test 7 의 `i32_below_gate_chunk` helper rename → `seed_low_score_chunk` +
반환 shape 을 `(chunk_id, doc_id)` tuple 로 확장. `i32` prefix 가 Rust
integer 타입과 충돌하던 가독성 문제 해소 + 호출자가 `id32("d_low")` 를
재계산하지 않도록 id 페어를 single source of truth 로 통합.
검증
- `cargo test -p kebab-rag -j 1 --test multi_hop` — 7 모두 통과.
- `cargo clippy -p kebab-rag --all-targets -j 1 -- -D warnings` clean.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
(A) ScriptedLm doc 의 `Arc<Vec<String>>` 표기 → 실제 구현 (`Vec<String>` +
`AtomicUsize`, 외부에서 `Arc::new(ScriptedLm::new(...))` 로 wrap)
반영.
(B) ScriptedLm::new doc 의 미존재 `with_*` builder 언급 제거.
(C) refuse path 의 hops 보존 회귀 핀 2 건 추가 (`tests/multi_hop.rs`):
- `multi_hop_refuse_no_chunks_preserves_hops_trace`: empty pool →
`refuse_no_chunks(Some(hops))` → Answer.hops = Some([Decompose,
Decide]).
- `multi_hop_refuse_score_gate_preserves_hops_trace`: top score 0.10
< 0.30 gate → `refuse_score_gate(Some(hops))` → 같은 shape.
refuse_* widening + ask_multi_hop 의 forwarding wiring 이 reverting
되면 두 test 가 회귀 잡음.
(D) test 5 의 redundant `assert_ne!(.., Some(MultiHopDecomposeFailed))`
제거 — `assert_eq!(.., None)` 이미 함의. 메시지에 의도 통합.
검증
- `cargo test -p kebab-rag -j 1 --test multi_hop` — 7 (5+2) 모두 통과.
- `cargo clippy -p kebab-rag --all-targets -j 1 -- -D warnings` clean.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
PR-3b 의 분할 첫 PR. ask_multi_hop 의 fixed depth=2 → dynamic N-hop.
ScriptedLm helper + 5+ integration tests (happy-path 통합 검증) 는
PR-3b-ii 분리. 본 PR 의 회귀 핀 = 기존 PR-2 의 2 integration test
통과 (decompose garbage refusal + multi_hop=false single-pass keep).
- `RagPipeline::multi_hop_decompose` 시그니처 변경 — `Result<
(Option<Vec<String>>, u32)>` (parsed result + LLM call latency_ms).
caller (`ask_multi_hop`) 가 hop trace 의 `llm_call_ms` stamp.
- `RagPipeline::multi_hop_decide` helper 신규. decide LLM call →
`parse_decompose_response` 으로 `Option<Vec<String>>` 반환. None
또는 empty array 가 stop signal (refusal 아닌 graceful degrade).
- `MULTI_HOP_DECIDE_SYSTEM_PROMPT` const 신규.
- `MULTI_HOP_DECOMPOSE_USER_TEMPLATE` const 제거 + `format!` named
arg 사용 (PR-2 회차 1 carry-over fix). compile-time substitution
check — 사용자 query 안에 `{max_sub_queries}` literal 있어도
mis-replace 회피.
- `ask_multi_hop` 의 §1 (Decompose) + §2 (Retrieve) 영역을 dynamic
loop 으로 재작성:
- iter 0 = decompose, HopRecord 추가 (kind=Decompose).
- iter 1..=max_depth = retrieve current_sub_queries → pool dedup
→ decide LLM call (forced_stop / pool_cap_hit 시 skip).
HopRecord 추가 (kind=Decide, sub_queries=new_sub_queries,
context_chunks_added, forced_stop, llm_call_ms).
- `max_pool_chunks` 도달 시 `pool_cap_hit = true` → 그 iter 의
HopRecord 가 `forced_stop = true` + decide LLM call skip.
- depth 도달 (`iter >= max_depth`) 시 동일하게 forced_stop.
- decide parse failure 또는 empty array → loop break (early
synthesize, NOT refusal — spec §9 graceful degrade).
- §6 (Generate) 시작 시 `synthesize_started: Instant::now()` 별
stamp → §8 Build Answer 직전 `HopRecord { kind=Synthesize,
llm_call_ms = synth_ms }` 추가. happy path 의 Answer literal
`hops: Some(hops)` 채움 (`hops: None` → `Some(...)` 변경).
