Identical-content files at different workspace paths share one assets row
(assets.asset_id = blake3 content hash, PRIMARY KEY). The UPSERT
`ON CONFLICT(asset_id) DO UPDATE SET workspace_path = excluded` made
twin files overwrite each other's workspace_path on every ingest, so
`get_asset_by_workspace_path(path1)` returned the OTHER twin's row (or
None) — break idempotent unchanged-detection for both files.
Fix: switch try_skip_unchanged to document-centric lookup. `documents.
workspace_path` is already UNIQUE (V001) and `id_for_doc(path, ...)`
includes path, so each twin has its own stable document row. Compare
`doc.source_asset_id` with the new asset's checksum instead of going
through the assets table.
Dogfood (multi-root: kebab-docs + httpx + zod + lodash) showed 27 of
726 docs marked Updated on every idempotent re-ingest — all 27 are
twin-file victims (empty `__init__.py` ×3, AGENTS.md ↔ CLAUDE.md
same content, duplicate logo PDFs/JPGs).
After: re-ingest reports 0 new / 0 updated / 726 unchanged.
No schema migration needed.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Dogfooding (PR #142 1B + multi-root corpus: kebab-docs + httpx + zod + lodash)
revealed schema.v1.repo_breakdown is always {} despite the 1A-2 Task 9
having added the code_lang_breakdown sibling. The schema.rs:171 placeholder
`BTreeMap::new()` was left in place. Mirror Task 9's code_lang_breakdown
query for the repo field — same metadata_json JSON-path pattern.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
p10-1A-1 (PR #139) added SearchFilters.code_lang + .repo fields and the CLI
--code-lang / --repo flags propagate them correctly into SearchFilters, but
neither the lexical retriever's FTS SQL nor the shared filter_chunks helper
(used by the vector retriever) ever applied them — so a code-lang-filtered
search returned all-doc hits (markdown / pdf / code mixed).
Discovered while dogfooding p10-1B with httpx + zod + lodash clones:
`kebab search 'AsyncClient' --code-lang python --json` returned markdown
hits from httpx/docs/ first.
Fix: add IN-list filters on json_extract(d.metadata_json, '$.code_lang')
and '$.repo' to both lexical.rs and filters.rs, mirroring the existing
media filter pattern. Two regression tests added in each crate covering
the new filter behavior.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Add `SqliteStore::code_lang_breakdown()` that queries
`json_extract(metadata_json, '$.code_lang')`, groups by it, and
skips NULL rows — returning `BTreeMap<String, u32>`.
Wire it into `collect_stats` in `kebab-app::schema`, replacing the
`BTreeMap::new()` placeholder inserted by 1A-1.
Test: `store::tests::code_lang_breakdown_counts_by_code_lang` asserts
rust=1 and that a null-code_lang doc does NOT appear in the map.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Four optional, serde-skipped-when-None fields added to `Metadata` for
code ingest context. All 11 downstream construction sites patched with
`repo: None, git_branch: None, git_commit: None, code_lang: None`.
Full workspace check (`--tests`) and per-crate test suite pass clean.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Adds five new u32 counters (skipped_gitignore, skipped_kebabignore,
skipped_builtin_blacklist, skipped_generated, skipped_size_exceeded)
and a SkipExamples struct (≤5 sample paths per category) to
IngestReport. All new fields are #[serde(default)] for backward-compat
deserialization. Downstream literal construction sites patched with
zeros/empty; snapshot re-baked.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Extends CountSummary with media_breakdown, lang_breakdown, stale_doc_count
fields populated via stats_ext::breakdowns(). Adds count_summary_with_threshold
for callers that need real stale counts. Mirrors all new fields onto the
wire-bound Stats struct in kebab-app::schema with #[serde(default)] for
backwards-compat. Also fixes search_budget_integration.rs for the trace field
added to SearchOpts in Task 1.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- ingested_after: convert OffsetDateTime to UTC before formatting
so non-Z offsets compare correctly against UTC TEXT storage
(lexical.rs + filters.rs)
- README: --tag is repeatable-only, not csv (only --media is csv)
- test(cli): add multi-value --tag OR-within IN-list coverage
- test(store): add UTC-offset regression test for ingested_after
- mcp: use ERROR_V1_ID const instead of hardcoded "error.v1"
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
filter_chunks helper in kebab-store-sqlite extended with the same 3
WHERE clauses as lexical. Vector still over-fetches k*2 then
post-filters via SqliteStore::filter_chunks; small k can return < k
hits when filters drop a lot — agent is expected to widen k or
paginate. AND combinator with existing filters.
