Files
EverOS/tests/integration/test_cli/test_backfill_phase2.py
T
Kendrick-SongandClaude Opus 4.7 b1441da607 refactor(config): make [embedding] and [rerank] soft dependencies (#361)
* refactor(config): make [embedding] and [rerank] soft dependencies

Make [embedding] and [rerank] soft runtime dependencies so a
freshly-onboarded user can run EverOS end-to-end with only [llm]
configured. Previously the server refused to start without embedding,
locking out anyone who just wanted keyword-only search.

## Capability tiers

- Tier 1 ([llm] only)                   KEYWORD search, add/flush, md
                                        writes, cascade sync
- Tier 2 ([llm] + [embedding])          + VECTOR / HYBRID search,
                                        reflection, skill extraction,
                                        backfill
- Tier 3 ([llm] + [embedding] + rerank) + AGENTIC search, knowledge

Tier upgrades require a server restart (capability accessors cache
for the process lifetime). Tier downgrades are read-safe: a Tier-3
user who drops [rerank] can still read/rename/delete existing
knowledge documents; only write/search endpoints return 422.

## What changed

- Component accessors — component/{embedding,rerank,llm}/accessor.py
  are the single process-wide provider singletons. service/* never
  maintains parallel singletons; it consumes get_embedding_capability()
  / get_rerank_capability() / get_llm_client() directly. Build-time
  ValueError from the factory is logged as capability_build_failed
  (was silently swallowed).
- Error mapping — ProviderNotConfiguredError -> 422 with everos.toml
  section hints (never EVEROS_* env-var strings).
  LanceDBMigrationError fails loud with escalating recovery guidance
  (restart -> wipe index). LLMNotConfiguredError in search maps to
  None for KEYWORD degradation.
- Nullable-vector LanceDB migration — schema v2 makes the vector
  column nullable so Tier-1 rows can land without embeddings.
  Migration is guarded by a cross-process memory_root_lock
  (fcntl.flock + anyio.to_thread) and runs optimize() per table
  after Phase-1 backfill to reclaim manifest bloat.
- Cascade — knowledge handlers register unconditionally (Tier-3 ->
  Tier-2/1 downgrade no longer strands DELETE); embed-requiring
  strategies use body-guards that check capability.available at
  execution time. _TABLE_SPECS has an import-time drift assertion
  against BUSINESS_SCHEMAS_WITH_VECTOR.
- `everos cascade backfill` CLI — Phase-1 (embed missing vectors) /
  Phase-2 (emit synthetic events for cascaded processing) / Phase-3
  (sync new skill files). Exit codes: 0 / 1 / 2 / 3 (server running
  preflight) / 4 (COMPLETED_WITH_FAILURES — per-row failures rolled
  up) / 130 (SIGINT). OMEConfig.crash_recovery_enabled=False in
  backfill engines prevents stale-RUNNING rows re-enqueuing into a
  smaller strategy registry.
- /health — reports capabilities + disabled_features per tier so ops
  can distinguish "boots but degraded" from "boots and full".
- Presentation split — memory / service / infra never import typer /
  click. TyperPresenter Protocol + run_backfill() live in
  entrypoints/cli/commands/_backfill_cmd.py. Enforced by
  import-linter.
- Startup hint — unconditional count_rows(filter="vector IS NULL")
  sweep emits unbackfilled_memory_rows (event name + hint text
  pinned) when Tier-1 rows exist. ParserLifespanProvider warms the
  everalgo.parser import at boot so /health doesn't block on first
  call.
- Knowledge upload UTF-8 short-circuit — _looks_like_utf8_text()
  routes text/* mime and known plaintext extensions (md/txt/rst)
  straight to UTF-8 decode instead of the parser. Prevents 503
  Multimodal-not-configured when Tier 3 sans [multimodal] uploads a
  markdown doc.