- doc comment 갱신: "PR-2 scope (fixed depth=2)" → "PR-3b-i
scope (dynamic N-hop)". refusal path 의 hops trace 손실 caveat
명시 (PR-3b-ii / follow-up 에서 helper signature 확장 시 해결).
기존 회귀 핀 (PR-2 의 2 integration test):
- `ask_multi_hop_dispatches_and_decompose_garbage_refuses`:
decompose garbage → RefusalReason::MultiHopDecomposeFailed +
정확히 1 LLM call. PR-3b-i 의 시그니처 변경 후도 통과.
- `ask_with_multi_hop_false_keeps_single_pass_path`: 영향 없음.
56 unit + integration test 모두 통과 (kebab-rag).
Wire 영향: `Answer.hops` 가 multi-hop happy path 에서 emit. JSON
Schema additionalProperties default `true` 라 wire breaking 아님
(PR-3a 의 review 확인). schema.json 명시 갱신은 별 PR
(PR-3b-ii 또는 PR-4 의 schema sweep).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- `parse_decompose_response_drops_partial_empty_keeps_valid` 신규
회귀 핀 — `["", "valid q", " "]` → `["valid q"]` (trim+filter
chain 동작 pin).
- `multi_hop_decompose` 의 `stop: Vec::new()` 옆 doc comment 추가 —
의도 명시 (instruction-following 모델 기대 + prose 추가 시
MultiHopDecomposeFailed refusal 가 policy). 회차 1 question 의
답변.
- plan 의 PR-3 implementation order 에 회차 1 carry-over 추가:
1) ask + ask_multi_hop 의 §4-§9 mirror → 공통 helper 추출,
2) decompose template 의 substitution corner case → format!
named arg 으로 교체.
회차 1 의 다른 suggestion (mirror refactor, substitution corner
case, history block helper) 는 PR-3 합리적 timing 으로 plan 에
명시 — 회차 2 reply 에 정리.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
PR-2 of fb-41 multi-hop RAG. Decompose + retrieve + synthesize 3-stage
pipeline가 `opts.multi_hop=true` 일 때 dispatch. Dynamic decide loop
는 PR-3.
- `AskOpts.multi_hop: bool` 필드 추가 + `impl Default for AskOpts`
도입 (HOTFIXES 2026-05-07 의 known limitation 해소). 9 explicit
init site 모두 `multi_hop: false` 추가 — Default 도입으로 향후
`..Default::default()` 점진 migrate 가능.
- `RagPipeline::ask` 의 entry 에 dispatcher 한 줄
(`if opts.multi_hop { return self.ask_multi_hop(...) }`).
- `RagPipeline::ask_multi_hop` 신규 method. 1) decompose LLM call
→ JSON array of strings parse, 2) 각 sub-query 로 retrieve +
chunk_id dedup pool, 3) score gate / no-chunks 가드, 4)
pack_context (single-pass 와 helper 공유), 5) synthesize LLM
call w/ MULTI_HOP_SYNTHESIZE_SYSTEM_PROMPT, 6) citation extract
+ Answer build. `prompt_template_version` = "rag-multi-hop-v1"
로 stamp — eval `compare` 가 single-pass vs multi-hop 분리.
- Prompt const 신규: MULTI_HOP_DECOMPOSE_SYSTEM_PROMPT +
MULTI_HOP_DECOMPOSE_USER_TEMPLATE + MULTI_HOP_SYNTHESIZE_SYSTEM_PROMPT
+ PROMPT_TEMPLATE_VERSION_MULTI_HOP + MULTI_HOP_MAX_SUB_QUERIES_DEFAULT.
- `kebab_core::RefusalReason::MultiHopDecomposeFailed` variant 신규.
Cascade: kebab-store-sqlite `refusal_reason_label` + kebab-tui `ask
refusal render` exhaustive match 갱신.
- `parse_decompose_response` + `strip_markdown_json_fence` helper —
markdown code fence (```json / ```) strip + JSON array of strings
parse + trim + drop empty + cap at MULTI_HOP_MAX_SUB_QUERIES_DEFAULT.
None 반환 시 caller 가 `MultiHopDecomposeFailed` refusal.
Tests (55 passing total, 8 신규):
- 6 unit (parse_decompose_response 의 bare array / fence variants /
garbage / cap / trim 회귀 핀).
