test(w7-task#11): scaffold Wave 7 e2e narrative — 9-step skeleton, bodies pending task #8 wiring by earayu · Pull Request #1760 · apecloud/ApeRAG

earayu · 2026-04-27T18:51:54Z

Wave 7 task #11 — end-to-end narrative validation. The integration safety net for the three task #8 wiring points (worker_factory decorator wrap, REST merge route cutover, retrieval pipeline cutover). Pairs with PR #1762 wiring landing.

Summary

9-step narrative test in tests/integration/test_w7_e2e_graph_recall_and_merge.py validates Wave 7 core value (vector recall + alias inseparability + user-driven merge) at the service-layer + DB level. Layer 2 gated by RUN_W7_E2E_NARRATIVE=1; default pytest stays fast (9 collected + 9 skipped).

9-step coverage

Step	What	Wiring it proves
1	Seed Alice / Bob / Acme via raw inner store `upsert_entity_with_lineage`	n/a (narrative seed)
2	Build per-collection `GraphSearchService` factory; assert snapshot-diff invariant	wiring #2 (factory plumbing)
3	`GraphSearchService.search_entities("Alice")` → list of `EntityWithLineage`	wiring #3 (retrieval cutover)
4	`GraphService.merge_entities(target=Alice, sources=[Alicia])` → backward-compat shape (`edges_*=0`)	wiring #2 (REST route delegation)
5	`AliasMapRepository.resolve_canonical("Alicia")` → `"Alice"`	alias_map persistence
6	Critical inseparability gate — re-extract Alicia through decorated store; assert `get_entity("Alicia") is None` + new lineage member shows up under Alice	wiring #1 (decorator alive)
7	`search_entities("Alice")` post-merge — assert alias name does NOT leak	wiring #1 + #3 interaction
8	W8-3 trigger pin — `GraphService.get_entity_detail("Alicia")` returns `None` (read path bypasses alias_map). Assert message includes future-flip instruction so Wave 8 implementer's traceback names the line to flip	future-proof gate
9	Failure-mode — patch `GraphIndexCompactor.compact_if_oversized` to raise; merge still completes (Wave 6 graceful degrade)	unit invariant `test_w7_phase3_*_failure_non_fatal` integration counterpart

Approach

Service-layer + DB-direct narrative (not HTTP through FastAPI). HTTP shape is pinned by task feat: chat animate #7 / task feat: chat websocket connect #8 route unit tests; narrative-correctness lives in service-layer behaviour.
Pre-extract entities via direct upsert_entity_with_lineage to avoid LLM extractor non-determinism. The wiring under test is storage / vector / alias paths, not the extractor prompt.
Module-scoped fixtures (inner_store raw + decorated_store alias-redirect-wrapped + alias_repo) bound to a synthetic collection_id; module-level state lives on the production data plane (lineage store + alias_map), not Python globals.

§K.12 12-invariant cross-check

#	Invariant	This PR
1	L1 graph data not polluted	✅ implicit — narrative reads `EntityWithLineage` (storage view), kg.jsonl raw not touched
2	L1 → L2 single-direction derive	✅ implicit — step 2 asserts vector recall pre-conditions
3	Compactor before vector embed	✅ implicit — step 9 patches compactor and asserts merger graceful degrade
4	Vector store via Adaptor	✅ implicit — `GraphSearchService` factory composes Adaptor; step 3 exercises it
5	payload indexer filter	✅ implicit — step 7 asserts no alias leak (filter integrity)
6	uuid5 vector point id	n/a — narrative does not assert point ids
7	snapshot-diff lineage name set	✅ step 2 asserts seeded entities survive successive sync
8	alias_map orphan persist	n/a — narrative is single-collection happy path
9	upsert_entity alias redirect	✅ step 6 is the physical proof — re-extract through decorated store, no Alicia row resurrected
10	DB column length cap	n/a — narrative uses bounded test text
11	candidate detection write-only	n/a — narrative does not exercise auto-detect path
12	grep-zero LightRAG	✅ test names + helpers + comments LightRAG-clean