- kebab-store-sqlite/src/filters.rs: media IN-list subquery, ingested_after
lexicographic >= compare, doc_id equality; mirrors lexical SQL arms
- 3 direct unit tests (filter_chunks_media_type/ingested_after/doc_id)
that run without AVX/Lance
- common/mod.rs: insert_doc / insert_doc_with_media / run_vector_search
helpers on HybridEnv for integration-test use
- hybrid.rs: 2 new #[ignore = "requires AVX..."] integration tests
(vector_filter_by_media, vector_filter_by_doc_id)
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- wire schema: relax effective_end.minimum 1 → 0 + expand
description to cover line-clamp + out-of-range sentinel
(panic-fix R1 emits Some(0) when line_start=1 and range is
beyond doc end — schema must accept it)
- tests: tighten first-chunk-target boundary test to assert ≤ 2
total neighbors (3-chunk doc, N=2). Strict "first chunk →
context_before empty" not assertable until chunks.ordinal
column lands (R1 #9 architectural caveat)
- store: trim contradiction in list_chunk_ids_for_doc warning
comment — drop "good enough for sequentially chunked
markdown" phrase that conflicts with "hash sort dominates"
paragraph above
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Chunk mode + +-N context. doc / span modes return placeholder
errors (filled by subsequent tasks). fmt_canonical_to_markdown
helper introduced now since doc mode (Task 4) consumes it.
Errors are typed StructuredError so classify preserves
chunk_not_found / doc_not_found through the wire layer.
Adds SqliteStore::list_chunk_ids_for_doc so the facade can derive
+-N neighbors without leaking direct rusqlite usage into kebab-app.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Replace kebab-app's private `KEBAB_PARSE_MD_VERSION` literal with a
direct reference to `kebab_parse_md::PARSER_VERSION` so the parser
version cascade has a single source of truth (design §9 invariant).
Add maintenance comment on schema.rs WIRE_SCHEMAS const pointing to
docs/wire-schema/v1/ + kebab-cli wire helpers as the authoritative
sources to keep in sync.
Tighten open_existing doc comment to match the actual SQLITE_OPEN_READ_WRITE
flag (needed for WAL pragma application) — callers should still avoid
issuing mutations through this connection.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Replace `path.exists()` + `Connection::open` (which silently CREATEs on
race) with `Connection::open_with_flags` using READ_WRITE|URI but NOT
CREATE. SQLite surfaces `SQLITE_CANTOPEN` for missing files; we wrap as
NotIndexed { found: None } as before.
Adds open_existing_does_not_create_missing_db regression test pinning
the no-side-effect invariant.
Also documents read-only intent on open_existing, the format contract
on NotIndexed.found, and removes scaffolding comments from kebab-app
error_signal that are no longer load-bearing.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
New `SqliteStore::open_existing` API + `NotIndexed` signal for the
missing-DB case. kebab-app re-exports the type via its `error_signal`
module so kebab-cli's `error_classify` can map it to
`error.v1 { code: "not_indexed" }`.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Add `DocumentStore::get_asset_by_workspace_path` trait method to
`kebab-core` and implement it on `SqliteStore` via a private
`asset_from_row` helper. Used by the incremental-ingest skip path to
compare a freshly-computed blake3 checksum against the persisted row
without a full round-trip through `put_asset_with_bytes`.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Add V006__incremental_ingest.sql to persist last_chunker_version and
last_embedding_version on the documents table. Wire both columns into
upsert_document (INSERT + ON CONFLICT UPDATE) and get_document (SELECT +
row mapper), replacing the previous hardcoded None. Add two round-trip
tests in tests/incremental_ingest.rs covering the set and None cases.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
reviewer-flagged: aa2a6ea claimed build clean but missed:
- crates/kebab-store-sqlite/tests/ingest_report_snapshot.rs (test fixture)
- crates/kebab-cli/src/wire.rs (test fixture)
- crates/kebab-store-sqlite/snapshots/ingest_report.snapshot.json (snapshot)
All three add `unchanged: 0` (or `\"unchanged\": 0`) to match the new
IngestReport.unchanged field. cargo clippy --workspace --all-targets
-- -D warnings now clean.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- `append_turn` 의 doc 은 "wrap in one transaction" 보장하지만 실제
코드는 auto-commit `conn.execute` 두 번이라 두번째 실패 시 first
row 가 commit 된 채 inconsistent 됨. 진짜 transaction 으로 교체:
`conn.transaction()` → `tx.execute(insert)` → `tx.execute(update)`
→ `tx.commit()`. SQLite BEGIN 으로 감싸 두 statement atomic.
