codemem-embeddings 0.9.1

Candle-based embedding service for Codemem using BAAI/bge-base-en-v1.5

Coverage
96%
24 out of 25 items documented0 out of 14 items with examples
Size
Source code size: 209.41 kB This is the summed size of all the files inside the crates.io package for this release.
Documentation size: 1.77 MB This is the summed size of all files generated by rustdoc for all configured targets
Ø build duration
this release: 2m 5s Average build duration of successful builds.
all releases: 2m 10s Average build duration of successful builds in releases after 2024-10-23.
Links
Homepage
cogniplex/codemem
4 0 0
crates.io
Dependencies
Versions
Owners

codemem-embeddings

Pluggable embedding providers with Candle (default), Ollama, and OpenAI-compatible backends.

Overview

Defines the EmbeddingProvider trait and a from_env() factory that selects the provider at runtime via environment variables.

Providers

Provider	Model	Use Case
`candle` (default)	BAAI/bge-base-en-v1.5 (768-dim)	Fully offline, pure Rust ML, Metal/CUDA GPU
`ollama`	nomic-embed-text	Local server, swap models freely
`openai`	text-embedding-3-small	OpenAI, Voyage AI, Together, Azure, any compatible API

Configuration

Variable	Default
`CODEMEM_EMBED_PROVIDER`	`candle`
`CODEMEM_EMBED_MODEL`	provider default
`CODEMEM_EMBED_URL`	provider default
`CODEMEM_EMBED_API_KEY`	reads `OPENAI_API_KEY`
`CODEMEM_EMBED_DIMENSIONS`	`768`

Key Features

CachedProvider wrapper adds LRU cache (10K entries) to any provider
embed_batch() for efficient bulk embedding
Safe concurrency — all lock acquisitions return CodememError::LockPoisoned instead of panicking