Add Yardstick workflow for continuous quality metrics#1275
Open
LeexLee94 wants to merge 2 commits intoruvnet:mainfrom
Open
Add Yardstick workflow for continuous quality metrics#1275LeexLee94 wants to merge 2 commits intoruvnet:mainfrom
LeexLee94 wants to merge 2 commits intoruvnet:mainfrom
Conversation
Single-action workflow that measures what matters on every PR and push: source lines, TypeScript errors, test pass/fail, line coverage, vulnerability counts, dependency totals, and build time. Results post as a GitHub Step Summary and an auto-updating PR comment so the state of the codebase is visible without digging. https://claude.ai/code/session_01BVecxWKHe7dRZ9GEPHDnJz
…onfig
The HNSW index already supported quantization via its internal `Quantizer`
class and the `HNSWConfig.quantization` field, but the public
`AgentDBAdapterConfig` had no way to enable it — users could not activate
the 4x-32x memory reduction or the corresponding search speedup without
patching internal code.
Changes:
- Add `hnswEfSearch: number` (default 100) to `AgentDBAdapterConfig`
so the search-time beam width is independently tunable from the
construction-time `hnswEfConstruction`
- Add `quantization?: QuantizationConfig` to `AgentDBAdapterConfig`
with inline docs explaining scalar/binary/product trade-offs
- Wire both fields through the `HNSWIndex` constructor
- Default `search()` to `config.hnswEfSearch` when callers do not
supply a per-query `ef`, removing the current fallback to the
(much higher) `efConstruction` value
- Add `optimization-presets.ts` with four ready-to-use configs:
- `speedPreset` — binary quant, M=8, ef=50 (10x faster, 32x smaller)
- `balancedPreset` — scalar quant, M=16, ef=100 (3x faster, 4x smaller)
- `accuracyPreset` — no quant, M=32, ef=200 (100% recall)
- `edgePreset` — binary quant, M=8, ef=50, cache=100 (mobile/WASM)
plus `withPreset(name, overrides)` helper for one-line setup
- Re-export presets from the package index
https://claude.ai/code/session_01BVecxWKHe7dRZ9GEPHDnJz
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
2 participants
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.
Summary
Introduces a comprehensive GitHub Actions workflow that automatically measures and reports code quality metrics on every PR and push to main.
Key Changes
.github/workflows/yardstick.ymlthat runs on pull requests, pushes to main, and manual dispatchImplementation Details
https://claude.ai/code/session_01BVecxWKHe7dRZ9GEPHDnJz