- 2 integration: `ask_multi_hop_dispatches_and_decompose_garbage_refuses`
(decompose garbage → MultiHopDecomposeFailed + 정확히 1 LLM call) +
`ask_with_multi_hop_false_keeps_single_pass_path` (회귀 핀, 기존
caller 자동 backwards-compat).
Happy-path multi-hop (decompose 성공 → synthesize) 의 integration
test 는 ScriptedLm helper 가 PR-3 의 decide loop 와 함께 도입될
때 같이 추가. 현 `MockLanguageModel` 는 canned single response 라
2-LLM-call sequence 핀 불가.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
PR-1 of fb-41 multi-hop RAG (spec: docs/superpowers/specs/2026-05-25-
p9-fb-41-multi-hop-rag-design.md, plan: docs/superpowers/plans/2026-
05-25-p9-fb-41-multi-hop-rag.md).
XL 작업의 첫 PR — baseline 측정 anchor 만 추가. RAG pipeline 미변경,
fixture file + parse 회귀 핀.
사용자 결정 4 axis (2026-05-25):
- approach: query decomposition (LLM 서브-질문)
- trigger: explicit `--multi-hop` flag
- MVP scope: dynamic N-hop (LLM 이 depth 결정, decompose seed +
ReAct-style decide loop hybrid)
- eval: multi-hop golden set 먼저 (본 PR)
본 PR:
- `fixtures/multi_hop_golden.yaml` 신규. 15 question (5 cross-doc +
5 intra-doc + 5 single-fact negative). 기존 `GoldenQuery` struct
그대로 사용 — 별 loader / type 변경 없음. `expected_chunk_ids`
비어 있어 curator 가 `kebab ingest` 후 채울 수 있는 template
형태. `must_contain` 으로 baseline 측정 가능 (P5-2 metric).
- `crates/kebab-eval/tests/loader.rs::loads_multi_hop_golden_fixture`
신규 회귀 핀. fixture parse OK + 15 question + 5/5/5 bucket
분포 + 모든 question 에 must_contain 최소 1 개.
baseline 측정 protocol (별 run, commit 에 artifact 안 포함):
1. v0.17.2 binary 로 single-pass `kebab eval run --fixture
multi_hop_golden.yaml` 실행
2. P@5, P@10, must_contain pass rate, citation_coverage 캡처
3. PR-3 (dynamic iter 머지) 후 동일 fixture + `multi_hop=true` 로
재실행 → Δ 비교
PR 분할 6 단계 (plan 참조): PR-1 (본 PR — fixture only), PR-2
(RagPipeline::ask_multi_hop fixed depth=2), PR-3 (dynamic iter),
PR-4 (CLI flag + wire), PR-5 (MCP + SKILL.md), PR-6 (TUI toggle +
trace render). 마지막 PR 후 v0.18.0 cut.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
v0.17.0 trigram tokenizer entry 가 미수정으로 남겨둔
heading_path_json JSON 노이즈 (HOTFIXES 2026-05-24) closure.
trigram 이 chunks_fts.heading_path 컬럼 (V002/V007 트리거가
chunks.heading_path_json 그대로 INSERT) 의 JSON 표기 + 안의 path
세그먼트 (app, src) 까지 3-gram 색인해서 query 가 우연히 false
positive hit 하는 문제. column filter 채택 — heading 색인 유지
(V007 verbatim 불변), 매칭 대상만 text 컬럼 한정.
- build_match_string 가 non-raw 분기에서 combined expression 을
`text : (<expr>)` 로 wrap. FTS5 column filter syntax 가 OR/AND
sub-expression 허용.
- Raw mode (`'...'`) 는 그대로 — 사용자가 명시 의도로
`'heading_path : agent'` 같은 explicit opt-in 가능 (escape hatch).
- 8 기존 build_match_string unit test expected string 갱신 +
`build_match_string_raw_mode_preserves_heading_filter` 신규.
- `lexical_heading_only_token_does_not_hit_default_mode` 신규 회귀 핀
(heading-only unique token 이 default mode 에서 0 hit).
- `lexical_raw_mode_can_opt_into_heading_path_filter` 신규 — 같은
fixture 가 raw mode 로 hit 확인 (escape hatch 동작 핀).
사용자 영향: lexical / hybrid 검색의 본문 precision ↑. recall
변화 없음 (text 본문 token 매칭은 동일). re-ingest 불필요 (FTS
query 시점 매칭만 변경). lexical_snapshot_run_1 + hybrid_snapshot
도 fixture regenerate 불필요 (text 본문 매칭 query 라 BM25 동일).