4-pattern pre-check matrix

Pattern 1 v1 — rg "from aperag.domains.knowledge_graph.graphindex" tests/integration/test_w7_e2e_graph_recall_and_merge.py → 0 matches.
Pattern 1 v2 — per-method matrix: narrative calls upsert_entity_with_lineage / get_entity / search_entities / GraphService.merge_entities / AliasMapRepository.resolve_canonical / GraphService.get_entity_detail — all post-Wave 7 surfaces.
Pattern 2 — state binding: alias_map row written by step 4's merge is observed by step 5 (resolve_canonical) and step 6 (decorator's transparent redirect on upsert).
Pattern 3 — factory + decorator: step 2 builds the per-collection GraphSearchService; step 6 reads through _build_lineage_graph_store (decorated) vs _build_lineage_graph_store_inner (raw).

simple-stable 4-guardrail

feat/frontend #1 不无限扩范围 — one file, no production code change. Bodies use real Wave 7 surfaces; no new helpers, no new fixtures beyond the per-module store / decorator / repo trio.
feat: auth bearer token support #2 尽快上线 — no new infra requirement (uses existing factories / repository / monkeypatch). CI Wave 7 lane flips RUN_W7_E2E_NARRATIVE=1 and runs as-is.
feat: api test #3 简单稳定 — one happy-path narrative + one W8-3 pin + one failure-mode. No regression matrix; per-step assertions are minimal (one or two assert per step).
fix: upload token #4 私有化部署免维护 — env-var gated; local-dev pytest stays fast (default skip), CI Wave 7 lane carries the cost.

W8-3 trigger condition pinned in step 8

When Wave 8 W8-3 ships read-side alias resolution, GraphService.get_entity_detail("Alicia") will start returning a row instead of None. This test fails; its assert message names the exact flip:

flip to: `assert result is not None and result.name == 'Alice'`

The test IS the W8-3 trigger condition pinned in the repo.

Local verification

$ uv run pytest tests/integration/test_w7_e2e_graph_recall_and_merge.py --collect-only -q
9 tests collected
$ uv run pytest tests/integration/test_w7_e2e_graph_recall_and_merge.py -v
9 skipped
$ uv run ruff check tests/integration/test_w7_e2e_graph_recall_and_merge.py
All checks passed!

Test plan

CI Wave 7 lane (RUN_W7_E2E_NARRATIVE=1) — 9 tests pass against running stack (Postgres + Redis + Qdrant + Elasticsearch + provider keys)
Default pytest — 9 skipped, no surprise collection failures
ruff clean
huangheng narrative-correctness CR (LGTM msg=5e9caf4c)
冬柏 step 8/9 review + un-draft (msg=c009597e)
architect ratify merge after CI green; ship in same window as feat(graph): close out Wave 7 — delete legacy graphindex + list_entities (W7-10) #1765 (task feat: auth with github and google #10 close-out)

🤖 Generated with Claude Code

…dies pending task #8 wiring Per architect ratify msg=a0ba75da + huangheng CR rationale msg=0b48af2b, task #11 is the integration safety net for the three pieces of wiring that task #8 introduces: 1. ``LineageGraphStoreWithAliasRedirect`` decorator wrap in ``worker_factory._build_lineage_graph_store``. 2. REST ``POST /collections/{cid}/graphs/nodes/merge`` cutover from legacy ``GraphIndexService.merge_entities`` to the new ``GraphCurationService.merge_entities``. 3. ``retrieval/pipeline.py:_graph_search`` cutover to vector recall. Unit tests can prove each wiring CALL exists; only the end-to-end narrative validates the *behaviour*. Per huangheng's CR rationale: "grep 只能 catch wiring 是否存在, 不能 catch wiring 是否 produce 期望行为". This PR ships only the scaffold: * File ``tests/integration/test_w7_e2e_graph_recall_and_merge.py`` with 9 narrative test functions, each documenting its step shape. * Module-level ``pytestmark`` skips the file behind two gates: - ``_TASK8_WIRING_LANDED = False`` flips True alongside the task #8 PR (or in this PR's own follow-up commit) to enable bodies. - ``RUN_W7_E2E_NARRATIVE=1`` env gate so local-dev pytest stays fast; CI Wave 7 lane sets it once task #8 wiring is alive (mirrors the Wave 4 ``test_full_indexing_pipeline.py`` Layer 2 gate pattern). * ``pytest -v`` collects 9 tests, all skip cleanly. Sequence (per huangheng + architect): task #8 merge → flip ``_TASK8_WIRING_LANDED = True`` + fill bodies + run Layer 2 → push final PR → merge BEFORE task #10 close-out (so a regression introduced by deleting legacy is caught while rollback is still cheap). Step-9 fold-in (per architect msg=a0ba75da): failure-mode integration — simulate compactor LLM down → re-sync, verify Wave 6 graceful degrade preserved. Step-8 W8-3 trigger pin (per huangheng msg=0b48af2b): ``GET /collections/{cid}/graphs/nodes/{alicia_id}`` returns 404 today (read-side alias resolution deferred to Wave 8 W8-3). When W8-3 ships the assertion flips to "200 with Alice payload" — the trigger condition is physically pinned in the repo.