`lock_conn()` 이 `MutexGuard<Connection>` 반환하므로 `let mut conn =
self.lock_conn(); let tx = conn.transaction()` 패턴 가능 (MutexGuard
의 DerefMut 활용).
9 chat_sessions 테스트 + clippy 통과.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
도그푸딩 item 15 — TUI / 같은 process 안에서 동일 query 반복 시 SQLite
FTS + Lance + RRF 재계산이 매번 발생하던 비용 해소. in-process LRU
캐시 + 모노토닉 corpus_revision 카운터로 ingest commit 발생 시 모든
entry 자동 stale.
## 핵심 변경
- **SQLite V004 migration**: `kv (key TEXT PRIMARY KEY, value TEXT)
STRICT` + `corpus_revision = '0'` seed. 미래의 다른 scalar 도 같은
테이블에 들어갈 수 있는 generic shape.
- **`SqliteStore::corpus_revision()` / `bump_corpus_revision()`** —
`UPDATE ... CAST AS INTEGER + 1` atomic. INSERT-OR-IGNORE 도 함께
실행 (V004 seed 가 무슨 이유로 누락된 케이스 paranoid).
- **`kebab-app::ingest_with_config_cancellable`** — `new + updated > 0`
시 bump, no-op (skipped-only) reingest 는 cache 보존.
- **`App.search_cache: Option<Mutex<LruCache<SearchCacheKey, Vec<
SearchHit>>>>`** — `config.search.cache_capacity` (default 256, 0
비활성). `lru = "0.12"` workspace dep 추가.
- **`SearchCacheKey`** = `query_norm` (NFKC + trim + lowercase) +
`mode` + `k` + `snippet_chars` + `embedding_version` (vector/hybrid
만, lexical 은 빈 문자열) + `chunker_version` + `corpus_revision`
snapshot.
- **`App::search`** rewrite — cache 활성 시 lookup → miss 면 기존
`search_uncached` 호출 후 put. cache 비활성이거나 lock 실패면
straight-line.
- **`App::search_uncached`** (rename of pre-fb-19 `search` body) +
`search_uncached_with_config` facade — CLI `kebab search --no-cache`
로 진입.
- **`Config.search.cache_capacity: usize`** field, `#[serde(default)]`
로 기존 config 호환.
- **CLI `--no-cache`** flag — 디버깅용 (CLI 는 매 호출이 새 process
라 사실상 no-op 이지만 spec 명시 + 향후 long-lived process 호환).
- **frozen design §9 versioning** 표에 `corpus_revision` row 추가
(기존 `index_version` 라벨과 다른 차원: 라벨은 retrieval 형상,
corpus_revision 은 ingest commit ack).
