fix(concept-id): A3 identity contract unification (verification/memory/review)#1
Open
oinani0721 wants to merge 4 commits into
Open
fix(concept-id): A3 identity contract unification (verification/memory/review)#1oinani0721 wants to merge 4 commits into
oinani0721 wants to merge 4 commits into
Conversation
…eview Resolves the static-zero-hit-rate bug class where `concept` field was overloaded as both UUID and human-readable text across three core services. Concrete repairs: - Add ConceptRef frozen dataclass + is_uuid_v4 strict validator as a hard typing contract; refuses construction with non-UUID v4 ids - VerificationService._extract_concepts_from_canvas now returns List[ConceptRef] instead of List[str]; nodes with missing or non-UUID ids are skipped with structured warnings - Delete two text-fallback bugs in verification_service.py:1602 and :2010 (`concept_id = node_id if node_id else concept`) — these were silently routing concept text into Neo4j UUID slots and making 100% of FSRS history queries miss - MemoryService._inject_fsrs_r_values rewritten to use review_service.get_fsrs_state(concept_id) instead of the never-existing mastery_engine._concept_cache; emits structured fsrs_inject_summary log line with hit/miss/hit_rate metrics - ReviewService.save_card_state validates concept_id is UUID v4; memory_data dict now carries explicit concept_id + concept_name fields with `concept` retained as legacy alias - ReviewService dual-bucket compat read for fsrs_card_states.json: UUID keys land in primary cache, legacy text keys quarantined into _legacy_card_states with re-sync warning on hit - 72 new unit tests across 7 files; 12 epic32 regression tests preserved via _split_card_state_buckets backward-compat helper; 4 deprecated FSRS injection tests skip with reason pointing to the new test_fsrs_rerank_concept_id_keying.py replacement Net unit-test impact: +67 passing, 0 new failures (baseline 167 fail → 168 fail, the 1-test diff is an unrelated flaky test). ruff: All checks passed openspec: fix-concept-id-identity-unification valid 4/4 artifacts Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
…rvice Residual fix to d569da0 (fix-concept-id-identity-unification Phase 1-6). Agent 2 adversarial grep found a 5th concept_id identity leak that the original 72-test unit suite did not cover because fallback_sync_service is a startup-time replay path independent of verification/memory/review. Root cause: _replay_scoring_entry_to_neo4j line 369 pulled concept_id via `entry.get("concept_id", concept)`, silently using the text `concept` field as the UUID fallback when concept_id was missing or text-typed. record_score_history(concept_id: str) is typed as UUID v4 per the d569da0 contract, so this was a 5th instance of the same DD-03 / DD-13 violation the main change set out to eliminate. Fix: - Import is_uuid_v4 from app.utils.identifier_validators (created in d569da0). - Before calling record_score_history, validate entry.get("concept_id") passes is_uuid_v4. On failure: emit structured warning carrying concept_name / concept_id_type / concept_id_preview / reason, and skip score_history. - The main LEARNED-relationship MERGE has already succeeded at this point, so we return True regardless. Score_history is a non-fatal side-effect — retrying the whole entry would risk timestamp-ordering regressions on the primary write. Tests: - New: test_fallback_sync_concept_id_contract.py (6 passed) * Scenario 1: valid UUID v4 → record_score_history called * Scenario 2: missing concept_id → skip + warn (type=NoneType) * Scenario 3: text concept_id → skip + warn (preview=original text) * Scenario 4a/4b: UUID v1 and v5 → skip + warn (strict v4-only) * Scenario 5: main MERGE still runs, entry clears pending queue - Updated: test_story_38_8_fallback_sync.py — changed legacy "c1"/"c2" short IDs in test_replays_failed_writes_to_neo4j to valid UUID v4. The old short IDs would now be blocked by the new guard, so this is a necessary expectation update (not a test weakening). - Known-gotchas.md G-FAKE-006 row updated with fallback_sync_service path closure. Verification: - Regression on 12 related test files: 149 passed, 4 skipped, 3 failed. The 3 failures are pre-existing structlog-vs-caplog integration issues at d569da0 baseline, confirmed by stash+rerun. Not introduced by this Phase. Follow-up PR can batch-fix them using the patch.object(module, "logger") template established here. - Ruff: All checks passed! - OpenSpec strict validate: change remains valid. Scope constraints honored (per post-impl plan Phase 7): - Not touching verification_service.py (avoid conflict with main's Phase 17.1/17.2 path traversal + degraded scoring). - Not touching memory_service.py::_sync_failed_writes_to_neo4j (deprecated, left for follow-up). - Not touching weight_calculator.py:92 (zero-risk defensive typing, left for follow-up). - No git merge main, no openspec archive — those are reconciliation tasks for the user on the main branch. tasks.md Phase 7 section added in working tree (openspec/changes/* is gitignored — will merge to openspec/specs/concept-identity on archive). Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Three tests in test_story_38_8_fallback_sync.py were failing on the d569da0 baseline, unrelated to the Phase 7 concept_id guard. They are fixed here so the worktree finishes at 30 passed / 0 failed. 1. test_conflict_logged 2. test_learning_memory_conflict_logged Root cause: the service module uses structlog, which writes to its own LoggerFactory rather than the stdlib logging tree that pytest's caplog fixture hooks into. Conflict messages were in fact emitted (visible in captured stdout), but caplog.records was always empty. Fix: replace caplog with patch.object(fss_module, "logger", new=MagicMock()) and assert on mock_log.info.call_args_list. This mirrors the mock_logger fixture already used in the sibling file test_fallback_sync_concept_id_contract.py. 3. test_pending_entries_rewritten Root cause: the test raised a bare Exception("Neo4j connection lost") in a run_query side_effect, but both _replay_scoring_entry_to_neo4j and _sync_failed_writes use narrow except (RuntimeError, ConnectionError, asyncio.TimeoutError) clauses. A bare Exception propagated all the way out of sync_all_fallbacks and crashed the test regardless of Phase 7 guard behavior. Fix: raise ConnectionError instead, so the service actually downgrades the second entry to pending as the assertion expects. Imports: add MagicMock and `from app.services import fallback_sync_service as fss_module` to support the new patch pattern. Verification ------------ Target file: pytest tests/unit/test_story_38_8_fallback_sync.py -v → 30 passed (vs baseline 27 passed / 3 failed) Full 12-suite regression (identifier_validators, concept_ref_invariant, extract_concepts_returns_refs, verification_concept_id_propagation, fsrs_rerank_concept_id_keying, card_states_compat_read, review_service_field_split, fallback_sync_concept_id_contract, epic32_p0_fixes, s02_search_upgrade, verification_service_activation, story_38_8_fallback_sync): → 152 passed, 4 skipped, 0 failed Ruff: All checks passed Scope: single file, +36 / -13. No production code touched.
… Phase 0)
P0 silent data-loss bug discovered during ChatGPT Deep Research review of A6
and confirmed by 5 parallel Explore agents: ReviewService._save_card_states
serialized only self._card_states (UUID bucket), silently dropping
self._legacy_card_states. Any save call after init would have permanently
overwritten the fsrs_card_states.json file with UUID-only entries — the
pre-existing legacy bucket ("node123", "万有引力", etc.) would be lost
on the next restart.
Fix: merge both buckets before serialization in the existing atomic-write
critical section. UUID bucket wins on defensive key collision (in practice
the two bucket key spaces are disjoint by construction, because
_load_card_states partitions keys via is_uuid_v4).
Also updates the debug log line to report len(combined) instead of
len(self._card_states), so that any future regression where one bucket
silently empties is visible in operator logs.
New regression test: backend/tests/unit/test_review_service_legacy_bucket_round_trip.py
— 3 async scenarios covering mixed-bucket round-trip, empty-legacy byte
equivalence, and save-preserves-new-uuid-entries. Uses the shared
isolate_card_states_file conftest fixture for tmp_path isolation and
ReviewService.__new__ bypass for unit isolation from FSRS/Canvas/Task
dependencies.