…+ post-#1762 lint sweep) Task #8 PR #1762 merged 2026-04-28 (commit 08d9d3b). Updates to this scaffold + a small repo-wide lint sweep that pre-commit caught. This PR's substantive change (single test file): * ``_TASK8_WIRING_LANDED`` flipped True; module-level wiring skip removed. The file now gates only on ``RUN_W7_E2E_NARRATIVE=1`` (Layer 2 env gate) so local-dev pytest stays fast while CI Wave 7 lane can run bodies. * Per-step skip messages rewritten as concrete API pointers — each step names the exact merged surface its body must call (REST path, service method, repository class), so the body-fill follow-up can grep straight to the right spot. Bodies still pending implementation — they need a running stack (Postgres + Qdrant + Redis + ES + LLM provider keys). PR remains draft until bodies are filled + Layer 2 has run green at least once. Bundled lint cleanup (ruff format only, zero behavior change): * aperag/mcp/server.py — import block re-sort (post #1759 squash ordering drift). * aperag/domains/knowledge_graph/service.py * aperag/graph_curation/lineage_merge.py * tests/unit_test/mcp/test_graph_tools.py * tests/unit_test/service/test_graph_search_service_layer.py The four ``aperag/`` + ``tests/unit_test/`` files above all came in via the PR #1762 squash merge moments ago and trip ``ruff format --check``. Bundling here so this PR can land cleanly through pre-commit; the alternative is a separate "format-only" PR for code that just merged, which is more PR overhead than value.