## Sync history with main (2 merges collapsed into this squash)

Merged origin/main at 6dcd3eb (v1.1.4 -> v1.2.0 adds OTel tracing,
/api/v2 alias, TracingLifespanProvider, per-cascade-embedding span
fix, memory-op instrumentation) and later at 42629df (PR #366
backfills v1.1.4 CWE-22 knowledge path traversal fix + cascade
retry-budget rework + errors.py -> core.errors.ExternalServiceError).
Key merge decisions:
- service/search.py adopts single wrap site — component.llm accessor
  already applies UsageRecordingClient when observability is on;
  service layer never keeps a parallel LLM singleton (Round-1 CR
  rule: "service layer never maintains parallel singletons").
- Knowledge router prefix moved to /knowledge; create_app() mounts
  it under both /api/v1 and /api/v2.
- Cascade retry classification uses ExternalServiceError from
  core.errors (cascade/errors.py deleted). _MAX_TOTAL_RETRIES=12
  cross-cycle budget preserved.
- Fixed backport typo: MemoryRoot.default() -> MemoryRoot.resolve()
  (no .default() classmethod exists — main PR #366 shipped a broken
  call).

## Verified layering

    $ git grep -l "^import typer\|^from typer" src/everos/{memory,service,infra}
    # empty
    $ git grep -l "^import click\|^from click" src/everos/{memory,service,infra}
    # empty

Memory / service / infra layers clean of CLI presentation libraries.

## Review history

Three rounds of Fable 5 (opus) code review across the pre-squash
commit history closed 38 findings total:
- Round 1: 10 findings (fail-loud migration, backfill hardening,
  knowledge router gate scoping, SearchManager guards, profile
  throttle lift)
- Round 2: 13 findings (hermetic test env, hot-reload doc drift,
  knowledge handler registration, Phase 3 sync guarantee, Phase 2
  idempotency, profile event-first path, OMEConfig crash-recovery
  gate, cross-process migration lock, batch embed per-row fallback,
  LanceDB optimize, typer/click layer split)
- Round 3: 15 Minor cleanups (accessor unification, marker revert,
  episode query hygiene, --verbose subcommand, parser lifespan warm,
  task-number scrub, temporal-overlap test, real-SIGINT slow mark,
  4 design-note back-references)

Full per-round context lives in the PR description on GitHub.

## Test plan

- make lint (ruff + import-linter 3 contracts + assets +
  deprecated-names + github-docs + datetime + OpenAPI drift)
- Hermetic env full pytest — 2027 passed / 7 deselected (7 = slow +
  live_llm markers)
- Manual e2e across Tier 1/2/3 (21/21 assertions across v1/v2
  double-mount and Tier 3 -> Tier 2 downgrade)
- /health reports correct capabilities + disabled_features per tier

## Known follow-ups

- .superpowers/sdd/followup-http-bridge.md (gitignored) — Path A for
  spec §10's "backfill 期间 EverOS 完全可用" promise
- _TABLE_SCHEMA_VERSION docstring — v3+ migrations need a version
  dispatch table
- extract_user_profile.py throttle-counter block — replace LanceDB
  count_by_owner with a sqlite memcell count

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

* fix(review): close 3 blockers surfaced by round-4 review

N1. cluster_repo.find_cluster_id_for_member was cross-owner-unsafe.
Its reverse index (member_type, member_id) alone cannot disambiguate
two owners whose entry_id happens to collide — entry_id is
deliberately only per-owner unique (see entries.py:47:
'Cross-user uniqueness is handled at the database layer via a
composite <user_id>_<entry_id> field; it is not encoded into the
EntryId string itself'). Phase 2's _scan_all_rows crosses all
owners, so on any multi-owner root, same-day seq=1 episodes under
different owners would either false-hit each other's cluster or
be silently skipped from clustering. Add required (app_id,
project_id, owner_id) keyword args + JOIN Cluster (which already
carries scope) to filter by parent scope. Prior signature had zero
production callers except two the same PR just added, so the
API break is contained. Regression test: two owners persist a
cluster each around the same entry_id, each lookup resolves to
its own owner's cluster, a third owner's lookup returns None.

N2. Ctrl-C / EOF at the y/N prompt was landing on the generic
except Exception branch (exit 2 with rich traceback) instead of
the exit-130 interrupt path. Root cause: typer 0.15+ vendored
click under typer._click, so typer.Abort and the standalone
click.exceptions.Abort are distinct classes. The interrupt-branch
catch only listed the standalone one; every existing 'abort'
test was manually raising click.exceptions.Abort so the miss
was a false-positive guard rail. Widen the catch to
(typer.Abort, click.exceptions.Abort) and declare click as a
first-class dependency (it was only pulled in via uvicorn).
Regression test: raise real typer.Abort() at the confirm step
and assert exit 130 + INTERRUPTED banner.