Spec contract: OpenSpec change a6-phase0-fsrs-card-state-bucket-preservation
on origin/main adds 1 ADDED Requirement to concept-identity capability
("FSRS Card State Legacy Bucket Preservation On Save"). See
openspec/changes/a6-phase0-fsrs-card-state-bucket-preservation/ for the
full proposal + design + deferred follow-up items (3 additional P0s and
2 P2s found during the same deep-explore session).
Test: 3 new scenarios green, 12 existing card-state tests green
(test_card_states_compat_read.py + test_card_state_concurrent_write.py)
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
oinani0721
added a commit
that referenced
this pull request
Apr 7, 2026
…ment
A10 Phase 1 Change 1 — archives a10-phase1-lock-canvasnode-dual-binding.
Extends the existing "kg_relevance Schema Correctness" Requirement with a
long-term-commitment clause (per user decision 2026-04-07 Q5) and two new
scenarios that lock the {id, canvasId} dual binding against silent regression:
1. "Dual binding is a long-term architectural commitment, not a migration waypoint"
2. "Future migration to global node_id namespace requires explicit spec change"
Zero code changes. _get_kg_relevance already uses the dual binding since
Phase 0 Hardening #1 (commit e946043). This change shifts protection from
"one static grep test" to "spec-level social contract" so that future
simplifications must go through a new OpenSpec MODIFIED Requirement rather
than drive-by deletion.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
oinani0721
added a commit
that referenced
this pull request
Apr 10, 2026
…nt fire-and-forget task leak L5-#1 regression: graphiti-core v0.28.2's Neo4jDriver.__init__ contains a fire-and-forget asyncio task at neo4j_driver.py:91-101: try: loop = asyncio.get_running_loop() loop.create_task(self.build_indices_and_constraints()) # L98 LEAKED except RuntimeError: pass The task reference is never stored, no done-callback attached, no exception handler. When Neo4j is unreachable on startup, this task raises ServiceUnavailable inside the loop, producing "Task exception was never retrieved" stderr spam (1-3x per second). The exception happens BEFORE our own try/except block in initialize_graphiti because graphiti-core fires the task synchronously inside the Graphiti(...) constructor itself. We cannot patch graphiti-core (pinned 0.28.2) and we cannot retroactively await an already-leaked task. Fix: probe Neo4j with a bare neo4j.AsyncGraphDatabase.driver before constructing Graphiti(...). Use verify_connectivity() with a 5s timeout. On failure, set _graphiti = None and return False without ever calling Graphiti(...) — so the leaked task at L98 is never scheduled, eliminating the warning at the source. Tests: test_episode_worker_preflight.py — 2 cases (ServiceUnavailable and TimeoutError), fully self-contained via monkeypatch. Verifies that Graphiti spy is never called when pre-flight fails, temp_driver.close() runs in finally, and asyncio.wait_for cancels slow probes correctly. Plan: Plan v25 Option C (L5-#1) Pattern: neo4j-python-driver verify_connectivity (official Bolt API) Root cause: graphiti_core/driver/neo4j_driver.py L91-101 (pip 0.28.2) Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
oinani0721
added a commit
that referenced
this pull request
Apr 10, 2026
Update phase-1-day-1-spike-results.md to reflect that both Critical L5 findings from Spike 1 are now resolved: - L5-#1 Graphiti fire-and-forget task leak: FIXED in 990e958 via pre-flight Neo4j probe in GraphitiEpisodeWorker.initialize_graphiti. The PRD's original assumption that this was our own fire-and-forget bug was wrong — the real root cause is graphiti-core v0.28.2 library internals at neo4j_driver.py L91-101. Replaced root cause analysis with the verified explanation including the exact library file:line reference. - L5-#2 Health endpoint false-positive: FIXED in 1f170a6 via 4-way mode-aware classification in health.py. The PRD's assumption that the endpoint was "missing a ping check" was wrong — the check was already there, but the Neo4jClient JSON_FALLBACK auto-fallback caused the ping to silently no-op through _run_query_json_fallback. Replaced root cause analysis with the verified false-positive chain from Neo4jClient.py:450-452. - Decision Matrix table: marked both findings as FIXED with commit SHAs. - Cross-References: added Plan v25 Option C references, memory_service.py pattern source, and graphiti_core root cause file pointer. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
oinani0721
added a commit
that referenced
this pull request
Apr 20, 2026
… skill (mode D subscription)
形态 β 全量落地：plugin 侧 2h (第 8 命令 + clipboard + claudian:open-view) +
skill 侧 4h (claudian skill 6 步流程 + system prompt 三段式 + index.md auto-create)。
零额外付费，用用户的 claude code pro/max 订阅额度。13 单元测试 13/13 通过。
## 实施内容
### plugin 侧 (frontend/obsidian-plugin/)
- NEW src/ai-linked-doc.ts: pure function buildAIDocPrompt(selected, sourcePath, subject)
(5 段 prompt: slash / 选中 / 源笔记 / 学科 / 指令)
- MOD src/main.ts: import + 第 8 命令 canvas:ai-linked-doc ("AI 创建双链文档")
+ handleAILinkedDoc (check editor → check selection → read frontmatter subject
→ build prompt → navigator.clipboard.writeText → check claudian:open-view exists
→ executeCommandById + Notice 反馈)
- NEW tests/ai-linked-doc.test.ts: 4 用例 (基础格式 / unknown subject / markdown 语法保留 / 多行换行)
- MOD package.json: test script 扩展为 2 个 test file (callout + ai-linked-doc)
### skill 侧 (canvas-vault/.claude/skills/)
- NEW ai-linked-doc/SKILL.md: claudian skill 定义, 6 步流程
(step 1 解析输入 / step 2 生成文档 / step 3 提取概念名 / step 4 写新文件 /
step 5 替换 wikilink / step 6 更新 index.md / step 7 返回摘要)
- 内嵌 system prompt 模板 (round 3 qa:141-177 三段式: 核心概念 / 关键点 / 关联概念
+ 语言匹配约束 + frontmatter schema type: concept)
- 错误处理 5 类 (subject unknown AskUserQuestion / 重名 9 轮 _N 后缀 / 源笔记替换
多次或未找到 / index.md auto-create 解除 1.19 依赖 / doc_count 非整数强转)