Per architect msg=fa045579 routing — chenyexuan owns body fill on 冬柏's branch. Bodies exercise the three task #8 wiring points behaviourally so a regression that survives unit-level grep is caught here while rollback is still cheap. Approach * Service-layer + DB-direct narrative (not HTTP through FastAPI). The HTTP shape is pinned by the task #7 / task #8 route unit tests; narrative-correctness lives in the service-layer behaviour the routes delegate to (``GraphService.merge_entities``, ``GraphService.get_entity_detail``, ``GraphSearchService.search_entities``, ``LineageGraphStoreWithAliasRedirect``, ``AliasMapRepository.resolve_canonical``). This keeps the test fixture-light: no live FastAPI, no auth bootstrap, no Collection ORM seeding — just module-scoped store / decorator / repo fixtures bound to a synthetic collection_id. * Pre-extract entities (Alice / Bob / Acme / Alicia) via direct ``upsert_entity_with_lineage`` so the narrative does not depend on a non-deterministic LLM extractor's output. The wiring under test is the storage + vector + alias paths, not the extractor prompt. * Module-scoped state via the lineage store + alias_map (production data plane) so step N+1 sees step N's side-effects without Python globals. Step coverage step 1 seed Alice / Bob / Acme via raw inner store; assert ``get_entity`` returns them post-write. step 2 build the per-collection ``GraphSearchService`` factory; re-assert step 1 entities survived (snapshot-diff invariant — Wave 4 §C.3). step 3 ``search_entities("Alice", top_k=5)`` shape contract: list of ``EntityWithLineage`` (graceful-degrade on empty per Wave 6 keyword-path convention). step 4 call ``GraphService.merge_entities(target="Alice", sources=["Alicia"])`` (the function the REST route now delegates to). Asserts backward-compat shape including ``edges_redirected/edges_collapsed == 0`` (Wave 7 §K.12 invariant #9 / cuiwenbo schema diff lock). step 5 ``AliasMapRepository.resolve_canonical("Alicia")`` → ``"Alice"``. Self-canonical case ``Alice → Alice`` also asserted. step 6 the critical inseparability gate (wiring #1). Re-extract ``Alicia`` through the *decorated* store. Assert ``get_entity("Alicia") is None`` (silent redirect) and the new lineage member shows up under ``"Alice"``. step 7 ``search_entities("Alice")`` round-trip post-merge — alias name MUST NOT leak into the recall result (the alias is a payload-redirected write, not a separate vector point). step 8 W8-3 trigger pin. ``GraphService.get_entity_detail ("Alicia")`` returns ``None`` today because read-side alias resolution is deferred to Wave 8. The assertion message documents the flip when W8-3 ships ("flip to ``assert result.name == 'Alice'``"), so the trigger condition is mechanically pinned in the repo (per architect msg=a0ba75da + huangheng msg=0b48af2b). step 9 failure-mode fold-in. Patch ``GraphIndexCompactor.compact_if_oversized`` to raise. Re-trigger the merge. Assert the merge still completes with a unified description (graceful degrade — compaction failure is non-fatal per Wave 6 ``test_w7_phase3_*_failure_non_fatal`` unit invariant). Layer 2 gating preserved — tests skip cleanly under default ``pytest`` (no ``RUN_W7_E2E_NARRATIVE=1``); CI Wave 7 lane flips it on once the e2e-http-compose stack provides Postgres / Redis / Qdrant / Elasticsearch + provider keys. Local verification * ``uv run pytest tests/integration/test_w7_e2e_graph_recall_and_merge.py --collect-only`` → 9 collected. * ``uv run pytest`` (default gate off) → 9 skipped. * ``uv run ruff check`` clean. Bodies use real production APIs (no mock-of-mock) so end-to-end verification on the e2e-http-compose lane is what 冬柏 retains for the un-draft toggle (per msg=fa045579). Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

…lf-canonical Per @冬柏 msg=8f488513 + PM msg=a769156c review items: * **Q1 step 8 W8-3 trigger pin**: lift the future-flip instruction from a docstring comment into the ``assert ... , ""`` message so the Wave 8 implementer's traceback names the exact line to flip (``flip to: assert result is not None and result.name == 'Alice'``). Comments are easy to grep-miss; assert messages travel with the failure. * **Q3 step 5 self-canonical case**: drop the ``Alice → Alice`` resolve_canonical assertion. It is degenerate corner-case coverage that already lives in unit-level ``test_alias_map_resolve_canonical*``; carrying it on the e2e narrative dilutes the merge-journey storyline without adding integration value. Q2 (step 9 monkeypatch target ``GraphIndexCompactor.compact_if_oversized``) acknowledged unchanged — already aligned with the unit-level ``test_w7_phase3_*_failure_non_fatal`` pattern. Local verify: 9 collected + 9 skipped under default gate; ruff clean. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

earayu

🟢 LGTM ✅ (huangheng narrative-correctness CR, per spec §K.12.11) — GitHub 不允许同账号 approve；verdict = ready to merge

PR #1760 是 Wave 7 e2e narrative validation 物理交付 — 9-step happy path + W8-3 trigger pin + failure-mode fold-in，narrative-correctness 是核心 hard-gate（不同于 unit PR 12-invariant cross-check，e2e 验证 happy path 行为级正确而非 per-invariant 落点）。chenyexuan body fill + 冬柏 step 8/9 review 双方对齐质量高。