HOTFIXES: 2026-05-24 v0.17.0 entry 의 `heading_path_json` 노이즈
항목 closure 표기 + 새 2026-05-25 post-v0.17.1 dogfood entry 추가.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- HOTFIXES 헤더 `v0.17.2` (vaporware) → `post-v0.17.1 dogfood`
로 변경, release tag 결정과 무관하게 정확한 anchor.
- HOTFIXES caller 수 `6 (5+3)` → `9 call site (6+3)` 으로 정정.
- OcrCfg.request_timeout_secs doc 의 edge case 가 LlmCfg sister
doc 과 동일한 구체 예제 (`u64::MAX`, `86400`) + reqwest 0.12.x
명시 주석으로 강화.
- LLM + OCR 양쪽의 legacy TOML fixture (78 줄 거의 동일) 를
module-level `LEGACY_PRE_TIMEOUT_TOML` const 로 추출. 두 test
가 동일 source 공유 → 옛 schema 가 또 변하면 한 곳만 수정.
reqwest::Duration::ZERO fact-check (회차 1 점 5) 는 회차 2
reply 에서 검증 결과 보고.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
v0.17.1 (PR #162) 가 LLM 쪽 hard-coded 300s 를 [models.llm]
request_timeout_secs 로 풀어준 것과 같은 패턴을 OCR 어댑터에 적용.
사용자 결정으로 별 노브 분리 ([image.ocr] request_timeout_secs) —
OCR 는 LLM 대비 cold start 패턴이 달라 독립 조절이 편함.
- OcrCfg.request_timeout_secs: u64 (serde default 300)
- KEBAB_IMAGE_OCR_REQUEST_TIMEOUT_SECS env override
- OllamaVisionOcr::build / from_parts 시그니처에 timeout 인자 추가
- REQUEST_TIMEOUT 상수 제거
- 3 신규 unit test (default / env / legacy parse) — LlmCfg 패턴 그대로
- HOTFIXES 2026-05-25 v0.17.1 entry 의 두 미진행 항목 모두 closure
(OCR timeout = 본 PR, --stream docs = PR #163 에서 이미 완료)
기존 config / 옛 KB 영향 없음 — 새 필드는 default 로 채워지고
동작도 동일 (300s). vision 모델 cold start 가 길면 env 또는
config 로 늘릴 수 있음.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
PR #162 워커 리뷰 반영.
- MEDIUM (W2) + LOW (W1): request_timeout_secs = 0 이 reqwest 의
의미상 disable 이 아닌 instant timeout (모든 요청 즉시 실패).
LlmCfg field rustdoc + ollama.rs module-level comment + README
세 군데에 명시 + u64::MAX / 86400 같은 large finite 값 권장.
- NIT (W1): HOTFIXES 2026-05-25 entry 의 '답변이 인 5분' typo →
'답변이 5분' (1자 삭제).
- NIT (W1): README + HOTFIXES 의 '확장 도그푸딩' 내부 jargon →
'후속 도그푸딩' 으로 통일.
코드 동작 변경 없음 — doc only. cargo test request_timeout 3 PASS.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
v0.17.0 확장 도그푸딩 (2026-05-25) 에서 발견된 두 가지를
한 PR 에 묶음.
(1) llm.generate_stream 의 hard-coded 300s timeout 을 config 노브로
빼냄. 8B+ 모델 (gemma4:e4b 등) 은 CPU only 환경에서 5분
안에 첫 RAG 답변 못 마치고 `error: kb-rag: llm.generate_stream`
으로 떨어지던 문제.
- kebab-config::LlmCfg 에 request_timeout_secs: u64 additive
필드 (#[serde(default = "default_llm_request_timeout_secs")]
default 300). 옛 config 가 키 누락해도 그대로 파싱 + 동일
동작.
- env override KEBAB_MODELS_LLM_REQUEST_TIMEOUT_SECS.
- kebab-llm-local::ollama.rs 의 REQUEST_TIMEOUT 상수 제거 →
OllamaLanguageModel::new 가 Duration::from_secs(
llm.request_timeout_secs) 로 reqwest client 빌드. doc
comment 도 동일 갱신.
- 신규 unit test 3 — default 300 핀 / env override / legacy
config (필드 누락) backward-compat.