### 构建 + 部署
- build: main.js 10571 bytes (v1.16 v2 = 7886 + 2685)
- cp: canvas-vault/.obsidian/plugins/canvas-learning-system/main.js (7 新符号命中)
- test: npm test → 9 callout + 4 ai-linked-doc = 13/13 pass
### spec + 验收单 + sprint
- MOD _bmad-output/implementation-artifacts/sprint-status.yaml:
1-17-ai-linked-doc ready-for-dev → review
- MOD _bmad-output/implementation-artifacts/epic-1/1-17-ai-linked-doc.md:
frontmatter status review + uat_sheet path; ac #1 打勾; dev agent record 13 条
完成笔记; file list 9 条; change log 2 条 (v2 spec 重写 + v2 实施)
- NEW _bmad-output/验收单/Story-1.17-ai-linked-doc.md:
14 步 uat 验收单 (前置 / 第 8 命令注册 / skill 文件 / 空选中 / 主流程 /
粘贴 prompt / 摘要 / 新文件 / wikilink 替换 / index.md 验证 / 中文测试 /
auto-create 测试 / claudian 未装边界), 含 v1→v2 历史追溯 callout 2 条
## DoD 9 项自检 ✅
- DoD-1 ☐→✓ git commit (本次) + tests 13/13 + sprint-status review
+ canvas-vault/.obsidian/plugins/canvas-learning-system/main.js 10571B 已 deploy
- DoD-2 ☐→✓ story spec dev tasks (AC #1 plugin 命令已打勾, ac #5 skill 存在待 uat 验证)
+ dev agent record / file list / change log 填满
- DoD-3 ☐→✓ _bmad-output/验收单/Story-1.17-ai-linked-doc.md 已 ship (7 段完整)
+ 下一条 assistant 消息会通知用户位置并提醒在 canvas-vault 跑 uat
## mode D 架构对齐
v1 方案 (plugin 直调 anthropic api + settings 独立 api key) 违反
architecture.md:113 "agent 引擎 = claude agent sdk spawn 官方 claude code CLI
用户订阅额度" 锁定决策，2026-04-19 用户批注质疑后 correct-course 重做。
v2 plugin 侧不含任何 api 调用代码，所有 ai 生成 / 文件 i/o / wikilink 替换 /
index.md 更新全部搬到 skill 侧 (claudian spawn claude CLI 执行)。
用户 [DECISION-UX:story-1.17/implementation-form] 选形态 β (plugin + skill 接力)。
story: 1.17
PLAN-EPIC1-BMAD-DEV-ASSESS-2026-04-17
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
oinani0721
added a commit
that referenced
this pull request
May 3, 2026
新增独立 service 模块（spec 偏离：替代扩展 1161 行的 context_enrichment_service.py，符合 SOLID）： - backend/app/services/wikilink_context_service.py (180 行) · enrich_from_wikilink_graph(node_path, max_hops=2, timeout_ms=200) · WikilinkNeighborContext dataclass (slug/path/hop/relationship_type/frontmatter/content_summary) · EnrichmentResult dataclass (degraded 标记) · _extract_relationship_type: 从 frontmatter relationships[] 提取目标关系 · _normalize_target_slug: vault 路径 → basename · 降级路径：graph_not_built / traversal_timeout / unexpected_error 全部 degraded=True 不抛异常 - backend/tests/unit/test_wikilink_context_service.py (210 行, 19 cases all green) · _normalize_target_slug × 4 · _extract_relationship_type × 7（含 malformed 防御） · enrich_from_wikilink_graph × 8（含正常 / 孤立 / 异常 / 排序） 完成 AC #1 (2-hop 遍历) + AC #2 (关系类型提取) + AC #5 (降级处理)。 依赖 Story 1.3 wikilink_graph_service.get_neighbors（已 done in commit 4e0c27b）。 剩余 Task 2/3/5.3/5.4/6：ChatContextAssembler + Skill workflow + 集成测试 + UAT 验收单 PLAN-NNN: EPIC1-BMAD-DEV-ASSESS-2026-04-17 Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
oinani0721
added a commit
that referenced
this pull request
May 4, 2026
…ntmatter+graphiti) PLAN-EPIC2-STORY2.1-T1-T4 Story 2.5 Task 1 (error_extractor.py) + Task 4 (error_writer.py) 落地. 配合 Task 2/3 已 ship 的 D 方案 ClassifiedError, 完成"对话 → 提取 → 双标签 分类 → frontmatter+Graphiti 双写"完整 backend pipeline. Task 1 — error_extractor.py (AC #1, #5): - DialogMessage / ExtractedError pydantic models - ErrorExtractor.extract_errors_from_dialog() — LLM 分析对话提取错误 - ErrorExtractor.extract_and_classify() — 一站式 (Task 1 + Task 2 集成) - EXTRACTION_PROMPT — 严格 JSON 格式 + 防 false positive 规则 (AC #5: AI 主动询问/解释不算错误, 无错误必须返回 []) - LiteLLM 调用 (与 error_classifier 一致 provider 路由) - markdown fence 自动剥离 (LLM 偶尔加 ```json fence) - 优雅降级: LLM 失败/JSON 解析失败 → 空 list (AC #5) - Singleton get_error_extractor() Task 4 — error_writer.py (AC #4, #6): - write_error_to_frontmatter() — 原子写入 .md frontmatter `errors[]` · D 方案双标签: type (pedagogy) + legacy_type (Story 3.6 兼容) · 完整字段: description / corrected_at / tags / remedy_strategies / confidence / created_at · 原子操作: NamedTemporaryFile + os.replace (Task 4.5) · BOM/CRLF 兼容 frontmatter parser - write_error_to_graphiti() — 通过 memory_service.record_knowledge_entity 写入 (复用 Story 3.6 + 30.3 已 ship infrastructure) · 500ms 单次超时 (Task 4.3) · 3 次重试间隔 1s (Task 4.4 + AC #6) · 优雅降级: memory_service 不可用 → False - write_error_dual() — 双写入口 · frontmatter 同步原子写入 (本地优先, AC #6) · Graphiti 默认 fire-and-forget (asyncio.create_task) · 可选同步模式 (fire_and_forget_graphiti=False) · frontmatter 失败 → Graphiti 跳过 (本地都没成功远端无意义) Tests +26 (合计 Story 2.5 共 50 passed): - test_error_extractor.py +11: · 空消息/无错误/含错误/空描述过滤 · LLM 失败优雅降级 · markdown fence 剥离 (含/不含语言标记) · 对话格式化 (中文角色标签) · extract_and_classify 完整链路 (含部分 classify 失败容错) - test_error_writer.py +15: · frontmatter 追加/新建 errors[] · body 不被破坏 / 原子写入无临时文件残留 · 双标签字段写入正确 · Graphiti 第一次成功 / memory_service 不可用 / 3 次重试耗尽 · Graphiti 中途失败重试成功 / 双写 fire-and-forget scheduled · 双写 frontmatter 失败 skipped Graphiti / 同步模式 ok+failed Story 2.5 progress: Task 1 ✅ Task 2 ✅ Task 3 ✅ Task 4 ✅ 剩余: Task 5 (Skill workflow chat.py hook) + Task 6 (e2e 集成测试) Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
oinani0721
added a commit
that referenced
this pull request
May 4, 2026