N3. _looks_like_utf8_text used mime.startswith('text/'), which
caught text/html as well. HTML uploads then bypassed everalgo's
_aparse_html — losing clean_html_for_llm (strips <script>/<style>
/<nav>/<iframe> + HTML comments) and the 1 MiB output cap. A
40 MiB .html with <script> bodies and <!-- prompt injection -->
comments would flow straight into the extraction LLM. Replace
with explicit allowlist {text/plain, text/markdown, text/x-rst,
text/x-markdown}; text/html and any future text/* mime now
default to the parser path. Test matrix asserts text/html →
False (was regressed as True by the earlier commit).

Hermetic env full pytest: 2033 passed / 7 deselected (+6 tests
from these regressions).

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

* fix(review): close round-4 major + minor + PR body errata

Round-4 review-driven cleanup. Blocker fixes (N1/N2/N3) landed in
561b5fe. This commit closes the remaining CONFIRMED items:

Major:
- J3 MemoryRoot.default() -> resolve() was a breaking public API
  rename this branch introduced (main still has default()). Adds
  default() as a backward-compat alias forwarding to resolve()
  with DeprecationWarning; CHANGELOG entry under Unreleased.
- J4 episode_repo.list_by_owner_after_ts(limit=N) truncates in
  fragment order (== insertion order), NOT newest-first. Docstring
  now spells out the trap so a future caller passing limit for a
  'newest N' window doesn't silently get the oldest N.
- J5.2 TyperPresenter.nothing_to_backfill picked colour via
  'could not be read' in message — a domain wording change
  would silently flip yellow -> green. Signature gains explicit
  scan_failed: bool kwarg; CLI colour-picks off the flag.
- J5.5 phase_header was Protocol-declared but never dispatched
  (run_backfill calls _print_phase_header directly). Removed the
  dead Protocol method + both no-op implementations.
- J6.3 3 inline from everos.core.errors import ... inside
  Phase 1/2/3 preflights promoted to a single top-level import.
- J7 subject-side embed failure was silently exit-0 because
  rows_processed advanced whenever any side wrote. Now: a row with
  a needed side still NULL counts as rows_failed (exit 4 =
  COMPLETED_WITH_FAILURES). Gated on spec.subject_of + row.needs_*
  + row.subject_text so non-Episode tables and subject-empty rows
  don't false-positive.
- J9 test_migration_cross_process.py did NOT actually test cross-
  process (all 5 tests mock memory_root_lock). Renamed to
  test_migration_lock_wiring.py; docstring now scopes it to
  'lock-invocation wiring' and points at test_core/…/test_locking.py
  for real flock coverage.
- J10 Phase 1 lacked the server-running preflight Phase 2/3 have.
  --phase all against a live server would burn Phase 1 embed API
  calls (real cost) before Phase 2 halted with exit 3. Phase 1 now
  probes _probe_ome_lock_available first; regression test in
  test_backfill_preflight.py; upgrade_path integration patches the
  probe so its in-process 'server + backfill' scenario stays valid.

Minor:
- M1 knowledge upload with NUL byte or filename > 255 bytes UTF-8
  used to raise ValueError/OSError at write_bytes → 500 with a
  half-written md left on disk. _safe_original_filename now
  rejects both up front with InvalidInputError (→ 400).
- M2 backfill optimize() now passes cleanup_older_than=timedelta(0)
  so older manifest versions are physically pruned (previous call
  compacted fragments but left the manifest chain on disk).
- M3 verify_business_schemas remediation text used to jump straight
  to 'rm -rf ~/.everos/.index/lancedb'; now walks restart → wipe.
- M5 multimodal/accessor.py capability_build_failed warning added
  so all four provider accessors log symmetrically (was silent).
- M7 test_knowledge_api parser-absence tests call
  parser_available.cache_clear() around the sys.modules patch so
  the lru_cache doesn't strand a stale True/False.
- M8 cascade_handler_embed_skipped (6 handlers) demoted INFO → DEBUG:
  Tier 1 imports were generating N × 6 handler-info lines per md.
- M10.1 test_drift_scenario_would_raise was a tautology (compared
  two hardcoded string sets, never touched the guard). Now
  monkey-patches BUSINESS_SCHEMAS_WITH_VECTOR to a superset and
  reloads _backfill, proving the import-time RuntimeError fires.
- M11 test_cascade_verbose_position subprocess.run calls gain
  env= — scrubs EVEROS_* from the developer environment so the
  same footgun as round-2 B1 doesn't re-appear inside subprocesses.
- M14 health() -> dict degraded the OpenAPI schema to
  additionalProperties: true. Introduce HealthResponse +
  HealthCapabilities Pydantic models so clients get real field
  shape; docs/openapi.json regenerated.
- M16 routes/knowledge.py:_require_knowledge_capabilities docstring
  claimed cascade.registry.build_handlers still gates
  knowledge_topic/knowledge_document, contradicting
  registry.py:177-194 (gate removed there, moved to HTTP layer).
  Rewritten to describe the current design accurately.