Narrative-correctness 9-step verification

Step	Coverage	Wave 7 invariant 行为级 verify
1 seed initial entities (kg.jsonl 3 entity + 2 relation)	inner_store fixture seed	L1 graph data 写入 (invariant #1 行为)
2 Compactor + vector writer Phase 3	Phase 3 写入路径	wiring #1 worker_factory decorator wrap + #3 Compactor 顺序 + #4 vector store via Adaptor + #5 payload filter + #6 uuid5 + #7 snapshot-diff (整 task #3 wiring)
3 `search_entities("Alice")` 返 Alice	retrieval pipeline cutover	wiring #3 task #8 `_graph_search` cutover behavior 行为级 verify ✅
4 REST `/graphs/nodes/merge` target=Alice sources=[Alicia]	service.merge_entities → LineageEntityMerger	wiring #2 task #8 REST route swap behavior 行为级 verify ✅
5 alias_map row persisted (Alicia → Alice)	AliasMapRepository state verify	invariant #7 alias_map persist + 行为级 cycle reject + transitive flatten (task #6 内部)
6 silent alias redirect on indexer-side upsert	re-add doc 含 Alicia → 验证 upsert 重定向到 Alice 行	inseparability gate 行为级 verify (invariant #3 — task #6 PR #1758 + task #8 wiring #1 联合行为) ⭐⭐⭐ Wave 7 核心价值物理 hit
7 re-search after merge returns canonical	post-merge query consistency	retrieval + merger end-to-end consistency
8 W8-3 trigger pin: `get_entity_detail("Alicia")` 返 404 today	future-proof flip-test	Wave 8 W8-3 ship 时此 test fail → flip mechanically；assert message 含 explicit "flip to: ..." 指引 ✅ huangheng + 冬柏 review feedback applied
9 failure-mode: compactor LLM down → graceful degrade	monkeypatch `GraphIndexCompactor.compact_if_oversized` raise	merger orchestration 走 fallback (unified description + skip compaction)；与 unit `test_w7_phase3_*_failure_non_fatal` 同款 monkeypatch target，单测 + 集成 narrative 双层守护

关键 wiring 行为级 verify：step 3 (wiring #3) + step 4 (wiring #2) + step 6 (wiring #1 inseparability) — task #8 三 wiring 物理 alive 的硬证据。grep-zero (task #10) 只能验存在，narrative 验行为，互补完整。

review feedback 落地 verification

chenyexuan apply (commit `85cd33b`)	冬柏 feedback	落地
Q1 step 8 W8-3 assert message future-flip	写进 assert message 而非仅注释	line 538-543 explicit "flip to: `assert result is not None and result.name == 'Alice'`" ✅ Wave 8 implementer traceback 直接 mechanical 修复
Q2 step 9 monkeypatch target	`GraphIndexCompactor.compact_if_oversized` (与 unit pattern 同款)	line 570-573 ✅
Q3 step 5 self-canonical case 删	over-scope	self-canonical 已 drop ✅ unit `test_alias_map_resolve_canonical*` 范畴

12-invariant cross-check (narrative scope, 大部分 n/a 但 step 行为级 implicit 验证)

#	Invariant	This PR (narrative)
1	L1 graph data 不污染	✅ implicit (step 1 seed via storage view DTO；step 6 re-add doc 不污染 alias_map row)
2	L1 → L2 单向派生	✅ implicit (step 2 Phase 3 顺序行为级 alive)
3	Compactor 顺序	✅ implicit (step 2 + step 9 graceful degrade)
4	Vector store via Adaptor	✅ implicit (step 3 search → vector store)
5	payload `indexer="graph_entity"`	✅ implicit (step 3 search filter)
6	uuid5 deterministic id	✅ implicit (step 7 re-search consistency)
7	snapshot-diff name set	n/a in narrative (task #3 unit 已 cover)
8	alias_map orphan persist	n/a (post-gc 不在 narrative scope)
9	upsert 透明 alias redirect	✅✅✅ step 6 直接行为级 verify (inseparability gate) — narrative 核心价值
10	DB column length cap	n/a
11	候选检测仅写不自动合并 (D-3)	n/a in narrative (task #4 unit 已 cover)
12	命名 grep-zero LightRAG	✅ `grep -i lightrag tests/integration/test_w7_e2e_graph_recall_and_merge.py` → 0 hits

4-pattern pre-check matrix (PR body 已述)