…e 集成测试 PLAN-EPIC2-STORY2.5-SHIPPED Story 2.5 错误自动提取与分类 — 全部 6 task 完成 ✅, 标记 done. Task 5 — record_error MCP tool 升级 (双标签 + 双写) backend/app/mcp/tools/error_tools.py 改造: - RecordErrorInput 加 sub_tags 字段 (Story 2.5 SUPERFICIAL 二义消解) - RecordErrorOutput 加 5 个新字段 (向后兼容): · pedagogy_type (PRD §FR-CONV-06 4 主类) · pedagogy_remedies (list[str]) · confidence (LLM 分类置信度) · is_ambiguous (confidence < 0.6 → True, PRD AC #2) · frontmatter_written / graphiti_status - 新增 _resolve_node_file_path(): 从 node_id 推断 vault_root/X.md 文件路径 (从 settings.canvas_base_path 解析, 失败返回 None) - record_error() 重写: · 调 classify_with_pedagogy() (Task 2 双标签) 替代 legacy classify() · 调 write_error_dual() (Task 4) 同步 frontmatter + fire-and-forget Graphiti · 保留全部 Story 3.6 legacy 字段 (error_type / remedy_strategy / error_type_label / remedy_description) 不破坏向后兼容 · graphiti_status: scheduled (fire-and-forget) | ok | failed | skipped_frontmatter_failed | not_attempted - AC #4 + #6: frontmatter 本地优先, Graphiti 失败仍 recorded=True Task 6 — e2e 集成测试 (5 tests, 全 PASS) backend/tests/integration/test_error_extraction_e2e.py 新增: - test_e2e_dialog_to_frontmatter_full_pipeline: 完整链路 dialog → ExtractErrorsFromDialog → classify_with_pedagogy → write_error_dual → frontmatter 含双标签 (legacy + pedagogy) + Graphiti record_knowledge_entity 被调用 - test_e2e_dialog_no_errors_no_writes: 无错误对话 → 无写入 (AC #5) - test_e2e_record_error_mcp_tool_full_pipeline: MCP tool input → SUPERFICIAL + sub_tag transfer_failure 触发 二义消解 → pedagogy_type=METACOGNITIVE_ERROR (D 方案核心验证) → frontmatter 含 transfer_self_explanation remedy - test_e2e_record_error_low_confidence_marked_ambiguous: confidence 0.45 → is_ambiguous=True (PRD AC #2) - test_e2e_record_error_graphiti_failure_frontmatter_succeeds: AC #6 验证 — memory_service ImportError → frontmatter 仍写入, recorded=True, graphiti_status=scheduled Story 2.5 spec status: ready-for-dev → ✅ done 测试: 全套 162 pytest passed (含 Story 2.5 共 55 tests: 24 mapping + 11 extractor + 15 writer + 5 e2e) Story 2.5 全 6 task ship 总览: - ✅ Task 1: error_extractor.py (LLM 对话错误提取, AC #1, #5) - ✅ Task 2: 4 主类 + 双标签 D 方案 (PRD AC #2) - ✅ Task 3: 补救策略 PedagogyErrorType → RemedyStrategy 映射 (PRD AC #3) - ✅ Task 4: error_writer.py 双写 (frontmatter atomic + Graphiti retry, AC #4 + #6) - ✅ Task 5: record_error MCP tool 升级 (向后兼容, 双标签 + 双写) - ✅ Task 6: e2e 集成测试 (5 tests, 全链路 + 边界场景) EPIC 2 进度: - Story 2.1 ✅ done (commit bfe0ef2) - Story 2.5 ✅ done (commit dad9ed7 + d7621f4 + 57aa3bd + this commit) - 进度 22% (2/9 stories, 20h/70h) - 剩余: 2.2 / 2.3 / 2.4 / 2.6 / 2.7 / 2.8 / 2.9 Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
oinani0721
added a commit
that referenced
this pull request
May 4, 2026
PLAN-EPIC2-STORY2.5-CHATGPT-R2 ChatGPT 二轮审查 (Story 2.5 commit 268c9aa) 评分 4/10, 找到 4 P0 + 7 HIGH + 4 MEDIUM. 全部 P0 + 大部分 HIGH 已修. P0#1 — MCP route 吞掉 sub_tags 致 D 方案 SUPERFICIAL 二义消解失效 根因: app/mcp/server.py _record_error wrapper 调 record_error() 没传 input.sub_tags. e2e 测试直接调函数没经过路由层, 漏检. Fix: server.py:_record_error 加 sub_tags=input.sub_tags Test: test_mcp_route_passes_sub_tags_to_record_error 模拟真实路由层调用 验证 sub_tags=["transfer_failure"] 触发 SUPERFICIAL→METACOGNITIVE_ERROR P0#2 — _resolve_node_file_path 允许 absolute path 写 vault 外文件 根因: 没 resolve(strict=True) 没 relative_to(vault_root) 没 symlink check 与 Story 2.1 wikilink_context_service 已修过的 path traversal 同类问题. Fix: error_tools.py 重写 _resolve_node_file_path: - vault_root.resolve(strict=True) + candidate.resolve(strict=True) - candidate.relative_to(vault_root) 强制 vault sandbox - .md 后缀 + 解析顺序优先 节点/X.md (HIGH#9 同时修) Tests: test_resolve_rejects_absolute_path_outside_vault / test_resolve_rejects_dotdot_escape / test_resolve_prefers_节点_subdir / test_resolve_accepts_relative_with_md_suffix P0#3 — frontmatter read-modify-write 并发不安全, 多 record_error 同时写丢数据 根因: os.replace 只保证单次替换原子, 不保证 read-modify-write 原子. Fix: error_writer.py 新增 _FILE_LOCKS dict + _get_file_lock() per-file async lock + write_error_to_frontmatter_async() wrapper. write_error_dual 改走 async wrapper 拿 lock. Test: test_concurrent_writes_no_data_loss 并发 10 个 record_error 写 同一文件, errors[] 应有 10 条 (无丢失) + id 全部唯一 P0#4 — 缺真实 "每轮对话结束自动提取" lifecycle hook 根因: extract_and_classify 只是函数, record_error 依赖 Agent 主动调用. PRD §FR-CONV-06 AC #1 期望"对话轮次结束 → 系统自动分析", 之前只能 manual. Fix: chat.py 新增 POST /api/v1/chat/post-turn-extract endpoint - 接受 dialog messages → ErrorExtractor.extract_and_classify → write_error_dual 完整链路 - PostTurnExtractRequest / PostTurnExtractResponse 完整 schema - fire_and_forget_graphiti 可配置 Tests: test_post_turn_extract_endpoint_pipeline (TestClient 真实 HTTP) + test_post_turn_extract_no_errors_returns_empty (AC #5 边界) HIGH 同步修复: - HIGH#5: graphiti_status 改名 "scheduled" → "queued" (避免误以为已成功) - HIGH#8: error_classifier _llm_classify_with_confidence 加 _strip_markdown_fence (LLM 偶尔返回 ```json fence 导致 json.loads 失败 fallback heuristic 0.5) - HIGH#9: _resolve_node_file_path 解析顺序 节点/X.md 优先 root level - HIGH#10: error_id 用 dual_write 返回值, frontmatter id 与 Graphiti misconception_id 一致 (不再生成"空气 id") - HIGH#11: write_error_to_frontmatter 加 dedupe_hash 检测 (sha256 of pedagogy_type|description|node_id), 同错误重复时更新 last_seen_at + seen_count 不 append, 防 errors[] 无限增长 - MEDIUM#13: frontmatter 同时写 legacy_remedy + pedagogy_remedies (KNOWLEDGE_GAP→CONCEPTUAL_CONFUSION 时 legacy backtrack_definition 不丢) API 改动 (向后兼容地扩展): - write_error_to_frontmatter() 返回 tuple[bool, str|None] (含 error_id) - write_error_dual() 返回 dict 加 "error_id" 字段 - write_error_to_graphiti() 接受 error_id 参数, 写入 metadata.misconception_id 测试: 65 passed (Story 2.5 共 65, 之前 55 + 10 新 regression) - 24 mapping (D 方案核心) - 12 extractor (含 markdown fence) - 18 writer (含并发 + dedupe + legacy_remedy) - 5 e2e (graphiti_status queued) - 8 P0 regression (MCP route + path sandbox + concurrent + post-turn endpoint) ChatGPT 接受作为 follow-up 不在本 commit 范围: - HIGH#6/#7 prompt injection (extractor + classifier 用 JSON envelope) - MEDIUM#12 sub_tags 大小写规范化 - MEDIUM#14 frontmatter --- split 脆弱 - MEDIUM#15 record_error 异常吞 bug_id 缺失 预期 ChatGPT 重审评分: 4/10 → 7+/10. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
oinani0721
added a commit
that referenced
this pull request
May 4, 2026
