feat(config,app): embedding provider=ollama 배선 + endpoint, version 0.26.0

kebab-config: EmbeddingModelCfg.endpoint: Option<String>(serde default, ollama용,
None→models.llm.endpoint 폴백) + provider 문서에 ollama + env
KEBAB_MODELS_EMBEDDING_ENDPOINT. kebab-app embedder(): provider match 에 ollama
분기(facade 경유). workspace member += kebab-embed-ollama, app dep 추가.
version 0.25.0 → 0.26.0(minor, +Cargo.lock) — 신규 임베딩 백엔드/모델은 CLAUDE.md
§Release 의 surface 변경 트리거.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
This commit is contained in:
2026-06-03 04:59:23 +00:00
parent cbcae69abf
commit 72c99c452c
5 changed files with 73 additions and 33 deletions

63
Cargo.lock generated
View File

@@ -4724,7 +4724,7 @@ dependencies = [
[[package]]
name = "kebab-app"
version = "0.25.0"
version = "0.26.0"
dependencies = [
"anyhow",
"base64 0.22.1",
@@ -4739,6 +4739,7 @@ dependencies = [
"kebab-embed",
"kebab-embed-candle",
"kebab-embed-local",
"kebab-embed-ollama",
"kebab-llm",
"kebab-llm-local",
"kebab-nli",
@@ -4771,7 +4772,7 @@ dependencies = [
[[package]]
name = "kebab-chunk"
version = "0.25.0"
version = "0.26.0"
dependencies = [
"anyhow",
"blake3",
@@ -4789,7 +4790,7 @@ dependencies = [
[[package]]
name = "kebab-cli"
version = "0.25.0"
version = "0.26.0"
dependencies = [
"anyhow",
"clap",
@@ -4810,7 +4811,7 @@ dependencies = [
[[package]]
name = "kebab-config"
version = "0.25.0"
version = "0.26.0"
dependencies = [
"anyhow",
"dirs 5.0.1",
@@ -4826,7 +4827,7 @@ dependencies = [
[[package]]
name = "kebab-core"
version = "0.25.0"
version = "0.26.0"
dependencies = [
"anyhow",
"blake3",
@@ -4840,7 +4841,7 @@ dependencies = [
[[package]]
name = "kebab-embed"
version = "0.25.0"
version = "0.26.0"
dependencies = [
"anyhow",
"blake3",
@@ -4854,7 +4855,7 @@ dependencies = [
[[package]]
name = "kebab-embed-candle"
version = "0.25.0"
version = "0.26.0"
dependencies = [
"anyhow",
"candle-core",
@@ -4864,6 +4865,7 @@ dependencies = [
"kebab-config",
"kebab-core",
"kebab-embed-local",
"kebab-embed-ollama",
"rayon",
"serde_json",
"tempfile",
@@ -4873,7 +4875,7 @@ dependencies = [
[[package]]
name = "kebab-embed-local"
version = "0.25.0"
version = "0.26.0"
dependencies = [
"anyhow",
"fastembed",
@@ -4884,9 +4886,24 @@ dependencies = [
"tracing",
]
[[package]]
name = "kebab-embed-ollama"
version = "0.26.0"
dependencies = [
"anyhow",
"kebab-config",
"kebab-core",
"reqwest 0.12.28",
"serde",
"serde_json",
"tokio",
"tracing",
"wiremock",
]
[[package]]
name = "kebab-eval"
version = "0.25.0"
version = "0.26.0"
dependencies = [
"anyhow",
"kebab-app",
@@ -4905,7 +4922,7 @@ dependencies = [
[[package]]
name = "kebab-llm"
version = "0.25.0"
version = "0.26.0"
dependencies = [
"anyhow",
"kebab-core",
@@ -4914,7 +4931,7 @@ dependencies = [
[[package]]
name = "kebab-llm-local"
version = "0.25.0"
version = "0.26.0"
dependencies = [
"anyhow",
"kebab-config",
@@ -4931,7 +4948,7 @@ dependencies = [
[[package]]
name = "kebab-mcp"
version = "0.25.0"
version = "0.26.0"
dependencies = [
"anyhow",
"kebab-app",
@@ -4949,7 +4966,7 @@ dependencies = [
[[package]]
name = "kebab-nli"
version = "0.25.0"
version = "0.26.0"
dependencies = [
"anyhow",
"hf-hub",
@@ -4964,7 +4981,7 @@ dependencies = [
[[package]]
name = "kebab-parse-code"
version = "0.25.0"
version = "0.26.0"
dependencies = [
"anyhow",
"gix",
@@ -4987,7 +5004,7 @@ dependencies = [
[[package]]
name = "kebab-parse-image"
version = "0.25.0"
version = "0.26.0"
dependencies = [
"ab_glyph",
"anyhow",
@@ -5011,7 +5028,7 @@ dependencies = [
[[package]]
name = "kebab-parse-md"
version = "0.25.0"
version = "0.26.0"
dependencies = [
"anyhow",
"kebab-core",
@@ -5028,7 +5045,7 @@ dependencies = [
[[package]]
name = "kebab-parse-pdf"
version = "0.25.0"
version = "0.26.0"
dependencies = [
"anyhow",
"blake3",
@@ -5043,7 +5060,7 @@ dependencies = [
[[package]]
name = "kebab-rag"
version = "0.25.0"
version = "0.26.0"
dependencies = [
"anyhow",
"blake3",
@@ -5065,7 +5082,7 @@ dependencies = [
[[package]]
name = "kebab-search"
version = "0.25.0"
version = "0.26.0"
dependencies = [
"anyhow",
"globset",
@@ -5084,7 +5101,7 @@ dependencies = [
[[package]]
name = "kebab-source-fs"
version = "0.25.0"
version = "0.26.0"
dependencies = [
"anyhow",
"blake3",
@@ -5102,7 +5119,7 @@ dependencies = [
[[package]]
name = "kebab-store-sqlite"
version = "0.25.0"
version = "0.26.0"
dependencies = [
"anyhow",
"blake3",
@@ -5122,7 +5139,7 @@ dependencies = [
[[package]]
name = "kebab-store-vector"
version = "0.25.0"
version = "0.26.0"
dependencies = [
"anyhow",
"arrow",
@@ -5146,7 +5163,7 @@ dependencies = [
[[package]]
name = "kebab-tui"
version = "0.25.0"
version = "0.26.0"
dependencies = [
"anyhow",
"crossterm",

