refactor(config): ExpansionCfg + [ingest.expansion] 제거

IngestExpansionCfg struct + IngestCfg.expansion 필드 + Default +
KEBAB_INGEST_EXPANSION_* env 파싱 + 테스트 제거. migrate.rs 의
ingest.expansion 섹션 주석 제거 — config migrate 테스트는 ingest.code 앵커로
정정(forward-compat: 기존 [ingest.expansion] 섹션은 serde 가 무시).

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
This commit is contained in:
2026-06-02 21:36:56 +00:00
parent ecaf224381
commit ec96648956
2 changed files with 10 additions and 101 deletions

View File

@@ -606,8 +606,6 @@ impl UiCfg {
#[serde(default)] #[serde(default)]
pub struct IngestCfg { pub struct IngestCfg {
pub code: IngestCodeCfg, pub code: IngestCodeCfg,
#[serde(default)]
pub expansion: IngestExpansionCfg,
} }
/// p10-1A-1: settings for the code ingest pipeline. All fields have /// p10-1A-1: settings for the code ingest pipeline. All fields have
@@ -648,34 +646,6 @@ impl Default for IngestCodeCfg {
} }
} }
/// doc-side expansion config. Default: disabled (requires explicit opt-in).
#[derive(Clone, Debug, PartialEq, Serialize, Deserialize)]
#[serde(default)]
pub struct IngestExpansionCfg {
/// Whether doc-side alias expansion is enabled during ingest.
pub enabled: bool,
/// Ollama model used for alias generation (empty = use LLM default).
pub model: String,
/// Maximum aliases generated per chunk.
pub max_aliases_per_chunk: usize,
/// Prompt template version tag.
pub prompt_version: String,
/// Whether alias embeddings are stored as separate dense vectors.
pub embed_aliases: bool,
}
impl Default for IngestExpansionCfg {
fn default() -> Self {
Self {
enabled: false,
model: String::new(),
max_aliases_per_chunk: 8,
prompt_version: "expansion-v1".to_string(),
embed_aliases: false,
}
}
}
impl Config { impl Config {
/// Defaults per design §6.4. /// Defaults per design §6.4.
pub fn defaults() -> Self { pub fn defaults() -> Self {
@@ -1166,25 +1136,6 @@ impl Config {
self.pdf.ocr.lang_hint = if v.is_empty() { None } else { Some(v.clone()) }; self.pdf.ocr.lang_hint = if v.is_empty() { None } else { Some(v.clone()) };
} }
// ingest.expansion
"KEBAB_INGEST_EXPANSION_ENABLED" => {
self.ingest.expansion.enabled = parse_bool(v);
}
"KEBAB_INGEST_EXPANSION_MODEL" => {
self.ingest.expansion.model = v.clone();
}
"KEBAB_INGEST_EXPANSION_MAX_ALIASES" => {
if let Ok(n) = v.parse::<usize>() {
self.ingest.expansion.max_aliases_per_chunk = n;
}
}
"KEBAB_INGEST_EXPANSION_PROMPT_VERSION" => {
self.ingest.expansion.prompt_version = v.clone();
}
"KEBAB_INGEST_EXPANSION_EMBED_ALIASES" => {
self.ingest.expansion.embed_aliases = parse_bool(v);
}
// Unknown KEBAB_* keys are silently ignored — see // Unknown KEBAB_* keys are silently ignored — see
// `env_unknown_key_is_ignored` test. // `env_unknown_key_is_ignored` test.
_ => {} _ => {}
@@ -1913,41 +1864,6 @@ max_context_tokens = 8000
assert_eq!(cfg.ingest.code.max_file_bytes, 524_288); assert_eq!(cfg.ingest.code.max_file_bytes, 524_288);
} }
#[test]
fn expansion_defaults_off() {
let cfg = Config::defaults();
assert!(!cfg.ingest.expansion.enabled);
assert_eq!(cfg.ingest.expansion.max_aliases_per_chunk, 8);
assert_eq!(cfg.ingest.expansion.prompt_version, "expansion-v1");
}
#[test]
fn expansion_env_override() {
let mut env = HashMap::new();
env.insert("KEBAB_INGEST_EXPANSION_ENABLED".into(), "true".into());
env.insert("KEBAB_INGEST_EXPANSION_MODEL".into(), "gemma3:4b".into());
env.insert("KEBAB_INGEST_EXPANSION_MAX_ALIASES".into(), "12".into());
env.insert("KEBAB_INGEST_EXPANSION_PROMPT_VERSION".into(), "expansion-v2".into());
let c = Config::defaults().apply_env(&env);
assert!(c.ingest.expansion.enabled);
assert_eq!(c.ingest.expansion.model, "gemma3:4b");
assert_eq!(c.ingest.expansion.max_aliases_per_chunk, 12);
assert_eq!(c.ingest.expansion.prompt_version, "expansion-v2");
}
#[test]
fn embed_aliases_defaults_off() {
let cfg = Config::defaults();
assert!(!cfg.ingest.expansion.embed_aliases);
}
#[test]
fn embed_aliases_env_override() {
let mut env = HashMap::new();
env.insert("KEBAB_INGEST_EXPANSION_EMBED_ALIASES".into(), "true".into());
let c = Config::defaults().apply_env(&env);
assert!(c.ingest.expansion.embed_aliases);
}
} }
#[cfg(test)] #[cfg(test)]