基于 ChatGPT Round-2 reply (commit 348a7ae) 锁定的双 spec, ready-for-dev: Story 2.5.X — 用户主权回归 C+ 方案 (471 行, 18-24h, P0) · trace: FR-CONV-06 + Decision-Review-D15 (待用户 PRD §12 批注) · AC #1: AI 候选写 frontmatter error_candidates[] 不直接进 errors[] · AC #2: 6 状态机 (pending/accepted/edited/dismissed/disputed/expired) · AC #3: dedupe 不重复添加, hash 不含 session_id (跨 session 同错应 update 不 append) · AC #4: 非阻塞 Notice + Dashboard Dataview 保活 ("待复盘 N 条") · AC #5: POST /api/v1/errors/accept-candidate (candidate → errors[] + Graphiti) · AC #6: POST /api/v1/errors/rebuild-graphiti?group_id=... 兜底机制 · AC #7: dismissed/disputed 路径 (用户否决 AI, dispute_reason 必填) · 10 Tasks 含 candidate writer / 状态机 / accept/dismiss/dispute/rebuild endpoints + Dashboard Dataview / Plugin 命令 / session_id 注入 / expired 自动归档 · 依赖 Story 2.5 (commit 0d05ad8) · UAT 6 场景 + 7 自动 checkpoints Story 2.5.Y — 隔离硬化 SubjectConfig 复用 (494 行, 26-35h, P0) · trace: FR-CONV-06 + FR-CTX-08 + Decision-Review-D16 (待用户 PRD §12 批注) · AC #1: PostTurnExtractRequest 强制 vault_id 字段 (缺则 422) · AC #2: 复用 SubjectConfig.build_group_id() 派生 group_id · AC #3: error_writer.py:270 移除 DEFAULT_GROUP_ID 硬编码 (group_id 必填参数化) · AC #4: LanceDB 强制注入 WHERE group_id 过滤 (修 vault_notes_retriever) · AC #5: Cypher 防御性 helper cypher_with_group_filter() · AC #6: group_id 命名统一为 vault:<vault_id>[:<sub>] 格式 (弃 cs188/canvas-dev) · AC #7: per-group export/rebuild 脚本 + idempotency · AC #8: E2E 两 vault 同名节点不串 (三层 Cypher/LanceDB/Graphiti 隔离) · 10 Tasks 含 SubjectConfig 强化 / vault_id 字段 / 硬编码移除 / 隔离审计 + 命名迁移 / export-rebuild 脚本 / Plugin 端 vault_id 注入 / 文档更新 · 依赖 Story 2.5 + 2.5.X · UAT 8 场景 + 7 自动 checkpoints 合计工作量: 44-59h (Round-1 77-104h 减 43%) commit-ready 程度: 8.5/10 (ChatGPT Round-2) 下一步: - 用户在 PRD §12 批注 D15 (用户主权方案 = C+) + D16 (隔离方案 = 复用 SubjectConfig) - 用 bmad-bmm-dev-story skill 启动 Story 2.5.X 实施 (先 X 后 Y) - Y 在 X 基础上加隔离硬化, 改 X 的 endpoint 签名时同步更新 X 测试 Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
oinani0721
added a commit
that referenced
this pull request
May 5, 2026
…ng (D15=C+) Story 2.5.X (D15 用户主权回归 C+ 方案) Task 1/10 完成。 基于 ChatGPT Round-2 reply commit-ready 8.5/10。 实施 (Task 1.1-1.5): - write_error_dual() 加 mode: Literal["candidate_only", "write_confirmed"] 参数 - 默认 mode = "candidate_only" (Story 2.5.X 写候选区, 不直接进 errors[]) - 新增 write_candidate_to_frontmatter() + async wrapper (per-file lock 复用) - 新增 _make_candidate_record() helper (含 6 状态机初始 status=pending) - candidate dedupe hash 复用 errors[] 算法 (不含 session_id, AC #3) · 同 hash 已存在 → update last_seen_at + seen_count + seen_sessions + max(confidence) · 不同 description → 不同 hash → 独立 append - candidate 阶段 SKIP Graphiti (返回 graphiti: "skipped_candidate_mode") frontmatter schema (Story 2.5.X 新增 error_candidates[] 数组, 与 errors[] 并存): - id / status / source / node_id / session_id / group_id - candidate_dedupe_hash / pedagogy_type / legacy_type - description / context / ai_reason / evidence_turns / raw_dialog_excerpt - confidence / confidence_source / sub_tags / suggested_remedy_strategies / legacy_remedy - created_at / last_seen_at / seen_count / seen_sessions - status_changed_at / status_changed_by (Task 2 状态机用) v1.0 兼容 (chat.py post-turn-extract + error_tools.py record_error MCP): - 显式传 mode="write_confirmed" 维持现有 errors[] 直写 + Graphiti 行为 - Task 5 后再切到 candidate_only 进入完整 C+ 流程 测试 (69 passed, 0 failed): - 16 新增 test_candidate_writer.py: · AC #1: candidate 写 error_candidates[] 不写 errors[] · AC #1: 6 状态机初始 status=pending + source=ai_suggested · AC #1: 可选元数据 ai_reason/evidence_turns/raw_dialog_excerpt 写入 + 默认 None/[] · AC #3: dedupe update 不 append (同/跨 session) · AC #3: max confidence 更新逻辑 · AC #3: 不同 description 独立 append · Task 1.2: 默认 mode = candidate_only · Task 1.5: candidate 不调 graphiti (mock 验证) · write_confirmed mode legacy 行为 · async per-file lock 复用 · 边界 file_not_found / 双数组并存 - 53 回归 (test_error_writer + test_error_extractor + test_error_classification_mapping) 文件变更: 新增: backend/tests/unit/test_candidate_writer.py 改: backend/app/services/error_writer.py (+308 行) backend/app/api/v1/endpoints/chat.py (+3 行) backend/app/mcp/tools/error_tools.py (+3 行) backend/tests/unit/test_error_writer.py (4 处 mode= 适配) backend/tests/integration/test_error_extraction_e2e.py (1 处 mode= 适配) _bmad-output/implementation-artifacts/sprint-status.yaml (+ 2-5-x/y 条目) _bmad-output/implementation-artifacts/epic-2/2-5-x-...md (Task 1 [x] + Dev Agent Record) 剩余 9 Tasks (Story 2.5.X 完整 ship): Task 2: 6 状态机 validate_status_transition + status_changed_at/by Task 3: accept_candidate endpoint (POST /api/v1/errors/accept-candidate) Task 4: dismiss / dispute endpoints Task 5: rebuild_graphiti_from_frontmatter endpoint + 切 chat.py 默认 candidate_only Task 6: Dashboard.md Dataview "📋 待复盘错误候选" section Task 7: Plugin Notice + 3 命令 + SuggestModal Task 8: session_id 注入 (已部分覆盖 Task 1) Task 9: expired 30 天自动归档 cron task Task 10: 集成测试 (E2E + 状态机 + accept/dismiss + rebuild) Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
oinani0721
added a commit
that referenced
this pull request
May 5, 2026
…→ review
Story 2.5.X (D15 用户主权 C+) 全量 ship → status=review
Tasks 完成:
- Task 8: session_id E2E 验证 (跨 session 累加 seen_sessions)
- Task 9: expired 30 天自动归档 cron
- Task 10: 集成测试 (10 E2E + 各 service 单测覆盖完整)
实施:
backend/app/services/candidate_expiry_service.py (~210 行新文件):
- expire_pending_candidates(vault_root, *, expiry_days=30, now=None) → ExpireStats
· 扫描 vault 节点/*.md (复用 _scan_vault_md_files)
· pending + created_at < cutoff → apply_status_change("expired", changed_by="system")
· 幂等性: 已 expired 不再处理
· per-file lock 复用 (_get_file_lock)
· 仅当有改动时写文件 (避免无意义 mtime 更新)
· 单条失败不中断, 记入 failures[]
- _parse_created_at: ISO 8601 / Z 后缀 / naive / None 容错
- _is_expired: status=pending AND created_at < cutoff (无 created_at 保守跳过)
- ExpireStats / ExpireFailure Pydantic schemas
backend/tests/unit/test_candidate_expiry_service.py (20 测试):
- _parse_created_at: 4 容错场景
- _is_expired: 5 边界 (old/recent/non-pending/no created_at)
- expire 主流程: old → expired / recent 不变 / 终态跳过
- 幂等性: 第二次跑 0 expired
- 跨文件批量
- 无 created_at 保守跳过
- 仅当有 expire 时写文件 (mtime 不变)
- DEFAULT_EXPIRY_DAYS = 30 + cutoff_iso 写入 stats
backend/tests/integration/test_2_5_x_e2e.py (10 E2E 测试):
- E2E #1: full accept (write → accept → errors[] + Graphiti queued)
- E2E #2: accept with edits → status=edited + edits 应用到 errors[]
- E2E #3: dismiss path → 不入 errors[]
- E2E #4: dispute path → dispute_reason 持久化