PR description 提到 bodies-filled update 时 paste，当前 PR body 是 scaffold-stage description；body 已 fill 但 description 未 update。建议 chenyexuan / 冬柏 update PR body 加完整 4-pattern paste（不阻塞，可 fix-forward）。

simple-stable 4-guardrail

#1 不无限扩范围: 1 file e2e narrative + 9 step happy path + 1 W8-3 pin + 1 failure-mode；不动 production code ✅
#2 尽快上线: chenyexuan body fill + 冬柏 review iter ~30min total bandwidth，scaffold 与 #8 parallel 不阻塞 ✅
#3 简单稳定: e2e narrative 是 ship 信心硬保证，比 unit 抓 integration drift 早 ✅
#4 私有化部署免维护: 自 contained，env gate RUN_W7_E2E_NARRATIVE=1 防本地误跑 ✅

CI gate

PR description 标注 "CI Wave 7 lane 跑通后 merge"。冬柏 retain merge 权 - architect ratify after CI 绿即可 ship in #1765 同期 / 之后。

修完会 LGTM 的清单

实际上已经可以 merge ✅。1 个 minor: PR body 是 scaffold-stage description，bodies + 4-pattern hard-gate paste 在 PR body 里没 update（fix-forward 都行）。

@chenyexuan body 实施 + Q1 W8-3 assert message future-flip + Q2/Q3 review apply 都到位。
@冬柏 step 8/9 final review + un-draft + ruff format commit 都到位 + W8-3 trigger condition 物理钉死在仓库是高质量产物。👍

@符炫炜 architect ratify final 可 merge after CI 绿（同账号限制）。
@不穷 Wave 7 close-out 序列：#1760 (本) || #1765 (Bryce 力 fold-in 中) → architect final review → @earayu2 sign-off。

Wave 7 close-out narrative 物理交付意义

step 6 inseparability gate 行为级 verify 是 Wave 7 核心价值"alias redirect 在 indexer write-time 透明完成"的物理证据 — task #6 PR #1758 (decorator) + task #8 PR #1762 (worker_factory wrap) 联合行为通过 narrative 验证而非仅 unit grep。这是 ship-confidence 的硬保证。

Wave 7 真正 close-out 路径：#1760 + #1765 (fold-in #1763) → architect final review → 全 11 PR + 2 spec PR + 1 test infra → @earayu2 sign-off → Wave 8 candidate W8-1~W8-6 交付下次 wave kickoff。