View File

@@ -15,7 +15,7 @@ pub const CURRENT_SCHEMA_VERSION: u32 = 2;
#[derive(Clone, Debug, PartialEq, serde::Serialize)] #[derive(Clone, Debug, PartialEq, serde::Serialize)]
pub struct MigrationChange { pub struct MigrationChange {
pub kind: ChangeKind, pub kind: ChangeKind,
/// dotted path, 예: `ingest.expansion`, `workspace.include`. /// dotted path, 예: `ingest.code`, `workspace.include`.
pub path: String, pub path: String,
/// 사람·wire 용 한 줄 설명. /// 사람·wire 용 한 줄 설명.
pub detail: String, pub detail: String,
@@ -83,7 +83,6 @@ fn section_comment(path: &str) -> Option<&'static str> {
"ui" => "# TUI 팔레트·role 스타일.", "ui" => "# TUI 팔레트·role 스타일.",
"ingest" => "# ingest 정책(code skip 등).", "ingest" => "# ingest 정책(code skip 등).",
"ingest.code" => "# code ingest skip 정책(.gitignore 자동 honor).", "ingest.code" => "# code ingest skip 정책(.gitignore 자동 honor).",
"ingest.expansion" => "# doc-side 별칭 확장(기본 off). 패러프레이즈 강건성↑, LLM 비용 큼.",
"pdf" => "# PDF ingest. scanned PDF OCR 은 기본 off(page 당 cost).", "pdf" => "# PDF ingest. scanned PDF OCR 은 기본 off(page 당 cost).",
"pdf.ocr" => "# scanned PDF page-단위 OCR(기본 off).", "pdf.ocr" => "# scanned PDF page-단위 OCR(기본 off).",
"logging" => "# ingest 로그(기본 on, ~/.local/state/kebab/logs).", "logging" => "# ingest 로그(기본 on, ~/.local/state/kebab/logs).",
@@ -259,7 +258,7 @@ mod tests {
// `[pdf]` 등은 안 나오고 `[pdf.ocr]` 같은 하위 테이블만 직렬화된다. // `[pdf]` 등은 안 나오고 `[pdf.ocr]` 같은 하위 테이블만 직렬화된다.
for section in [ for section in [
"[workspace]", "[workspace]",
"[ingest.expansion]", "[ingest.code]",
"[pdf.ocr]", "[pdf.ocr]",
"[logging]", "[logging]",
"[ui]", "[ui]",
@@ -273,8 +272,8 @@ mod tests {
#[test] #[test]
fn reconcile_adds_missing_section_preserving_user_values_and_comments() { fn reconcile_adds_missing_section_preserving_user_values_and_comments() {
// ingest 는 code 만 있고 expansion 누락(v0.21.0 동기 시나리오), // ingest 통째 누락(→ ingest.code 추가), logging 통째 누락,
// logging 통째 누락, score 는 사용자가 바꿈, 주석 보유. // default_k 는 사용자가 바꿈, 주석 보유.
let user_text = "\ let user_text = "\
schema_version = 1 schema_version = 1
@@ -283,9 +282,6 @@ root = \"/my/notes\" # 내 워크스페이스
[search] [search]
default_k = 25 default_k = 25
[ingest.code]
skip_generated_header = true
"; ";
let mut user: DocumentMut = user_text.parse().unwrap(); let mut user: DocumentMut = user_text.parse().unwrap();