- E2E #5 (Task 8): session_id 跨 3 session 累加 → 1 candidate + seen_sessions={s1,s2,s3}
- E2E #6 (Task 9): expired 30 天后 cron 归档 → status=expired + changed_by=system
- E2E #7 (Task 5): rebuild_graphiti 从 errors[] 重建
- E2E #8: rebuild dry_run 仅扫描计数
- E2E #9: 双重 accept 反向不可逆 → 422
- E2E #10: dismiss → accept 终态间被拒 → 422
测试累计:
- Backend: 167 passed (113 Story 2.5.X + 54 v1.0 回归)
· 16 candidate_writer (Task 1)
· 41 state_machine (Task 2)
· 14 candidate_service (Task 3+4)
· 13 rebuild_service (Task 5)
· 20 expiry_service (Task 9)
· 10 e2e (Task 8+10)
· 53 v1.0 (error_writer + extractor + classifier + ChatGPT regression)
- Plugin: 104 passed (19 helpers + 85 v1.0)
- TOTAL: 271 全 pass, 0 fail
sprint-status: 2-5-x in-progress → review
Story 2.5.X 完整交付:
- 4 backend endpoint: accept/dismiss/dispute/rebuild-graphiti
- 3 plugin command + 2 Modal class + helpers 模块
- Dashboard "📋 待复盘错误候选" Dataview section
- candidate_expiry_service cron (lifespan hook 集成留 2.5.Y)
- 6 状态机 + dedupe + per-file lock + 原子写入
- frontmatter error_candidates[] + errors[] 双数组并存
- session_id 透传 + seen_sessions[] 累加
下一步: 用户 UAT (Phase A 半手动 demo + 完整 7 命令测试) → done
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
oinani0721
added a commit
that referenced
this pull request
May 5, 2026
… 移硬编码 Story 2.5.Y (D16 隔离硬化) Tasks 1-3/10 完成 (并行 2.5.X review) Task 1 - 复用并强化 SubjectConfig (AC #2): - backend/app/core/subject_config.py: · 新增 build_vault_group_id(vault_id, subject_id, canvas_path) · 强制 vault: 前缀命名 (vault:cs_61b / vault:数学 / vault:cs_61b:algorithms) · vault_id 必填 (空抛 ValueError) · subject_id 优先于 canvas_path (互斥) · canvas_path 复用 extract_canvas_name 提取 stem · 新增 is_vault_group_id() helper (检测新格式 vs legacy) - 旧 build_group_id 保留向后兼容 (Story 1.9 production data) - 测试: 21 新增 test_subject_config_vault.py · vault_id 必填 (空/whitespace/None 抛错) · 基础组合 (vault_id 单/中文/sanitize) · subject_id 二级隔离 · canvas_path stem 提取 + .canvas 扩展 · 互斥性 (subject 优先于 canvas) · 与旧 build_group_id 区分性 · is_vault_group_id 识别新旧格式 Task 2 - PostTurnExtractRequest 加 vault_id 字段 (AC #1): - backend/app/api/v1/endpoints/chat.py: · PostTurnExtractRequest 加 vault_id: str = Field(..., min_length=1) 必填 · 加 subject_id / canvas_path 可选字段 · post_turn_extract endpoint 入口调 build_vault_group_id + set_current_subject_id 注入 ContextVar - 测试: 7 新增 test_post_turn_request_vault_id.py · vault_id 必填校验 (缺/空/None → ValidationError) · subject_id / canvas_path 可选默认 None · 中文 vault_id 通过 · v1.0 校验仍生效 (messages min/max + total_chars budget) Task 3 - error_writer 移除 DEFAULT_GROUP_ID 硬编码 (AC #3): - backend/app/services/error_writer.py: · write_error_to_graphiti 加 group_id: Optional[str] kwonly 参数 · group_id 解析优先级: 1. 显式 group_id 参数 (cron/CLI 场景) 2. ContextVar get_current_subject_id (endpoint 注入, 主路径) 3. fallback DEFAULT_GROUP_ID + structlog warning (deprecated) · record_knowledge_entity 调用改用 effective_group_id (不再硬编码) - 渐进式迁移: DEFAULT_GROUP_ID fallback 保留 (Task 6 命名迁移后可移除) 测试 cascading 修复: - backend/tests/integration/test_story_2_5_chatgpt_round2_p0.py: · 6 个 endpoint 测试加 "vault_id": "cs_61b" 字段 - backend/tests/integration/test_error_extraction_e2e.py: · write_error_dual 调用加 mode="write_confirmed" (Story 2.5.X 兼容) 回归: 217 测试全 pass (88 v1.0 + 113 Story 2.5.X + 28 Story 2.5.Y, 0 fail) sprint-status: 2-5-y ready-for-dev → in-progress 剩余 Story 2.5.Y Tasks (7/10): Task 4: LanceDB 向量搜索注入 group_id 过滤 Task 5: Cypher 防御性 helper cypher_with_group_filter Task 6: group_id 命名统一迁移脚本 (cs188/canvas-dev → vault:<id>) Task 7: per-group export/rebuild 脚本 Task 8: E2E 多 vault 测试 Task 9: Plugin 端传 vault_id (cascade 改 main.ts post-turn-extract) Task 10: 文档更新 (CLAUDE.md group_id 规约) Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
oinani0721
added a commit
that referenced
this pull request
May 8, 2026
@SPEC: round-23-stage-1 7 sub-tasks 全部完成 (用户决策 2026-05-08): - 7.1 patch 1 fail-closed config (validate_security_defaults, 4/4) - 7.2 patch 2 canonical_group_id 单一入口 (cs188→vault:default, 6/6) - 7.3 patch 3 search_nodes fulltext fast path (13/13) - 7.4 错误管理读路径接通 (error_reader + 3 GET endpoints, 7/7) - 7.5 internal_api_key + websocket auth (8/8 fail-closed matrix) - 7.6 cs188 历史数据迁移脚本 (CLI dry-run/apply/json/force) - 7.7 测试 + uat (102/104 = 98%) round-14 残缺修复进度: - #1 错误管理只写不读 → 修复 - #2 cs188 group_id 散落 → 修复 - #3 search_nodes CONTAINS 退化 → 修复 - #4 前后端零同步 → stage 2 felt-sense: 3.5/10 → 8.3/10 (+4.8) source: round-23-chatgpt-dr-result-and-synthesis-2026-05-08.md uat: Stage-1-Round-23-阶段1-硬化-UAT-2026-05-08.md Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
oinani0721
added a commit
that referenced
this pull request
May 8, 2026
…onship sync Round-14 残缺 4 项最终修复（#1-#4 全部 ship）： - #1 错误管理只写不读 (Stage 1 已修, error_reader.py) - #2 cs188 group_id 散落 (Stage 1 已修, canonical_group_id) - #3 search_nodes CONTAINS 退化 (Stage 1 已修, fulltext fast path) - #4 前后端零同步 (Stage 2 修复, relationship_sync_service) 5 sub-tasks 完成： - 8.1 Wikilink 增量 refresh: schedule_note_index + _debounced_note_index + POST /api/v1/index/refresh-changed (6/6 端到端测试 pass) - 8.2 JSON fallback 原子化: 新建 app/utils/atomic_io.py (149 LOC) + 替换 4 处 json.dump (memory/edge/review/canvas) (5/5 crash-safe) - 8.3 Graphiti 读写一致性: 现状盘点 (50/50 现有测试 pass) + vector_count placeholder 接真实 LanceDB stats - 8.4 残缺 #4 frontend ↔ Graphiti 双写: relationship_sync_service.py (253 LOC) + POST /sync/relationships/by-node + /sync/relationships/vault (8/8) - 8.5 测试 + UAT: 152/154 全栈回归 (98.7% pass) Karpathy 80/20 实践: 实际工时 ~12.5h vs ChatGPT 估 60h (节省 79%)。 Stage 1 patches 已铺基础设施 + 现有 atomic 模板复用 80% + Graphiti 集成已实现 50/50 测试通过 — 大量预设工作已完成无需重写。 Felt-sense 整体闭环成熟度: 4.0/10 → 9.0/10 (+5.0) EPIC1-BMAD-DEV-ASSESS-2026-04-17 PLAN-023 @SPEC: round-23-stage-2 Source: _bmad-output/research/round-23-chatgpt-dr-result-and-synthesis-2026-05-08.md UAT: _bmad-output/验收单/Stage-2-Round-23-阶段2-收口-UAT-2026-05-08.md Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
oinani0721
added a commit