(2) docs — README 사전 요구 절 + docs/SMOKE.md ollama 안내에 한 단락:
CPU only / RAM ≤ 16 GB 환경 ⇒ ≤ 4B Q4 모델 권장
(gemma3:4b / qwen2.5:3b / phi3:mini). 8B+ 시도 시 timeout
패턴 사전 안내. request_timeout_secs 노브 사용법.
HOTFIXES 2026-05-25 entry — 위 두 변경 + 미진행 사항
(kebab-parse-image OCR 의 같은 hard-coded 300s 는 scope 외
follow-up 으로 등재 + ask --stream 권장 강조 후속) 기록.
workspace cargo test -j 1 + clippy 통과. 코드 변경은 backwards-compat
(additive serde field) 라 기존 사용자 영향 없음.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
실제로 doc count 임을 명시 (PR #161 워커 리뷰 MEDIUM 반영)
JSON schema description 은 PR-C 본체에서 'code chunk count' →
'doc count' 로 정정했으나 Rust struct field 의 rustdoc 은 같은
오기재를 그대로 carry — Gemini round 2 가 JSON schema 만 봤고
rustdoc 은 miss. 워커 둘 다 동일 finding (MEDIUM).
implementation 변경 없음 — 의미가 doc count 였던 사실이 처음부터
일관. wording 만 맞춤.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
closure of HOTFIXES 2026-05-21. C typedef-wrapped anonymous
struct/enum/union 이 typedef alias 이름으로 symbol unit 방출.
- crates/kebab-parse-code/src/c.rs: type_definition 분기 추가.
inner anonymous struct_specifier / enum_specifier / union_specifier
탐지 → declarator field 의 type_identifier 재귀 추출 → synthetic
unit (typedef alias). named inner aggregate / plain alias 는
기존대로 glue. PARSER_VERSION code-c-v1 → code-c-v2.
recover_typedef_alias + extract_typedef_alias_name helper 추가.
- crates/kebab-store-sqlite/src/store.rs: 두 helper 신규
(parser_version bump cascade 용 doc-id 기반 orphan purge).
- stale_chunk_ids_for_workspace_path_except_doc_id(workspace_path,
keep_doc_id) — sister of stale_chunk_ids_at, doc_id 기반.
- purge_document_at_workspace_path_except_doc_id(workspace_path,
keep_doc_id) — CASCADE document/chunks 제거, assets 보존.
keep_doc_id="" 가 "모든 doc 제거" 사용.
- crates/kebab-app/src/lib.rs: try_skip_unchanged 의 parser_mismatch
분기에서 purge_workspace_path_for_parser_bump 호출. helper 가
app.vector() 로 lazy 접근 + delete_by_chunk_ids + SQLite document
row 제거. Ok(None) 반환 전 cleanup 끝나서 caller 의 새 INSERT 시
idx_docs_workspace_path UNIQUE 충돌 회피.
- tests:
- c.rs unit tests 4 신규 — typedef_struct_emits_unit /
typedef_enum_emits_unit / typedef_union_emits_unit /
typedef_to_existing_type_stays_glue (negative).
- tier1_c_ingest_searchable: parser_version assertion code-c-v1 →
code-c-v2.
- 회귀: bytes-edit 경로 (asset_id 변경) 의 기존 purge_orphan_at_workspace_path
+ purge_vector_orphans_for_workspace_path 는 그대로 — 신규 분기와
공존, 기존 test 모두 PASS.
미해결 (Risks): nested typedef (typedef struct { struct {...} inner; }
Outer;) 의 inner 익명 struct 는 여전히 glue — v2 의 1차 범위는
top-level typedef alias 만.
cargo test --workspace --no-fail-fast -j 1 + clippy 통과.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
PR #159 worker-1 review 의 LOW 가독성 nit 반영 — CLI stderr [hint]
line + --json hint shape 통합 test 가 없었음.
- search_plain_emits_short_query_hint_to_stderr — 빈 KB + 2자 query
→ stderr 가 "[hint]" + "3자 이상" 포함 확인.
- search_json_emits_hint_field_for_short_query — 동일 입력 --json
→ search_response.v1.hint 필드 set + 표준 advisory 문자열 정합.
- search_json_omits_hint_field_when_query_is_long_enough — 3자
query → hint 필드 absent (additive serializer 의 None 제외 동작).
wire_search_response 5 → 8 PASS.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
trigram tokenizer 가 snippet 단위 + 단어 경계 + BM25 raw score 분포를
모두 바꿔서 unicode61 assumption 기반의 3 test 가 regression.