## 테스트
- `kebab-store-sqlite` 신규 3 unit (fresh=0, monotonic bump, persist
across reopen)
- `kebab-app` 신규 4 integration (cached repeat 같은 hits, NFKC 정규화
로 case/whitespace collapse, --no-cache parity, first ingest bumps
corpus_revision)
- 워크스페이스 전체 `cargo test --workspace --no-fail-fast -j 1` exit 0
- `cargo clippy --workspace --all-targets -- -D warnings` clean
## 문서
- README `kebab search` 행: 캐시 동작 + `--no-cache` 안내 + corpus_
revision 무효화 메커니즘
- docs/SMOKE.md `[search]` 절에 `cache_capacity` 라인 추가
- HANDOFF: 2026-05-03 entry
- spec status planned → in_progress
## Out of scope
- patch-and-merge incremental (RRF 정규화 전체 hit set 기준이라 어려움)
- SQLite 영속 cache (P+)
- 다른 process 간 cache 공유 (in-process 만 — corpus_revision 이
cross-process 무효화는 O(1))
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
`kebab-normalize::derive_title(frontmatter_title, blocks, file_stem)` 가
다음 단계로 비어있지 않은 첫 결과를 사용:
1. frontmatter `title` (trim 후)
2. 첫 H1 텍스트
3. 첫 H2 텍스트
4. 첫 Paragraph (Quote / List / Code / Table / ImageRef 제외) 의 첫 80 자
5. 파일 stem (확장자 제외)
6. (sentinel) `"untitled"` — 위 다섯 단계가 모두 blank 인 병적 케이스
선택된 문자열은 NFC 정규화. 빈 문자열은 절대 반환하지 않음.
`build_canonical_document` 가 metadata lift 직후 helper 호출. 기존 단순
lift 로직 (metadata.user["title"] → CanonicalDocument.title) 은 fallback
chain 의 1 단계 입력으로 자리 이동.
`KEBAB_PARSE_MD_VERSION` 상수를 `pulldown-cmark-0.x` → `md-frontmatter-v2`
로 bump. parser_version 변경 → §4.2 doc_id 입력 변화 → 기존 markdown
doc 의 `doc_id` 갱신, 다음 ingest 시 idempotent upsert 로 자동 재처리
(design §9 cascade). `kebab-store-sqlite` 의 snapshot fixture 도 같은
literal 로 갱신.
기존 M7 정책 ("metadata.user[\"title\"] = '' 가 빈 title 로 lift") 은
폐기. 빈 문자열 입력은 fallback chain 을 타고 file stem 까지 떨어진다.
spec p9-fb-07 line 37: "빈 문자열 반환 금지".
테스트 (kebab-normalize):
- 8 개 단위 테스트 (각 fallback 단계 + NFC + sentinel)
- `build_canonical_document` 통합 테스트 2 개 (H1 / file stem)
- 기존 M7 테스트 2 개를 새 정책에 맞춰 갱신
문서:
- README: `kebab ingest` 행에 "title 자동 채움" 안내 + 기존 doc 도
다음 ingest 에서 갱신
- HANDOFF: 2026-05-03 머지 후 발견 entry
- spec status: `planned` → `in_progress`
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Spec PR #59 의 §3.8 multi-turn behaviour 구현. RAG facade 가 prior
turns 받아 prompt 에 prepend, retrieval query expansion 적용,
Answer 에 conversation_id / turn_index 채움.
신규 (kebab-core):
- Answer 에 conversation_id (Option<String>) / turn_index (Option<u32>)
field 추가. serde skip_serializing_if 로 single-shot 의 wire
output 변경 0 (기존 외부 wrapper 영향 없음).
- Turn struct (question + answer + citations + created_at).
- RefusalReason::LlmStreamAborted variant.
신규 (kebab-rag):
- AskOpts 에 history (Vec<Turn>) / conversation_id / turn_index 3 field.
- AskOpts::single_shot(mode) helper.
- RagPipeline::ask_with_history(query, history, conversation_id,
turn_index, opts) — combined opts 로 ask 호출.
- expand_query_with_history: history.last() 의 answer 첫 200 자
concat 해 SearchQuery.text 확장 (spec §3.8 의 \"cheap concat\";
LLM-based standalone-question rewriting 은 P+).
- serialize_history + remaining_history_budget_chars: spec 의 priority
enforcement — system+question 필수, retrieved chunks 가 차지한
뒤 남은 char budget 안에서 newest 우선, oldest drop.
