chore(release): merge v0.15.0 to main - #551
Merged
Merged
Conversation
* feat(brain): opt-in vecq vector store backend Add `brain.vector_store: usearch|vecq` (default usearch). vecq-core 0.2.0 replaces the HNSW store with a 4-bit quantized brute-force scan: pure Rust (no C++ FFI), bit-identical deterministic results, ~6x smaller index, 14x faster builds — query latency stays sub-ms at cora's typical index sizes and the recall trade is absorbed by Brain's RRF fusion (#542). - CodeVectorIndex gains an enum backend (Usearch | Vecq); every backend owns its file extension (.usearch / .vecq) so switching stores never misreads the other's format - config applied at every Brain entry point (index, watch, brain cmd, MCP brain_search) so CLI and MCP cannot silently diverge - vecq insert/remove map symbol ids to native u64 keys (tombstones); search converts similarity to the cosine-distance contract Known limitation (documented + tested): vecq-core 0.2.0 from_bytes does not serialize the keyed map (vecq#32), so reload rebuilds fresh instead of serving a silently-empty index; `cora index` re-embeds. Persistence lands when vecq ships key serialization. Regression tests: backend roundtrip (insert/replace/remove/search + distance conversion), per-backend persistence extensions, lenient store-kind parsing. Signed-off-by: ajianaz <ajianaz@users.noreply.github.com> * fix(brain): wire brain section config, self-heal vecq reload, dedupe config load Three fixes from PR #543 review: 1. Config wiring: .cora.yaml 'brain:' section was never parsed — CoraFile had no brain field, so vector_store: vecq silently fell back to usearch and embedding mode was ignored. Add BrainSection + field-wise merge_into (matches other sections). 2. Vecq reload self-heal: a reloaded vecq index is rebuilt empty (upstream vecq#32 lacks key serialization) and arrived with dirty=false, so a no-change 'cora index' run saved an empty index over the stored one — permanently losing vector signal. Now reload marks the index dirty, embed_project clears stale embed fingerprints when it sees dirty+empty, and cora index forces re-embed when the on-disk vecq file needs a rebuild. 3. Brain command: load_config was called twice; fold into one load shared by resolve_backend and apply_config_store. Verified: 937/937 tests pass; e2e with CODECORA_HOME sandbox — vector_store: vecq now creates cora_index.vecq config now works, and a no-change re-run logs 'cleared=2, re-embedding all symbols' instead of silently persisting an empty index. --------- Signed-off-by: ajianaz <ajianaz@users.noreply.github.com> Co-authored-by: ajianaz <ajianaz@users.noreply.github.com>
VECTOR_CACHE was only populated by embed_project (i.e. 'cora index'), so a fresh search process — 'cora brain' CLI or MCP brain_search — always saw an empty cache and the vector signal never fired; results silently degraded to FTS + graph only. Verified on develop and on the vecq PR branch with the same sandboxed e2e run. vector_search now lazy-loads the on-disk index once per process via ensure_vector_cache (double-checked read/write locking; a concurrent double load is idempotent). Load failures degrade gracefully with a warn log. Also guard dimension mismatch (embedding provider switched since the index was built) — previously a panic inside the backend. Note: the vecq backend still degrades to FTS-only in fresh search processes — that is the documented upstream limitation (vecq#32, no key serialization), tracked separately. Regression tests (tests/brain_search_process.rs): fresh-process brain search must emit the vector signal (default usearch backend), and brain.vector_store: vecq must produce a .vecq file — both run with isolated CODECORA_HOME sandboxes. Co-authored-by: ajianaz <ajianaz@users.noreply.github.com>
…rain.vector_bits knob (#548) Upgrade the opt-in vecq store to vecq-core 0.3.0 and close #547. - Keyed persistence (upstream vecq#32 closed, format v1.3+): symbol-id keys survive reload. load_or_create_vecq now deserializes for real and keeps a healthy keyed index; legacy keyless files, corrupt files, and dim mismatches rebuild dirty so the next re-embeds. A dim guard is mandatory: add_keyed panics on wrong-dim vectors. - brain.vector_bits config (residual|4|5|6, lenient, default residual): residual is best recall@10 at 4-bit scan speed in a recall study on cora's own embeddings (ahead of plain 5-bit at 1k/5k/13k scales). A width change rebuilds the index once so the config takes effect; vecq_file_needs_rebuild is width-aware. - BrainConfig gets a manual Default (derived Default produced empty strings, inconsistent with the serde defaults). - Cross-project heal: the vector index is one shared file; a rebuild (width/dims/legacy/corrupt) now clears embed fingerprints for ALL projects, not just the active one, and the usearch dims-mismatch deletion path joins the same heal (pre-existing fingerprint leak). Tests: keyed roundtrip, legacy keyless rebuild, corrupt rebuild, dim-mismatch rebuild, width-switch rebuild, roundtrip across all four widths, lenient parsing, config merge. 947 pass incl. the fresh-process integration tests from #546. Closes #547 Co-authored-by: ajianaz <ajianaz@users.noreply.github.com>
Add a Self-Upgrade section to docs/cli-reference.md (--check, default flow with SHA-256 checksum verification, -y for CI), a README install/ upgrade note and a command-table row, per #541. Correction vs the issue body: the background update check is disabled via the CORA_NO_UPDATE_CHECK=1 environment variable — no --no-update-check flag exists. Closes #541 Co-authored-by: ajianaz <ajianaz@users.noreply.github.com>
Co-authored-by: ajianaz <ajianaz@users.noreply.github.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What
GitFlow release merge: develop → main for v0.15.0. Tag
v0.15.0will be pushed on main after this merges, triggeringrelease.yml.Why
Ship 0.15.0: opt-in vecq vector store with keyed persistence, residual default width, and the
brain.vector_bitsknob (#543, #548, #547); fresh-process vector signal fix (#546, #545); cross-project fingerprint heal;cora upgradedocs (#549, #541).Testing
Full suite green on develop (947 tests); CI validates this PR.