View File

@@ -12,6 +12,7 @@ members = [
"crates/kebab-embed",
"crates/kebab-embed-local",
"crates/kebab-embed-candle",
"crates/kebab-embed-ollama",
"crates/kebab-llm",
"crates/kebab-llm-local",
"crates/kebab-rag",
@@ -31,7 +32,7 @@ edition = "2024"
rust-version = "1.85"
license = "MIT OR Apache-2.0"
repository = "https://github.com/altair823/kebab"
version = "0.25.0" # v0.25.0 — doc-side expansion(별칭) 기능 완전 제거: Chunk.aliases / expansion.rs / IngestExpansionCfg / alias lexical arm / expansion_progress wire kind 제거, 신규 마이그레이션 V013 이 chunk_aliases_fts + chunks.aliases DROP. AssetTimings.expansion_ms 는 wire 호환 위해 값 0 유지. 별칭 default-off 였어 사용자 체감 0. — CLAUDE.md §Release
version = "0.26.0" # v0.26.0 — arctic-embed-l-v2.0 임베더 통합: kebab-embed-candle 다중 모델 레지스트리(e5 mean + arctic CLS, 모델별 pooling/prefix 분기) + 신규 kebab-embed-ollama 크레이트(provider="ollama", POST /api/embed, L2 정규화, batch+fail-soft). config models.embedding.provider 에 "ollama" 추가 + endpoint: Option<String>. 기본 동작 불변(provider=fastembed e5), arctic 은 opt-in, embedding_version cascade(arctic-cls / ollama:{model} 태그). — CLAUDE.md §Release
# pre-v0.18 workspace-wide cleanup: enable clippy::pedantic group with
# intentional allow-list. The allowed lints are either cosmetic (doc style),

View File

@@ -19,6 +19,7 @@ kebab-search = { path = "../kebab-search" }
kebab-embed = { path = "../kebab-embed" }
kebab-embed-local = { path = "../kebab-embed-local" }
kebab-embed-candle = { path = "../kebab-embed-candle" }
kebab-embed-ollama = { path = "../kebab-embed-ollama" }
kebab-llm = { path = "../kebab-llm" }
kebab-llm-local = { path = "../kebab-llm-local" }
kebab-rag = { path = "../kebab-rag" }