- ask 본문: history 가 비어있지 않으면 [이전 대화] 블록을 user
prompt 위에 prepend. Answer 생성 site 3 곳 (정상 / NoChunks /
ScoreGate refuse) 모두 conversation_id / turn_index 채움.
신규 (kebab-store-sqlite):
- refusal_reason_label 가 LlmStreamAborted → 'llm_stream_aborted'.
기존 caller 변경 (single-shot 동작 동일):
- kebab-cli main.rs Cmd::Ask: AskOpts 에 history=Vec::new(),
conversation_id=None, turn_index=None 명시 (CLI multi-turn 은
p9-fb-18 의 --session/--repl 가 채움).
- kebab-tui src/ask.rs spawn site 동일 (multi-turn UI 는 p9-fb-16).
- kebab-eval runner.rs golden eval 동일 (single-shot per query).
- kebab-app tests/ask_smoke.rs / kebab-tui tests/ask.rs / kebab-rag
tests/pipeline.rs / kebab-eval metrics.rs Answer literal 갱신.
Test:
- 9 신규 lib unit (expand_query 4 / serialize_history 3 / remaining_budget 2).
- 기존 12 PASS 회귀 0.
Plan 갱신:
- p9-fb-15 status planned → in_progress. 머지 후 한 줄 commit
으로 completed flip.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- (critical) embeddings.rs: truncate_embedding_records 위치 이동.
mark_embedding_records_committed 함수 위에 끼워 넣었더니 위쪽
mark_committed 의 14 줄짜리 doc comment (`WHERE status='pending'`
의 design rationale 등) 가 truncate 의 doc 으로 흡수되고
mark_committed 자체는 doc 없이 남는 버그. impl block 끝 (mark_committed
의 닫는 } 다음) 으로 옮겨 plan 의 원래 의도와도 일치.
- (nit) tests/reset_cli.rs: removed_paths 의 idempotency 검증 보강.
data dir 은 reported, cache dir 은 omit (생성 안 했으니)
되어야 함을 strict 하게 assert. state dir 은 logging init 의
side-effect 로 자동 생성되어 둘 다 가능하므로 허용.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Wipes every row from embedding_records and returns the deleted row
count. Used by the upcoming `kebab reset --vector-only` to keep SQLite
consistent after the on-disk Lance store is removed.
Plan deviation from the original spec (task 1):
- Original test plan opened SqliteStore with a raw path; the actual
signature is `SqliteStore::open(&Config)`, so the integration test
builds a Config with `storage.data_dir` pointed at a tempdir.
- Original return type was Result<()>; bumped to Result<u64> so the
caller (kebab-app::reset) can surface the truncated count in the
reset_report.v1 wire payload without a separate COUNT query.
p9-fb-06 task 1.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
P7-3 의 storage UNIQUE bug fix 가 SQLite 측 (documents → blocks /
chunks / embedding_records) 만 sweep 했음. LanceDB 의 vector 는 별도
store 라 옛 chunk_id 를 가진 row 가 디스크에 잔존. 검색에는 영향 없지만
디스크는 무한 누적. HOTFIXES `2026-05-02 P7-3` caveat 의 "P+ task" 약속을
같은 후속 PR 안에서 닫음.
변경:
- `VectorStore::delete_by_chunk_ids(&[ChunkId])` trait method 추가 (default
no-op 제공 — 테스트 fake / 기존 impl 이 그대로 컴파일).
- `LanceVectorStore::delete_by_chunk_ids` 가 connection 의 모든
`chunk_embeddings_*` 테이블을 순회 + `Table::delete("chunk_id IN (...)")`
를 batch=200 단위로 실행. 다중 모델 워크스페이스 (마이그레이션 중간 등)
에서도 안전.
- `SqliteStore::stale_chunk_ids_at(workspace_path, new_asset_id)` 가
read-only SELECT 로 옛 chunk_id 들 반환. CASCADE 가 흐르기 *전* 에
caller 가 호출.
- `kebab-app::purge_vector_orphans_for_workspace_path` 가 위 두 단계를
orchestrate. 세 ingest path (markdown / image / pdf) 의
`put_asset_with_bytes` 호출 직전에 한 줄로 호출.