Hermetic env full pytest: 2037 passed / 7 deselected.

Explicitly deferred to followup:
- J1 lazy multimodal client (needs everalgo signature change)
- J2 tier definition (knowledge = Tier 3 whole)
- J5.1/3/4 broader presentation-split refactor
- J6.1/2 backfill dispatch and _backfill_table refactor
- J8 Phase 1 keyset pagination for bulk migration OOM
- M4 flock timeout/waiting log/re-entry (core/persistence refactor)
- M6 UTF-8 codec strategy (BOM/UTF-16/GBK)
- M9 count_by_owner monotonicity (latent, INTERVAL=1 short-circuits)
- M13 --phase all multi-phase combined-outcome test coverage
- M15 PR-marker rationale comments (16 occurrences, all inert)

M12 was refuted (both event names exist).

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-07-29 11:05:23 +08:00

315 lines
11 KiB
Python

"""Integration test for ``everos cascade backfill --phase clusters``.
Exercises the real Phase 2 scan + synthetic-event-emission path (see
``memory.cascade._backfill._run_phase_clusters``): seeds real episode /
agent-case rows in LanceDB, drives ``run_backfill(phase="clusters", ...)``,
and asserts the phase synthesizes exactly one ``EpisodeExtracted`` /
``AgentCaseExtracted`` event per row and fans them into the OME engine.
The OME engine itself is replaced with an in-memory spy
(``_FakeClusterEngine``, injected via monkeypatching
``_backfill._build_cluster_engine``) so this test exercises the scan +
event-synthesis contract without paying for a real APScheduler + sqlite
jobstore runtime — that end-to-end wiring is already covered by
``tests/integration/test_ome_strategies_integration.py``.
Covers: empty DB (nothing to backfill, no engine touched), declined
confirmation (no events emitted, exit 1), a small DB with episodes +
agent cases (correct event types / counts / fields, exit 0), and a
missing embedding capability (exit 2, mirroring Phase 1's guard).
One test (``test_real_offline_engine_grows_cluster_count``) does NOT
install the fake spy: it drives Phase 2 through the actual
``_build_cluster_engine`` path — a real ``OfflineEngine`` + APScheduler
runtime with ``trigger_profile_clustering`` registered — and asserts
``cluster_repo.count()`` grows. This is the one check the fake-engine
tests above cannot make: they prove the scan + event-synthesis contract,
but never exercise whether the real engine wiring actually drives a
cluster into existence.
"""
from __future__ import annotations
import hashlib
from collections.abc import AsyncIterator
from pathlib import Path
import pytest
from everos.component.embedding import EmbeddingCapability, EmbeddingProvider
from everos.component.utils.datetime import get_utc_now
from everos.config import load_settings
from everos.entrypoints.cli.commands._backfill_cmd import run_backfill
from everos.infra.ome.events import BaseEvent
from everos.infra.persistence.lancedb import (
AgentCase,
Episode,
agent_case_repo,
dispose_connection,
episode_repo,
)
from everos.infra.persistence.sqlite import cluster_repo, sqlite_manager
from everos.memory.cascade import _backfill as backfill_mod
from everos.memory.events import AgentCaseExtracted, EpisodeExtracted
_DIM = 1024
class _StubEmbedder(EmbeddingProvider):
dim = _DIM
async def embed(self, text: str) -> list[float]:
return [float(len(text) % 7)] * self.dim
async def embed_batch(self, texts): # type: ignore[no-untyped-def]
return [[float(len(t) % 7)] * self.dim for t in texts]
class _FakeClusterEngine:
"""Spy replacing the real OME engine: records emitted events instead of
dispatching them through APScheduler."""
def __init__(self) -> None:
self.emitted: list[BaseEvent] = []
self.started = False
self.stopped = False
async def start(self) -> None:
self.started = True
async def emit(self, event: BaseEvent) -> None:
self.emitted.append(event)
async def wait_idle(self, *, timeout: float = 30.0) -> bool: # noqa: ASYNC109