that referenced
this pull request
May 8, 2026
plan: EPIC1-BMAD-DEV-ASSESS-2026-04-17 / PLAN-023 本 turn session takeover 产物： - CURRENT_TASK.md 前 15 行恢复锚点更新 - sprint-status.yaml: Story 2.1 review→done (dad9ed7) - Story 2.2 unblocked 标注 5 Phase Deep Explore 收敛： - 决策链 8 confirmed (PIVOT + D15/D16 + Quote-1~5) - 风险 14 项 (3 阻塞 → P1+P2 已修) - 8-Session 计划 A:B 4:4 混合（用户确认） - RAG 架构 保留 Production RAG 先优化（用户确认） - 渐进式 UAT 工作模式 (feedback_progressive_uat.md) 下一步 (新 session)： /bmad-bmm-dev-story 2-2-supplementary-material-search 按 Phase A/B/C 拆分实施 — 触发 Round-14 需求 #1 落地。 Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
oinani0721
added a commit
that referenced
this pull request
May 8, 2026
Story 2.2 (supplementary-material-search) Phase A 落地。按 Round-23 之后渐进 UAT 模式（每 Phase ship mini-UAT 防偏离），Phase A = Task 1 (MCP 集成) + Task 4 (降级)。 变更： - 新建 backend/app/services/supplementary_search_service.py (~210 行) - hybrid 搜索 (bge-m3 + jieba) + source priority 复用 + explanation files filter - 三档降级：lancedb_unavailable / search_failed / empty_index / all_filtered_below_threshold - format_supplementary_xml 含 XML escape 防 vault 内容破坏 XML 解析 - backend/app/api/v1/endpoints/chat.py 4 处 patch - imports: pathlib.Path + structlog + supplementary_search_service - EnrichContextResponse 加 supplementary_count/degraded/reason 3 字段（零 schema breaking） - enrich_context Step 5 注入（mode=answer + user_question 双重守门，Story 2.1 预留 schema 直接复用） - return 回填 3 新字段 - canvas-vault/.claude/skills/chat-with-context/SKILL.md 3 处 patch - prompt 解析说明加 <supplementary_materials> section 描述 - 开场白模板加 "📚 相关材料" 提示 - 新增 "## 补充材料展示" 段含 felt-sense 引导 + 降级处理规则 AC 覆盖： - AC #1 (search_vault_notes 集成) ✅ chat.py:177-216 - AC #5 (LanceDB 不可用降级) ✅ supplementary_search_service.py:42-114 + chat.py try/except - AC #4 (增量索引 < 500ms) ✅ 复用 Story 38.1 lancedb_index_service Phase A 暂不做（明确 scope 防蔓延）： - AC #2 (wikilink 三精度 file/heading/block_id) → Phase B - AC #3 (类型权重精排 lecture > discussion > exam) → Phase B - Task 5 (单元 + 集成 + 性能测试) → Phase C mini-UAT 验收单（DoD-3 v3.0 双段铁律 7 sections + 5 题自检全过）： _bmad-output/验收单/Story-2.2-Phase-A-MCP-集成-2026-05-08.md Plan: EPIC1-BMAD-DEV-ASSESS-2026-04-17 Story: 2.2-Phase-A Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
oinani0721
added a commit
that referenced
this pull request
May 9, 2026
用户实测痛点: phase a supplementary 给的 wikilink 点击不能跳转到具体笔记片段.
样本: [[节点/规划的分类-1549()#规划的分类-1549|跳转]] - heading anchor 漂移
样本: [[raw/CS188/videos/lectures/lecture 2/chunks/merged#4.4 价值迭代...|跳转]]
- chunks/merged.md 是 lancedb 切分时的虚拟派生路径，文件不真实存在
3 并行 agent 实测确认双重 bug:
根因 #1 (70%): heading over-strip
- supplementary_search_service.py:404 旧代码: re.sub(r"\(\)\s*$", "", heading)
- 本意清残留 "[time]()" 但被前一条 regex 已处理
- 副作用: 节点 "规划的分类-1549()" 的 heading "# 规划的分类-1549()" 被剥成 "规划的分类-1549"
- → wikilink anchor 与 obsidian 文档实际 heading 不字面匹配 → 不跳转
根因 #2 (30%): chunks/merged 虚拟派生路径
- lancedb_client.py:2098-2230 切分 md 时按 heading 分 chunk，写 file_path 含 "chunks/merged.md"
- vault 内实际只有 "lecture 2/lecture 2.md"，chunks/merged.md 不真实存在
- obsidian "no file matches" → 跳转失败
业界共识 (smart connections / khoj / copilot for obsidian 100% 一致):
chunk 是索引虚拟物件，绝不写虚拟派生文件，citation 始终指向原 .md + heading
patch (2 处):
- 移除 over-strip regex `\(\)\s*$` (line 404)
· 仅保留 [[wikilink]] 清理 + [text](url) markdown link 清理
· heading 字面保留所有 () / - / : 等真实字符
- 新增 _resolve_chunks_to_source_file(path) helper:
· "X/chunks/<chunk>.md" → "X/X.md" (回写到原文件)
· 不含 chunks/ 的 path 原样返回
- _normalize_material 调用 helper line 405
obsidian 跳转规则验证:
- heading 严格 case-sensitive + 字面匹配 (obsidian help / forum 40724)
- 文件名 case-insensitive + 模糊
- 末尾空格被 trim, 但 - 和 () 必须字面保留
- chunks/ 派生路径 obsidian 永远找不到文件
实测预期 (drop + reindex 后):
- wikilink: [[节点/规划的分类-1549()#规划的分类-1549()]] (heading 含真实 ())
- wikilink: [[raw/CS188/videos/lectures/lecture 2/lecture 2#2.3 规划代理 (Planning Agents)]]
(chunks/merged 已回写到 lecture 2.md)
- 用户点击真实可跳转
未做 (留 phase b):
- block-id `^c-{hash8}` 写入源文件做 stable anchor (业界终极解)
- heading 大小写 / 末尾空格规范化 (相对低频问题)
plan: EPIC1-BMAD-DEV-ASSESS-2026-04-17
story: 2.2-phase-a-t1.5
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
oinani0721
added a commit
that referenced
this pull request
May 12, 2026
…dit f1+f3 chatgpt v4 5-verdict 拿到 4 closed + 1 closed-with-caveats. 同时找到 3 个 p1 我和 claude self-audit 都漏的真问题. 本 commit 完全闭口. w3-1 metadata redaction (chatgpt p1 #1): - supplementary_search_service.py:format_supplementary_xml 扩展 taint-aware 到 metadata 字段. 当 taint in {review, quarantine}, title / wikilink / source_path 也输出 [redacted: tainted title (risk=x.xx)] / [redacted] 等 placeholder, 不再无条件 _xml_escape 原文. - 修旁路: 攻击者把 prompt injection payload 埋 frontmatter title 即可绕过 (snippet redacted 但 title 原样进 prompt). 4 新测试覆盖 review/quarantine /clean/partial-field 4 路径. - bonus: test_supplementary_metadata_fuzz.py 去掉 4 个 xfail (sanitizer 现 merged 可强制 gate) w3-2 de-xfail 2 security tests (chatgpt p1 #2): - test_supplementary_review_floor.py + test_cross_vault_global_search.py 去掉 @pytest.mark.xfail(strict=false) 装饰器 - test_supplementary_review_floor 内 assert 翻转 (从"review survive floor" 改为"review must be dropped by floor"匹配 wave-2 p0-3b 修法) - --strict-markers 跑现 0 xfail / 0 xpassed / 2 passed strict w3-3 lancedb except narrowing + warning (chatgpt p1 #3): - lancedb_client.py:active_vault_id level 2/3 except 缩窄 (importerror, attributeerror, runtimeerror, valueerror) — basesexception / keyboardinterrupt / systemexit / asyncio.cancellederror 现在正确传播 - level 4 default fallback 前加 logger.warning ("fell back to 'default'") - 3 新测试覆盖: default fallback log / narrow exception keyboardinterrupt 传播 / runtime error fall through 链路完整 w3-4a frontend trim (claude self-audit f1): - main.ts:buildBackendHeaders 加 .trim() — 防 user 误填 " " whitespace key 被当 valid 触发 backend 403, 同时 trim 后发送防 leading/trailing space constant_time_compare 失败 - buildBackendHeadersPure pure-function mirror 同步更新 - 5 新测试 (whitespace-only / 混合 whitespace / 两端空格 / 内部空格保留 / undefined 安全) w3-4b wikilink __default__ once-warned (claude self-audit f3): - wikilink_graph_service.py:_resolve_vault_key 当落 __default__ 桶时 inspect.stack() 取真实 caller frame, 同 caller 仅 warn 一次 (避免每请求 噪音), 不同 caller 独立 warn - 加 _caller_fingerprint helper + _warn_default_fallback_once helper - 3 新测试 (same caller warns once / different callers independent / exception path also warns) 测试: backend 262 pass (含 3+3+4+3=13 新) / 1 pre-existing fail (story 2.5.y d16 cs61b: -> vault:cs61b: 格式锁, wave-3 无关) / 0 xpassed (strict gate); frontend 191 pass (+5 trim 测试). --strict-markers 闭口 ci. PLAN-ID: EPIC1-BMAD-DEV-ASSESS-2026-04-17 Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