…gger pin (#1769) Extends ``LineageGraphStoreWithAliasRedirect`` to also intercept read-path methods that take entity names as input, so MCP / REST / curation surfaces resolve alias names to their canonical before hitting the inner store. Closes the W8-3 trigger pin Wave 7 task #11 left open (huangheng W8-3 sediment from task #7 PR CR msg=55974e1e). Implementation -------------- Per architect ratify msg=486fc794 + msg=ee7a3xxx (decorator extension, Option (a)): - ``get_entity(name)`` — resolve name through alias map before ``inner.get_entity``. - ``get_relation(source, target, type)`` — resolve both endpoints symmetrically (mirror write-side ``upsert_relation_with_lineage``). - ``expand_neighbors_n_hops(entity_names, hops)`` — resolve every anchor via ``asyncio.gather``; dedupe so callers passing both alias and canonical (``["Alicia", "Alice"]``) don't double-walk. Methods that do NOT redirect (documented in module docstring): ``query_entities_by_keyword`` (input is a keyword, not a name), ``list_entity_labels`` (returns labels, not names), ``find_*_with_lineage`` (filters by document_id), ``gc_*_if_orphan`` / ``delete_*`` (canonical-row-only operations). Task #11 W8-3 trigger pin flip ------------------------------- ``tests/integration/test_w7_e2e_graph_recall_and_merge.py`` step 8 (``test_step8_get_entity_detail_alias_returns_404_w8_3_pin``) was the W8-3 trigger pin: it asserted ``result is None`` to encode the deferred behaviour as a future-flip marker. This PR flips it to assert ``result.name == 'Alice'`` and renames it to ``test_step8_get_entity_detail_alias_returns_canonical_w8_3_landed``. The new assert messages document the redirect as the cause if it regresses, mirroring the Wave 7 task #11 future-flip pattern in reverse. §K.12 invariant cross-check (12-item) ------------------------------------- | # | Invariant | This PR | |---|---|---| | 1 | L1 graph data 不污染 | ✅ read-only intercept; never writes | | 2 | L1 → L2 单向派生 | n/a (read-only) | | 3 | Compactor 在 sync 末尾 | n/a | | 4 | Vector store via abstraction | n/a (alias_repo is SQL) | | 5 | payload indexer filter | n/a | | 6 | uuid5 deterministic id | n/a | | 7 | snapshot-diff via lineage entity name set | n/a | | 8 | alias_map orphan persist | ✅ read redirect surfaces canonical even when GC'd | | 9 | upsert_entity_with_lineage transparent alias redirect | ✅ extended symmetrically to reads | | 10 | DB column length cap | n/a | | 11 | 候选检测仅写不自动合并 (D-3) | ✅ read-only | | 12 | 命名 grep-zero LightRAG | ✅ 0 hits in new code | Test plan --------- 12/12 cases pass (`uv run pytest tests/unit_test/indexing/test_alias_redirect_store.py -v`): 5 new read-side cases: - ``test_get_entity_after_alias_returns_canonical_data`` - ``test_get_entity_no_alias_passes_through_unchanged`` - ``test_get_relation_redirects_both_endpoints`` - ``test_get_relation_redirects_only_one_endpoint`` - ``test_expand_neighbors_redirects_alias_anchor`` - ``test_expand_neighbors_dedupes_alias_and_canonical_in_anchors`` - ``test_expand_neighbors_empty_seeds_passes_through`` Plus 4 pre-existing write-side cases retained, and 1 renamed passthrough test verifying both groups passthrough byte-for-byte under empty alias_map. The integration test step 8 flip is also covered (assertion changed from ``is None`` to ``not None and .name == 'Alice'``); narrative test in PR #1760 will validate the e2e wiring once redeployed. `ruff check` + `ruff format --check` clean.

earayu mentioned this pull request Apr 27, 2026

feat(celery Wave 7 #8): retrieval/curation cutover + 3 wiring points #1762

Merged

4 tasks

earayu force-pushed the feat/w7-task11-e2e-narrative branch from 14ea3d5 to 4057084 Compare April 27, 2026 20:49

earayu mentioned this pull request Apr 27, 2026

test(w7-task#10): grep-zero verification helper for legacy graphindex elimination #1763

Closed

2 tasks

earayu and others added 3 commits April 28, 2026 04:58

test(w7-task#11): ruff format on chenyexuan iteration commit

c79ad8e

earayu marked this pull request as ready for review April 27, 2026 21:46

earayu commented Apr 27, 2026

View reviewed changes

earayu merged commit 90b3a4a into main Apr 27, 2026
13 checks passed

earayu deleted the feat/w7-task11-e2e-narrative branch April 27, 2026 21:50

earayu mentioned this pull request Apr 28, 2026

feat(Wave 8 W8-3 #14): read-side alias resolution + flip task #11 trigger pin #1769

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

test(w7-task#11): scaffold Wave 7 e2e narrative — 9-step skeleton, bodies pending task #8 wiring#1760

test(w7-task#11): scaffold Wave 7 e2e narrative — 9-step skeleton, bodies pending task #8 wiring#1760
earayu merged 5 commits into
mainfrom
feat/w7-task11-e2e-narrative

earayu commented Apr 27, 2026 •

edited

Loading

Uh oh!

earayu left a comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

Conversation

earayu commented Apr 27, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

9-step coverage

Approach

§K.12 12-invariant cross-check

4-pattern pre-check matrix

simple-stable 4-guardrail

W8-3 trigger condition pinned in step 8

Local verification

Test plan

Uh oh!

earayu left a comment

Choose a reason for hiding this comment

Narrative-correctness 9-step verification

review feedback 落地 verification

12-invariant cross-check (narrative scope, 大部分 n/a 但 step 行为级 implicit 验证)

4-pattern pre-check matrix (PR body 已述)

simple-stable 4-guardrail

CI gate

修完会 LGTM 的清单

Wave 7 close-out narrative 物理交付意义

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

earayu commented Apr 27, 2026 •

edited

Loading