return True
async def stop(self) -> None:
self.stopped = True
@pytest.fixture
async def backfill_runtime(
tmp_path: Path, monkeypatch: pytest.MonkeyPatch
) -> AsyncIterator[Path]:
"""Tmp memory root + stub embedder + isolated sqlite engine.
Phase 2 touches both LanceDB (episode / agent_case scan) and sqlite
(``cluster_repo.count()``), so both process-wide singletons are
disposed and rebuilt against ``tmp_path`` around each test — mirrors
``backfill_runtime`` in ``test_backfill_phase1.py`` plus the sqlite
reset from ``test_ome_strategies_integration.py``.
"""
monkeypatch.setenv("EVEROS_ROOT", str(tmp_path))
load_settings.cache_clear()
await dispose_connection()
if sqlite_manager._engine is not None:
await sqlite_manager.dispose_engine()
monkeypatch.setattr(sqlite_manager, "_engine", None, raising=False)
monkeypatch.setattr(sqlite_manager, "_session_factory", None, raising=False)
import everos.component.embedding.accessor as acc
monkeypatch.setattr(
acc, "_capability", EmbeddingCapability(provider=_StubEmbedder())
)
yield tmp_path
await dispose_connection()
if sqlite_manager._engine is not None:
await sqlite_manager.dispose_engine()
@pytest.fixture
def fake_engine(monkeypatch: pytest.MonkeyPatch) -> _FakeClusterEngine:
"""Install the spy engine in place of ``_build_cluster_engine``."""
engine = _FakeClusterEngine()
monkeypatch.setattr(backfill_mod, "_build_cluster_engine", lambda: engine)
return engine
def _episode(entry_id: str, *, owner_id: str = "u1") -> Episode:
return Episode(
id=f"{owner_id}_{entry_id}",
entry_id=entry_id,
owner_id=owner_id,
owner_type="user",
session_id="s1",
timestamp=get_utc_now(),
parent_id=f"mc_{entry_id}",
sender_ids=[owner_id],
episode=f"episode body {entry_id}",
episode_tokens=f"episode body {entry_id}",
md_path="users/u1/episodes/episode-2026-01-01.md",
content_sha256=hashlib.sha256(entry_id.encode()).hexdigest(),
vector=[0.1] * _DIM,
)
def _agent_case(entry_id: str, *, agent_id: str = "agent1") -> AgentCase:
return AgentCase(
id=f"{agent_id}_{entry_id}",
entry_id=entry_id,
owner_id=agent_id,
owner_type="agent",
session_id="s1",
timestamp=get_utc_now(),
parent_id=f"mc_{entry_id}",
quality_score=0.8,
task_intent=f"task intent {entry_id}",
task_intent_tokens=f"task intent {entry_id}",
approach=f"approach {entry_id}",
approach_tokens=f"approach {entry_id}",
md_path="agents/agent1/.cases/agent_case-2026-01-01.md",
content_sha256=hashlib.sha256(entry_id.encode()).hexdigest(),
vector=[0.1] * _DIM,
)
async def test_empty_db_reports_nothing_to_backfill(
backfill_runtime: Path,
fake_engine: _FakeClusterEngine,
capsys: pytest.CaptureFixture[str],
) -> None:
code = await run_backfill(phase="clusters", auto_yes=True)
out = capsys.readouterr().out
assert code == 0
assert "Nothing to backfill" in out
assert fake_engine.emitted == []
assert fake_engine.started is False
async def test_declined_confirmation_emits_nothing_and_exits_one(
backfill_runtime: Path,
fake_engine: _FakeClusterEngine,
monkeypatch: pytest.MonkeyPatch,
) -> None:
await episode_repo.add([_episode("ep1")])
monkeypatch.setattr("typer.confirm", lambda *a, **k: False)
code = await run_backfill(phase="clusters", auto_yes=False)
assert code == 1
assert fake_engine.emitted == []
assert fake_engine.started is False
async def test_episodes_and_cases_synthesize_matching_events(
backfill_runtime: Path,
fake_engine: _FakeClusterEngine,
capsys: pytest.CaptureFixture[str],
) -> None:
await episode_repo.add([_episode("ep1"), _episode("ep2")])
await agent_case_repo.add([_agent_case("ac1")])
code = await run_backfill(phase="clusters", auto_yes=True)
out = capsys.readouterr().out
assert code == 0
assert fake_engine.started is True
assert fake_engine.stopped is True
assert len(fake_engine.emitted) == 3
episode_events = [e for e in fake_engine.emitted if isinstance(e, EpisodeExtracted)]