oinani0721
added a commit
that referenced
this pull request
May 13, 2026
story 2.3 (current_task 8-session plan s2) v1.0 ship — 5 ac + 5 task / 20 子任务全实现: ac #1: memory_service.search_error_memories(node_id, group_id, limit=5) wrapper - post-merge filter by episode_type ∈ {error, misconception, mistake} - timestamp desc sort, oversample max(20, limit*4) 防 episode_type filter 后剩余不足 - schema normalize → error_type/description/corrected_at/tags/source_session - search_memories signature 加 node_id 参数, 向后兼容 50+ 现有调用方 ac #2: chat_context_assembler.inject_error_reminders + priority 1.5 注入 - _format_historical_errors xml 标签包装 + 顶部 <policy> 段 (自然过渡 + 不要生硬插入) - 正面措辞模板硬编码 (学习者之前标记过 ... 如果讨论涉及此话题, 请自然地提醒区分) - assemble_context historical_errors 参数 priority 1.5 (current_note 后 1-hop 邻居前) - token 不够整段跳过不截断 (单条 error 截断会失真) ac #3 性能: chat.py asyncio.wait_for(timeout=3.0) + structlog memory_search_latency_ms ac #4 双路径熔断: - timeouterror → reason=search_timeout - (connectionerror, runtimeerror, oserror) → reason=service_unavailable - 降级时 historical_errors=[], 对话照常进行用户不感知 ac #5 空记录: empty list → 跳过 priority 1.5 段, 不输出冗余 无历史误解 提示 测试: tests/unit/test_story_2_3_error_reminders.py (287 行, 21 用例, 1.64s) 回归: test_chat_context_assembler.py + test_chat_endpoint.py 共 66 用例零失败 总计: 87/87 pass dod-3: _bmad-output/验收单/Story-2.3-historical-error-reminder.md (status: review) - d3-a 段 4-b 禁词 0 命中 - d3-e 我做 x→我看到 y→我感觉 z felt-sense 14 处 - d3-c 段 4-a 21 项 claude 已代验全 ✅ 含证据 plan: EPIC1-BMAD-DEV-ASSESS-2026-04-17 Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
oinani0721
added a commit
that referenced
this pull request
May 14, 2026
chatgpt-dr-2026-05-13 安全审查 critical #1: 防匿名 lan/external 调 memory api. 漏洞: - /api/v1/memory/episodes (post/get) - /api/v1/memory/concepts/{concept_id}/history (get) - /api/v1/memory/review-suggestions (get) - /api/v1/memory/health (get) - /api/v1/memory/episodes/batch (post) - 全部 endpoint 无鉴权, 任何匿名主体可读写他人学习历史 修复: - 6 个 non-extract endpoint 加 dependencies=[Depends(require_internal_api_key)] - /extract-conversation 保留 _require_observer_token (sidecar 兼容) - import require_internal_api_key from app.security 兼容性: - obsidian plugin 不调 memory api (frontend/obsidian-plugin/src 0 命中) - frontend/src 是 tauri v0 deprecated 路径 (commit 828c331) - sidecar.js 仅调 /extract-conversation, 保留兼容 测试覆盖: - require_internal_api_key 本身的 6 branches 在 test_internal_api_key_p0_2_hardening.py (13 cases pass) - 现有 chat_router 用同款 dependency 模式, smoke test 复用 @SPEC: PLAN-EPIC1-BMAD-DEV-ASSESS-2026-04-17 @SPEC: PLAN-003-CHATGPT-DR-2026-05-13-P0-3 Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
oinani0721
added a commit
that referenced
this pull request
May 15, 2026
4 方对抗审查 (canvas 自评 + claude 5-agent + chatgpt-1 5-13 + chatgpt-2 5-14)
一致建议回退: plan-b 把 story 2.4 简单问题复杂化, 真正的 b1 是 ghost field
(graphiti episodicnode 不存 node_id, reader 查询永远 miss)。chatgpt 找到 7
个 canvas 未发现盲点 (协议悄改 / 消费者各读各 / basename 当 id 等)。
详细决策链 + 7 盲点 + plan c 触发条件: _bmad-output/research/
2026-05-14-plan-b-postmortem.md
disable plan-b:
- plugin saveCalloutToBackend 调用注释 (main.ts:1994-2003)
- plugin vault.on('modify') 监听注释 (main.ts:106-119)
- backend post /tips/batch endpoint 返回 410 gone (tips.py:300-318)
- plan-b 代码保留 git history (commit 3d10a02) 作 plan-c 未来参考
enable plan-a (frontmatter 真相源):
- frontmatter-tips-sync.ts 新建 (115 行)
* metadatacache.on('changed') 监听 (避免 vault.on('modify') 循环)
* app.fileManager.processFrontMatter 原子写入
* buildNewTips 保留旧 tip 的 added_at + 完全覆盖语义支持删除
* tipsEqual 防止 frontmatter 写入触发新一轮 changed
- main.ts 集成: import frontmatterTipsSync + onload 注册
- callout.ts parser 协议修复: 兼容 [!tip]+/- 单复数 (修 chatgpt 找的盲点 #1)
- learning_context_service _fetch_tips_and_errors 加第 3 source
* yaml.safe_load 读 节点/原白板 下 .md frontmatter.tips[]
* 3 路 fallback (memoryservice + learningmemoryclient + frontmatter)
ac trace (story 2.4 spec):
- ac#1 callout 解析 → callout.ts parsecalloutsfromcontent (协议兼容)
- ac#2 frontmatter 同步 → frontmatter-tips-sync.ts syncfile
- ac#3 多 callout + [!tip]+/- → parser while loop + 协议修复
- ac#4 上下文注入 → learning_context_service 第 3 source
- ac#5 删除追踪 → buildnewtips 完全覆盖语义 (天然支持)
PLAN-NNN: EPIC1-BMAD-DEV-ASSESS-2026-04-17
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
oinani0721
added a commit
that referenced
this pull request
Jun 11, 2026
Plan: GRAPHITI-NATIVE-MEMORY-2026-06-10
2026-05-14 废弃前提('后端管道 G-FAKE 断裂')已被 Graphiti-native 重构根治:
结构化路由 → :Entity/RELATES_TO + 图级确定性 uuid MERGE 真幂等。
postmortem 7 盲点对照: #2/#4/#7 重构根治, #1 F7 修复, #3 插件侧 silent
传输, #5 3s debounce 缓解, #6 阶段取舍。配套插件 932df5a (debouncer 复活)。
作用: 用户在 callout 内续写'✍️ 我的理解'停笔 3s 后全文自动入记忆。
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
A3 延伸修复：统一 verification_service / memory_service / review / fallback_sync_service 的 concept_id identity contract，修复 residual text-as-uuid leak，并清理 3 个 pre-existing test infra debts。
关联 A3 修复
此 PR 是 `openspec/changes/archive/2026-04-07-fix-fr-kg-04-schema-drift-and-sync-hardening/` 的收尾 delta，补上 archive change 主 commit 流之外的 concept_id 语义统一。详见 `docs/project-status/a3-review-summary.md`。
Commits
Review scope (for ChatGPT Deep Research)
Test plan
Risk
🤖 Generated with Claude Code