Smoke 검증 (release binary, fastembed enabled):
- whitepaper.pdf 첫 ingest → chunk_ids = {f616…, 4e0f…}, vector store 에
그 두 ID 의 row 존재.
- byte 변경 후 re-ingest → 새 doc_id (3741…) + 새 chunk_ids
(ed0c…, e13c…). vector search "REWRITTEN chapter two" → 새 chunk_ids 만
hit. 옛 query "Edited page two body" 시도해도 옛 chunk_ids 는 vector
store 에 더 이상 없음 (의미적으로 가장 가까운 새 chunks 가 hit).
HOTFIXES `2026-05-02 P7-3` 의 \"vector store cleanup\" 항목이 \"deferred\" →
\"closed by follow-up PR\" 로 갱신. SMOKE.md 의 알려진 동작 (\"옛 vector
잔존\") 도 \"두 store 정합\" 으로 갱신.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
P7-3 통합 테스트가 노출한 storage 레이어 버그 fix.
`assets.workspace_path` 의 UNIQUE 제약과 `upsert_asset_row` 의
`ON CONFLICT(asset_id)` 만 처리하던 gap 사이 — byte 가 변경된 자산
re-ingest 시 새 asset_id 가 같은 workspace_path 에서 secondary UNIQUE
충돌. md / image / pdf 모두 영향.
Fix:
- 새 helper `purge_orphan_at_workspace_path` 가 같은 `workspace_path`
의 *다른* `asset_id` 를 발견하면 documents → assets 순서로 sweep.
documents 의 ON DELETE RESTRICT 회피 + CASCADE 로 blocks / chunks /
embedding_records 정리. copied 모드면 storage_path 의 byte 파일도
best-effort 삭제.
- `put_asset_with_bytes` 의 두 분기 (copy / reference) + `DocumentStore
::put_asset` 모두 호출.
- 회귀 테스트 `put_asset_with_bytes_sweeps_workspace_path_orphan` (이전
의 "UPSERT 실패시 orphan 청소" 테스트가 더 이상 doable 하지 않으므로
대체).
- `re_ingest_edited_pdf_produces_new_doc_id` integration `#[ignore]` 해제 →
9 통합 테스트 모두 default 로 통과.
Vector store orphan 은 별도 P+ task — LanceDB 가 SQLite cascade 와 무관하게
운영되므로 stale chunk_id vector 가 디스크에 남음. 검색에는 영향 없음 (search 가
SQLite join 통해 surface).
Smoke 검증 (release binary, markdown 2 + image 1 + PDF 2):
- doctor pass
- 첫 ingest: 5 new
- list docs: 5 docs all media types
- search lexical "pdf-page-v1 chunker" → whitepaper.pdf hit
- search hybrid → cross-media 결과
- inspect doc PDF: parser_version=pdf-text-v1, blocks 가 SourceSpan::Page
- 동일 byte re-ingest: 5 updated, 0 errors (P1 idempotency)
- byte 수정 후 re-ingest: 1 new (해당 PDF) + 4 updated, 0 errors (storage fix)
- corrupt PDF 추가: errors+=1 + IngestItem.error 메시지 정확, 다른 자산 영향 0
- 정리 후 다시 ingest: errors=0
- RAG ask: PDF 인용 + `citations[].citation` 에 `kind: "page"` + `page: <N>` +
`path: <pdf_path>` 정확히 노출
운영 fixture 보조:
- `crates/kebab-parse-pdf/examples/gen_smoke_pdf.rs` — `cargo run --release
--example gen_smoke_pdf -p kebab-parse-pdf -- <out.pdf> <text-pages>` 로
reportlab/qpdf 없이 in-tree PDF 생성.
- `crates/kebab-parse-image/examples/gen_smoke_png.rs` — 동일 방식의 PNG
fixture 생성.
- SMOKE.md 가 두 example 사용법 + 갱신된 HOTFIXES 동작 (byte 수정 시
errors+=1 → new+=1) 반영.
HOTFIXES `2026-05-02 P7-3` entry 가 \"deferred\" → \"fixed in same PR\" 로
업데이트, vector store orphan caveat 만 남음.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