let reference = annotated_default_document(); let reference = annotated_default_document();
@@ -294,25 +290,22 @@ skip_generated_header = true
reconcile(&ref_tbl, user.as_table_mut(), "", &mut changes); reconcile(&ref_tbl, user.as_table_mut(), "", &mut changes);
let out = user.to_string(); let out = user.to_string();
// 부분 존재하는 [ingest] 에 expansion 만 주석과 함께 추가. // 누락된 [ingest.code] 가 주석과 함께 추가.
assert!(out.contains("[ingest.expansion]"), "expansion not added:\n{out}"); assert!(out.contains("[ingest.code]"), "ingest.code not added:\n{out}");
// 통째 누락된 logging 추가. // 통째 누락된 logging 추가.
assert!(out.contains("[logging]"), "logging not added"); assert!(out.contains("[logging]"), "logging not added");
// 사용자 값/주석/기존 섹션 보존. // 사용자 값/주석/기존 섹션 보존.
assert!(out.contains("root = \"/my/notes\"")); assert!(out.contains("root = \"/my/notes\""));
assert!(out.contains("# 내 워크스페이스")); assert!(out.contains("# 내 워크스페이스"));
assert!(out.contains("default_k = 25")); assert!(out.contains("default_k = 25"));
assert!(out.contains("skip_generated_header = true"));
// 새 섹션 주석 부착. // 새 섹션 주석 부착.
assert!(out.contains("doc-side 별칭")); assert!(out.contains("code ingest skip 정책"));
// 부분 존재 부모로 재귀해 leaf 경로를 기록. // 통째 누락 부모는 부모 경로로 한 번 기록.
assert!( assert!(
changes changes
.iter() .iter()
.any(|c| c.kind == ChangeKind::AddedSection && c.path == "ingest.expansion"), .any(|c| c.kind == ChangeKind::AddedSection && c.path == "ingest")
"changes: {changes:?}"
); );
// 통째 누락 부모는 부모 경로로 한 번 기록.
assert!( assert!(
changes changes
.iter() .iter()
@@ -381,7 +374,7 @@ include = [\"*.md\"]
assert_eq!(outcome.to_schema_version, CURRENT_SCHEMA_VERSION); assert_eq!(outcome.to_schema_version, CURRENT_SCHEMA_VERSION);
assert!(outcome.changed()); assert!(outcome.changed());
assert!(!outcome.new_text.contains("include")); assert!(!outcome.new_text.contains("include"));
assert!(outcome.new_text.contains("[ingest.expansion]")); assert!(outcome.new_text.contains("[ingest.code]"));
assert_eq!(read_schema_version(&outcome.new_text), CURRENT_SCHEMA_VERSION); assert_eq!(read_schema_version(&outcome.new_text), CURRENT_SCHEMA_VERSION);
let again = migrate_document(&outcome.new_text); let again = migrate_document(&outcome.new_text);