case_events = [e for e in fake_engine.emitted if isinstance(e, AgentCaseExtracted)]
assert len(episode_events) == 2
assert len(case_events) == 1
assert all(e.source == "pipeline" for e in episode_events)
assert all(e.event_id.startswith("backfill_") for e in fake_engine.emitted)
assert {e.owner_id for e in episode_events} == {"u1"}
assert case_events[0].agent_id == "agent1"
assert case_events[0].task_intent == "task intent ac1"
assert case_events[0].quality_score == pytest.approx(0.8)
assert "phase 2 complete" in out
async def test_missing_embedding_capability_returns_exit_two(
backfill_runtime: Path,
fake_engine: _FakeClusterEngine,
monkeypatch: pytest.MonkeyPatch,
) -> None:
"""Mirrors Phase 1's own guard: no embedding provider configured means
the strategies can't re-embed, so the phase fails clean with exit 2
rather than emitting events into strategies that will crash."""
import everos.component.embedding.accessor as acc
await episode_repo.add([_episode("ep1")])
monkeypatch.setattr(acc, "_capability", EmbeddingCapability(provider=None))
code = await run_backfill(phase="clusters", auto_yes=True)
assert code == 2
assert fake_engine.emitted == []
assert fake_engine.started is False
async def test_reflection_merged_episodes_excluded_from_synthesis(
backfill_runtime: Path,
fake_engine: _FakeClusterEngine,
) -> None:
"""Reflection-merged episodes (``parent_type="cluster"``) must not be
re-synthesized as pipeline events.
``trigger_profile_clustering`` excludes Reflection output via
``applies_to=lambda e: e.source == "pipeline"`` — but that guard is
defeated if the Episode scan hands it a synthetic event whose
``memcell_id`` is actually a cluster id (see Important 1 review
finding on ``_backfill._scan_all_rows`` / ``_run_phase_clusters``).
Only the pipeline-sourced row should reach the fake engine.
"""
pipeline_episode = _episode("ep1")
merged_episode = _episode("ep_merged").model_copy(
update={
"id": "u1_ep_merged",
"parent_type": "cluster",
"parent_id": "cluster_abc123",
}
)
await episode_repo.add([pipeline_episode, merged_episode])
code = await run_backfill(phase="clusters", auto_yes=True)
assert code == 0
episode_events = [e for e in fake_engine.emitted if isinstance(e, EpisodeExtracted)]
assert len(episode_events) == 1
assert episode_events[0].memcell_id == pipeline_episode.parent_id
async def test_real_offline_engine_grows_cluster_count(
backfill_runtime: Path,
) -> None:
"""Phase 2 through the REAL ``_build_cluster_engine`` — no fake spy.
Deliberately does not use the ``fake_engine`` fixture: it drives an
actual ``OfflineEngine`` + APScheduler runtime with
``trigger_profile_clustering`` registered (see
``_backfill._build_cluster_engine``), proving the phase's own engine
wiring — not just the scan/synthesis contract the other tests in
this module pin — actually grows ``cluster_repo.count()``.
``trigger_profile_clustering`` re-embeds each episode's text via
``get_embedding_capability().require()`` — which ``backfill_runtime``
already stubs with the deterministic ``_StubEmbedder``, so no
per-strategy patch is needed here. No agent cases are seeded, so
``trigger_skill_clustering`` (the engine's other registered
strategy, whose ``cluster_by_llm`` fallback needs an LLM) never
fires, keeping this test LLM-free.
"""
await episode_repo.add([_episode(f"ep{i}") for i in range(5)])
# ``run_backfill`` itself creates the ``cluster`` table lazily (see
# ``_ensure_cluster_schema``) — call it here too so the "before" count
# can be taken without racing that lazy create.
await backfill_mod._ensure_cluster_schema()
clusters_before = await cluster_repo.count()
assert clusters_before == 0
code = await run_backfill(phase="clusters", auto_yes=True)
clusters_after = await cluster_repo.count()
assert code == 0
assert clusters_after > clusters_before