View File

@@ -45,6 +45,7 @@ use kebab_core::{
};
use kebab_embed_candle::CandleEmbedder;
use kebab_embed_local::FastembedEmbedder;
use kebab_embed_ollama::OllamaEmbedder;
use kebab_llm_local::OllamaLanguageModel;
use kebab_parse_code::{
CAstExtractor, CppAstExtractor, GoAstExtractor, JavaAstExtractor, JavascriptAstExtractor,
@@ -834,11 +835,13 @@ impl App {
if let Some(e) = self.embedder.get() {
return Ok(Some(e.clone()));
}
// Provider branch (Track 1 spec §3). `embeddings_disabled()` above
// already handled `"none"`; here we route the live providers.
// `fastembed`/`onnx`/(empty) keep the default onnxruntime path
// (vectors unchanged — `embedding_version` is preserved); `candle`
// selects the pure-Rust NUMA-safe backend.
// Provider branch (Track 1 spec §3 + arctic-embedder spec). The
// `embeddings_disabled()` check above already handled `"none"`; here we
// route the live providers. `fastembed`/`onnx`/(empty) keep the default
// onnxruntime path (vectors unchanged — `embedding_version` is
// preserved); `candle` selects the pure-Rust NUMA-safe backend (e5 or
// arctic via its model registry); `ollama` offloads to a remote
// `/api/embed` daemon.
let provider = self.config.models.embedding.provider.as_str();
let emb: Arc<dyn Embedder + Send + Sync> = match provider {
"fastembed" | "onnx" | "" => Arc::new(
@@ -847,10 +850,13 @@ impl App {
"candle" => Arc::new(
CandleEmbedder::new(&self.config).context("kb-app: load CandleEmbedder")?,
),
"ollama" => Arc::new(
OllamaEmbedder::new(&self.config).context("kb-app: load OllamaEmbedder")?,
),
other => {
return Err(anyhow!(
"kb-app: unknown embedding provider {other:?}; expected one of \
`fastembed` (default), `candle`, or `none` (lexical-only)"
`fastembed` (default), `candle`, `ollama`, or `none` (lexical-only)"
));
}
};

View File

@@ -155,9 +155,10 @@ impl NliCfg {
#[derive(Clone, Debug, PartialEq, Serialize, Deserialize)]
pub struct EmbeddingModelCfg {
/// `fastembed` (default, onnxruntime) or `candle` (pure-Rust,
/// NUMA-safe). `none` disables embeddings (lexical-only). Unknown
/// values error at embedder construction.
/// `fastembed` (default, onnxruntime), `candle` (pure-Rust, NUMA-safe),
/// or `ollama` (remote HTTP embedding endpoint). `none` disables
/// embeddings (lexical-only). Unknown values error at embedder
/// construction.
pub provider: String,
pub model: String,
pub version: String,
@@ -170,6 +171,13 @@ pub struct EmbeddingModelCfg {
/// provider. Defaulted on load so pre-0.22 config files still parse.
#[serde(default)]
pub num_threads: u32,
/// HTTP endpoint for the `ollama` embedding provider (e.g.
/// `"http://127.0.0.1:11434"`). `None` (or a missing key in TOML) means
/// "fall back to `models.llm.endpoint`" — same convention as the OCR /
/// vision endpoints. Ignored by the `fastembed` / `candle` providers.
/// Defaulted on load so pre-0.26 config files still parse.
#[serde(default)]
pub endpoint: Option<String>,
}
#[derive(Clone, Debug, PartialEq, Serialize, Deserialize)]
@@ -688,6 +696,7 @@ impl Config {
dimensions: 1024,
batch_size: 64,
num_threads: 0,
endpoint: None,
},
llm: LlmCfg {
provider: "ollama".to_string(),
@@ -950,6 +959,12 @@ impl Config {
self.models.embedding.num_threads = n;
}
}
"KEBAB_MODELS_EMBEDDING_ENDPOINT" => {
// Empty value → None (= fall back to models.llm.endpoint),
// mirroring the OCR endpoint override semantics.
self.models.embedding.endpoint =
if v.is_empty() { None } else { Some(v.clone()) };
}
// models.llm
"KEBAB_MODELS_LLM_PROVIDER" => self.models.llm.provider = v.clone(),