- wire_search_response::search_json_truncates_with_max_tokens +
search_plain_emits_truncated_hint_to_stderr: 단일 doc + 작은
max_tokens 로는 snippet 이 짧아서 budget loop 가 trip 안 함.
다중 doc fixture (5 doc) + budget 30 token 으로 hit-pop 경로
통해 truncated=true 보장.
- fetch_integration::fetch_chunk_with_context_returns_neighbors:
fixture body 의 2-char tokens (A1/A3 등) 가 trigram 비호환으로
0-hit. apples/banana/cherry/durian/elder 5-char unique words
로 갱신, query 도 cherry 로 deterministic pin.
- eval/runner::runner_per_query_snapshot_matches_fixture: trigram
token stream 으로 BM25 raw score 변동. UPDATE_SNAPSHOTS=1 로
regenerate.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
3개 신규 unit tests in tests/fts.rs §7:
1. fts_trigram_korean_3char_substring_hits — Codex sqlite 3.45.1 검증
동작 5개 assert pin: raw 3자 substring hit (충돌은/발생한),
quoted phrase hit (\"해시 충돌\"/\"시 충\"), raw 해시충 0-hit (원문
미존재).
2. fts_trigram_korean_short_query_zero_hit_pinned — 2자 한국어 query
(충돌·키) 0-hit 회귀 감지. trigram 구조 변경 시 먼저 fail.
3. fts_trigram_english_substring_hits — substring recall 동작 변경
pin (token→tokenizer, to 0-hit).
검증: cargo test -p kebab-store-sqlite --test fts → 13/13 PASS
(신규 3 + 기존 10).
Step 1c (multi-token 한국어 query e.g. \"해시 충돌\") 와 Step 5
(lexical BM25 snapshot 갱신) 는 Task A5 의 build_match_string()
재설계 후 진행.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
P10 dogfooding found that a k8s manifest with 2+ documents (e.g.
Deployment + Service in one file) fails to ingest:
UNIQUE constraint failed: chunks.chunk_id
Root cause: tier2_shared::push_chunks_with_oversize's non-oversize branch
hardcoded split_key = None. K8sManifestResourceV1Chunker calls it once per
resource; with split_key None every resource from the same document gets
the same id_hash (= base_policy_hash) → identical chunk_id. p10-3's
code_text_paragraph_v1 had the same bug (fixed in df3c5b8) but it calls
build_chunk_no_symbol directly — the push_chunks_with_oversize path was
never fixed.
Fix: push_chunks_with_oversize gains a base_split_key parameter for the
non-oversize single-chunk case. k8s chunker passes Some(resource.line_start)
so each resource gets a distinct chunk_id; dockerfile / manifest pass None
(1 chunk per file — no sibling collision, chunk_id stays stable).
Regression coverage: k8s_multi_doc_emits_one_chunk_per_resource now asserts
chunk_id distinctness; new integration test
tier2_k8s_multi_resource_yaml_ingests_without_collision ingests a real
2-document YAML end-to-end.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Reviewer nit #3: the hand-built fixed_doc() only verified chunker 1:1
mapping. New tests invoke CppAstExtractor against tests/fixtures/sample.cpp
and snapshot the real extractor → chunker pipeline (14 blocks emitted
covering namespace::chunk::Class, ctor/dtor/operator/template/free-fn
convention, glue <top-level> blocks between units).
Adds kebab-parse-code as a dev-dep of kebab-chunk (same precedent as
kebab-parse-md). Both the existing hand-built test AND the new
extractor-driven tests are kept — the former for fast chunker-only
validation, the latter for end-to-end regression detection.
Added tests:
- code_cpp_ast_extractor_snapshot: asserts all 8 named symbol units are present
- code_cpp_ast_extractor_chunks_deterministic: chunker output is stable
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
PR #155 (p10-3) merged WITHOUT the reviewer's required Option B1 fix —
the implementer reported a commit SHA (2a39513) that never made it to main.
Result: every reingest of a Tier 3-fallback file (non-k8s YAML, invalid
YAML, AST extractor failure) re-runs full extract + chunk + embed because
the parser/chunker version comparison can never match (stored is
code-text-paragraph-v1 / none-v1, but caller uses Tier 1/2 dispatch
values).
