Feature request: AdaptiveEmbedder Rust binding under crates/
Nobody has claimed this yet.
- Dominant language
- Rust
- Stars
- 4.5k
- Forks
- 603
- Avg merge
- 23h 32m
- Merged PRs (30d)
- 59
Description
Context
A downstream Rust-only consumer (SONA-Trader v2, ADR-002 Rust-only stack) consumes RuVector via a pinned vendor symlink. During empirical vector-tuning prep (Sprint 15.7), we walked vendor/ruvector/crates/ looking for an AdaptiveEmbedder Rust crate and found it absent.
What we found
The AdaptiveEmbedder API exists in the npm package only:
npm/packages/ruvector/src/core/adaptive-embedder.jsnpm/packages/ruvector/src/core/adaptive-embedder.d.ts
Closest hits under crates/:
crates/ruvector-core/src/advanced_features/matryoshka.rs— Matryoshka adaptive-dim search (different concept)crates/ruvllm/src/bitnet/rlm_embedder.rs— "Domain-adaptive" RLM embedder (different concept)
Ask
Expose AdaptiveEmbedder as a Rust crate under crates/ so Rust-only consumers can use it without a Node sidecar.
Current workaround
Embedding-model sweep restricted to 5 alternative candidates (fastembed bge-small, candle MiniLM-L6, OpenAI text-embedding-3-{small,large}, Voyage finance-2, Cohere embed-v3.0).
Happy to test a draft binding against a deterministic 10K-pair labeled eval-set + LHS-24 sweep we use downstream.
Contributor guide
No contributing guide indexed for this repository
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start by reading npm/packages/ruvector/src/core/adaptive-embedder.js and adaptive-embedder.d.ts to define the existing API, then compare the related Rust implementations in crates/ruvector-core/src/advanced_features/matryoshka.rs and crates/ruvllm/src/bitnet/rlm_embedder.rs. Determine the appropriate crate boundary and Rust-facing API. Done means Rust-only consumers can use AdaptiveEmbedder under crates/ and the draft can be evaluated against the stated deterministic 10K-pair set and LHS-24 sweep.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- javascript, rust
- Domain
- ai, machine-learning
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Quiet
- Clarity
- Mostly clear
- Newbie friendliness
- 45/100