altair823-org/kebab
0
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases 46 Wiki Activity

refactor(config): env 97→~25 + 노출 키 109→~30 (표면 정리)

Browse Source 
apply_env 의 KEBAB_* 매치 암을 103개→22개로 정리. 삭제된 암은 런타임에서
바꿀 일이 거의 없는 per-field 튜닝 노브(score_thresh, rrf_k, temperature,
multi_hop_*, nli_threshold, chunker_version, context_tokens 등) — struct
field 와 serde default 는 그대로 유지하므로 TOML 로는 여전히 설정 가능.

유지한 22개: endpoint×3, 모델명/프로바이더×5, 경로×2, 병렬도×2,
청킹 target/overlap×2, OCR 엔진/모델/언어+per-medium enabled×5,
caption enabled×1, search default_k×1, rag prompt_template_version×1.

영향 범위: struct field 삭제 없음(defaults 불변) → search/ask/chunk 출력
byte-identical 확인 (parity gate u2-surface IDENTICAL ✓).

docs: README KEBAB_* 항목을 실제 노출 키 목록으로 교체.
      SMOKE.md config 예시를 ~30개 공통 키로 슬림화 + env 설명 갱신.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012Mc6W1fgsrbFKTsqA6P8La
This commit is contained in:
altair823
2026-06-24 11:21:11 +00:00
parent da323af87b
commit bf7769cf5a
4 changed files with 119 additions and 297 deletions
Download Patch File Download Diff File Expand all files Collapse all files

13
crates/kebab-config/tests/pdf_ocr.rs
View File

@@ -63,14 +63,17 @@ fn pdf_ocr_defaults_off_with_qwen_3b() {
assert_eq!(cfg.ingest.pdf.ocr.lang_hint.as_deref(), Some("kor"));
}
// Test 3: env var override — pdf-only keys + shared engine knob.
// Test 3: env var override — pdf enabled + shared engine knob.
// v5: `model` moved to the shared `KEBAB_OCR_MODEL` (sets both mediums);
// `enabled`/`always_on`/`valid_ratio_threshold` stay pdf-specific.
// `enabled` stays per-medium addressable.
// KEBAB_PDF_OCR_ALWAYS_ON and KEBAB_PDF_OCR_VALID_RATIO_THRESHOLD removed
// from env surface (config-only) in Unit 2 surface trim.
#[test]
fn pdf_ocr_env_overrides() {
let mut env: HashMap<String, String> = HashMap::new();
env.insert("KEBAB_PDF_OCR_ENABLED".to_string(), "true".to_string());
env.insert("KEBAB_OCR_MODEL".to_string(), "qwen2.5vl:7b".to_string());
// always_on and valid_ratio_threshold arms gone — silently ignored below.
env.insert("KEBAB_PDF_OCR_ALWAYS_ON".to_string(), "true".to_string());
env.insert(
"KEBAB_PDF_OCR_VALID_RATIO_THRESHOLD".to_string(),
@@ -81,8 +84,10 @@ fn pdf_ocr_env_overrides() {
assert!(cfg.ingest.pdf.ocr.enabled);
assert_eq!(cfg.ingest.pdf.ocr.model, "qwen2.5vl:7b");
assert!(cfg.ingest.pdf.ocr.always_on);
assert!((cfg.ingest.pdf.ocr.valid_ratio_threshold - 0.75).abs() < 1e-6);
// always_on arm gone — stays at default (false).
assert!(!cfg.ingest.pdf.ocr.always_on);
// valid_ratio_threshold arm gone — stays at default (0.5).
assert!((cfg.ingest.pdf.ocr.valid_ratio_threshold - 0.5).abs() < 1e-6);
// 다른 env var 가 default 보존
assert_eq!(cfg.ingest.pdf.ocr.engine, "ollama-vision");