Blomios
32398827fb
Remplace les StubEmbedder pour les stratégies localServer/api/localOnnx par de vrais moteurs, chacun derrière une feature cargo off-by-default — la posture fondatrice « rien d'imposé, zéro dépendance » (défaut none → rappel naïf) reste byte-for-byte inchangée. C1a (feature vector-http, reqwest rustls optional): - HttpEmbedder couvrant localServer (Ollama/llama.cpp) et api (OpenAI/Voyage…), payload OpenAI-compatible /v1/embeddings, ordre restauré par index, bearer token lu via env var (jamais en clair), timeout client 30s. - detect_ollama() pour la détection de l'existant (C3). C1b (feature vector-onnx, fastembed v5 optional): - OnnxEmbedder en-process (e5-small, dim 384), init paresseuse + spawn_blocking, cache modèle sous <app_data>/embedders/onnx — aucun download au build ni au first-run, uniquement à la demande au 1er embed. - Catalogue RECOMMENDED_ONNX_MODELS + ONNX_CACHE_SUBDIR + onnx_model_is_cached exposés (sans feature) pour la config (C2) et la popup (C3). embedder_from_profile(profile, onnx_cache_dir) dispatche feature-gated ; sans la feature, retombe sur StubEmbedder (Unsupported) → fallback naïf via AdaptiveMemoryRecall. Composition root (build_memory_recall) propage le cache dir. Tests: 10 HTTP + 6 ONNX (dont 2 #[ignore] download réel) + 26 vectoriels, verts en défaut, --features vector-http et --features vector-onnx. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
43 lines
2.2 KiB
TOML
43 lines
2.2 KiB
TOML
[package]
|
|
name = "infrastructure"
|
|
version = "0.1.0"
|
|
edition.workspace = true
|
|
license.workspace = true
|
|
rust-version.workspace = true
|
|
description = "IdeA — infrastructure layer: concrete adapters implementing the domain ports (fs, event bus, clock, id)."
|
|
|
|
[dependencies]
|
|
domain = { workspace = true }
|
|
# The orchestrator filesystem watcher (driving adapter, ARCHITECTURE §14.3) drives
|
|
# the application's `OrchestratorService`; infrastructure may depend on application.
|
|
application = { workspace = true }
|
|
# `process` (additive) powers LocalProcessSpawner; the workspace baseline keeps
|
|
# rt/macros/sync/fs/io-util.
|
|
tokio = { workspace = true, features = ["process", "time"] }
|
|
uuid = { workspace = true }
|
|
async-trait = { workspace = true }
|
|
serde = { workspace = true }
|
|
serde_json = { workspace = true }
|
|
portable-pty = "0.9"
|
|
git2 = { workspace = true }
|
|
# Filesystem change notifications used to *wake* the orchestrator poll loop early
|
|
# (the poll loop remains the robust cross-platform correctness guarantee).
|
|
notify = "6"
|
|
# Optional HTTP client for the real `localServer`/`api` embedders (LOT C1a,
|
|
# §14.5.3). `default-features = false` + `rustls-tls` keeps it OpenSSL-free so the
|
|
# AppImage stays portable; pulled in *only* under the `vector-http` feature so the
|
|
# zero-dependency default (`none` strategy) compiles nothing extra.
|
|
reqwest = { version = "0.12", default-features = false, features = ["json", "rustls-tls"], optional = true }
|
|
# Optional in-process ONNX embedder (LOT C1b, §14.5.3). rustls all the way down
|
|
# (HF model download + ort binaries fetched at build time, NOT load-dynamic) so the
|
|
# AppImage stays self-contained and OpenSSL-free. Pulled in *only* under the
|
|
# `vector-onnx` feature; the zero-dependency default compiles nothing extra.
|
|
fastembed = { version = "5", default-features = false, features = ["hf-hub-rustls-tls", "ort-download-binaries-rustls-tls"], optional = true }
|
|
|
|
[features]
|
|
# Real HTTP-backed embedders (`localServer` Ollama/llama.cpp, `api` OpenAI/Voyage…).
|
|
# OFF by default: the founding posture is `none` ⇒ naïve recall, zero dependency.
|
|
vector-http = ["dep:reqwest"]
|
|
# Real in-process ONNX embedder (`localOnnx`). OFF by default, same posture.
|
|
vector-onnx = ["dep:fastembed"]
|