC)
|
|
state = {
|
|
"index_revision": 1,
|
|
"state": "idle",
|
|
"last_success_at": now - timedelta(minutes=1),
|
|
}
|
|
remote = {
|
|
"updated_at": now,
|
|
"isProcessing": "false",
|
|
"processingCount": "0",
|
|
}
|
|
|
|
assert _mapping_is_due({}, state, remote, force_scan=False) is True
|
|
assert (
|
|
_mapping_is_due(
|
|
{},
|
|
state,
|
|
{**remote, "isProcessing": "true"},
|
|
force_scan=False,
|
|
)
|
|
is False
|
|
)
|
|
assert (
|
|
_mapping_is_due(
|
|
{},
|
|
state,
|
|
{**remote, "isProcessing": "true"},
|
|
force_scan=True,
|
|
)
|
|
is False
|
|
)
|