This commit:
1. Adds the 7th param `fallback_chunker_version: Option<&ChunkerVersion>`
to try_skip_unchanged + the stored_is_tier3_fallback detection branch
(skip parser/chunker equality, keep embedder check).
2. Threads `None` through non-code call sites (md / image / pdf).
3. Code call site computes tier3_fallback_cv covering all Tier 1/2 langs
that can fall back: rust / python / ts / js / go / java / kotlin /
yaml / dockerfile / toml / json / xml / groovy / go-mod / c / cpp
(p10-1D additions).
4. Adds tier3_yaml_fallback_reingest_is_unchanged + tier3_shell_reingest_is_unchanged
regression tests (the originally-promised PR #155 regression coverage
that also never made it to main).
Smoke tests: 14 + 2 = 16 PASS.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Extends 4-arm match (parser_version / chunker_version / extract / chunks)
+ allowlist + tier3_fallback_cv with "c" + "cpp" arms. C uses CAstExtractor
+ CodeCAstV1Chunker; C++ uses CppAstExtractor + CodeCppAstV1Chunker. Both
langs are Tier 3-fallback-eligible (e.g. .h file with C++ syntax may fail
tree-sitter-c parse → Tier 3 paragraph fallback per p10-3 wrapper).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Mirrors code-go-ast-v1's chunker pattern. Snapshot test against
tests/fixtures/sample.c (function + typedef struct + typedef enum +
preprocessor) verifies symbol list + lang=c stamping.
Chunks produced (4 total):
- <top-level> glue: includes, defines, static vars, typedefs (lines 1-18)
- parse_record function (lines 20-23)
- print_record function (lines 25-27)
- main function (lines 29-33)
All chunks stamped with lang=c and chunker_version=code-c-ast-v1.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Symbol = namespace::Class::method via recursive build_blocks. namespace_definition
pushes namespace name (anonymous → <anonymous>). nested_namespace_specifier
(outer::inner) flattens all segments and pushes them. class_specifier / struct_specifier
(named) emit class unit + recurse with class name pushed. function_definition emits
method unit; symbol resolution unpacks declarator chain (pointer_declarator /
reference_declarator → function_declarator → identifier / field_identifier /
qualified_identifier / operator_name / destructor_name).
operator_cast (conversion operators, e.g. operator bool) handled as a direct
declarator kind on function_definition. template_declaration recurses with same
prefix (template params NOT in symbol). enum_specifier + concept_definition emit
type-level units. linkage_specification (extern "C") recurses into body with same
prefix. Other top-level nodes → <top-level> glue.
All 15 unit tests pass; build and clippy clean.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Top-level units: function_definition (symbol = fn name from declarator's
innermost identifier), struct_specifier, enum_specifier, union_specifier
(each emits 1 unit with the named identifier as symbol). Preprocessor
directives + top-level declarations group into a <top-level> glue chunk.
Empty file or zero units → <module> post-pass.
C symbol = function name only — no namespace, no class nesting (design §3.4).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Standard crate names resolved cleanly: tree-sitter-c v0.24.2 and
tree-sitter-cpp v0.23.4 are both compatible with workspace tree-sitter 0.26.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Two new tests verify end-to-end Tier 3 wiring:
- tier3_shell_ingest_searchable: .sh file → --code-lang shell search →
Citation::Code { symbol: None, lang: "shell" }, chunker_version
"code-text-paragraph-v1".
- tier3_yaml_fallback_picks_up_non_k8s_yaml: docker-compose-shaped yaml
(no apiVersion/kind) triggers k8s chunker's Ok(vec![]) result, fallback
retries with Tier 3 → Citation::Code { symbol: None, lang: "yaml" } and
chunker_version "code-text-paragraph-v1".
Also fixes a bug in CodeTextParagraphV1Chunker (Task B): short paragraphs
(≤80 lines) were emitted with split_key=None, causing all paragraphs from the
same document to share the same chunk_id (UNIQUE constraint violation at
put_chunks). Fix: always use para.line_start as split_key so every paragraph
gets a distinct id regardless of size.
Brings code_ingest_smoke to 14 tests (Tier 1: 9, Tier 2: 3, Tier 3: 2).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
After the chunks match resolves, an Ok(empty) result (Tier 2 invalid YAML
/ non-k8s YAML / similar) or Err (Tier 1 extractor / chunker failure) is
retried against CodeTextParagraphV1Chunker. On retry, chunker_version is
swapped to "code-text-paragraph-v1" and canonical.parser_version to
"none-v1" so downstream stamping + try_skip_unchanged remain consistent.
Extract failure is handled similarly — when a Tier 1 extractor errors
(e.g. tree-sitter parse failure), a synthesize_tier2_document-shaped
fallback doc is built from raw bytes and routed through Tier 3 chunker
directly (extract_fell_back guard).
shell direct path + Tier 2 extract synthesize_tier2_document failures
are exempted from the fallback chain (they ARE Tier 3 already, or the
error is real).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Extends ingest_one_code_asset's allowlist + 4-arm match (parser_version /
chunker_version / extract / chunks) to admit code_lang "shell" and route it
to CodeTextParagraphV1Chunker. parser_version "none-v1" + synthesize_tier2_document
reused.
Tier 1/2 fallback wrapper lands in the next commit.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Blank-line paragraph segmentation (whitespace-only lines as boundaries,
blank lines themselves never in any chunk's range). Paragraphs > 80 lines
split into 80-line windows with 20-line overlap (stride 60), sharing the
input lang and symbol=None per spec §9.3. tier2_shared exposes a new
build_chunk_no_symbol helper so Chunk id/hash/token semantics stay
identical with Tier 1/2. Extracts build_chunk_from_span as private core
so build_chunk and build_chunk_no_symbol share mechanics without drift.
4 unit tests cover multi-paragraph shell (4 paragraphs, blank-line
boundaries verified), 200-line oversize line-window split (chunks
1-80 / 61-140 / 121-200), empty file, and lang preservation when
input is yaml.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Tier 3 chunker (next task) needs to call the same Chunk-construction helper
to keep id / hash / token-count / policy_hash semantics identical with
Tier 2. Visibility-only change; signature and body unchanged.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Spec p10-2 risks section calls out "거대 ConfigMap" but no test exercised
the line-window split branch of tier2_shared::push_chunks_with_oversize.
This adds a 256-line ConfigMap fixture (generated inline) and asserts:
- ≥2 chunks emitted (split happened),
- all chunks share symbol `ConfigMap/prod/big`,
- chunk_ids all distinct (id_for_chunk's #L{k} suffix disambiguation),
- line ranges form a contiguous partition (prev.line_end + 1 == next.line_start).
Reviewer nit #1 (PR #153 code-reviewer).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Earlier comment claimed the function "mirrors RustAstExtractor pattern" but
the two differ: RustAstExtractor joins ctx.workspace_root to handle relative
paths, while Tier 2 trusts FsSourceConnector's absolute-path invariant.
Rephrase to document the actual rationale + the Kb URI fallback.
Reviewer nit #3 (PR #153 code-reviewer).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Adds yaml / dockerfile / toml / json / xml / groovy / go-mod arms to the
existing 7-arm AST match. parser_version unified to "none-v1" for Tier 2.
synthesize_tier2_document builds a minimal Document (single Block::Code
with raw file text) since Tier 2 has no parse step. allowlist in
ingest_one_asset extended to admit Tier 2 langs.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Reads entire Dockerfile / Dockerfile.* / *.dockerfile content and emits a
single Chunk with symbol "<dockerfile>", code_lang "dockerfile", line range
1..EOF. Oversize >200 lines splits into line-windows sharing the symbol via
tier2_shared::push_chunks_with_oversize.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Splits multi-document YAML by ^---\s*$, requires apiVersion + kind string
fields per document, emits 1 chunk per recognized k8s resource. Symbol =
<kind>/<namespace>/<name> or <kind>/<name> (cluster-scoped). Invalid YAML
returns 0 chunks (handled by p10-3 paragraph fallback). Oversize >200 lines
splits into line-windows sharing the same symbol.
tier2_shared module hosts the oversize fallback + Chunk-construction helper
mirroring code_rust_ast_v1's Chunk shape. Task E (dockerfile) and Task F
(manifest) will reuse it.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Replaces 1A-1 era inline match block with a single call to
kebab_parse_code::code_lang_for_path, per design §3.5 single-source-of-truth
rule. Adds Tier 2 routing test (yaml / dockerfile / toml / json / xml /
groovy / go-mod) and preserves all non-code extension branches.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Mirrors code_go_ast_snapshot pattern. In-memory CanonicalDocument (no
kebab-parse-code dep — boundary §6.3).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
