Skip to content

feat: adaptive retrieval and benchmark execution improvements - #10

Open
outbounder wants to merge 2 commits into
mainfrom
92-optimize-kp
Open

feat: adaptive retrieval and benchmark execution improvements#10
outbounder wants to merge 2 commits into
mainfrom
92-optimize-kp

Conversation

@outbounder

@outbounder outbounder commented Apr 17, 2026

Copy link
Copy Markdown
Member

Summary

  • add adaptive fact retrieval internals with confidence-aware graph expansion, write-time ingest signals, first-wave relation confidence metadata, and retrieval trace logging
  • improve benchmark execution ergonomics and reliability with quick stack scripts, benchmark env setup/warmup helpers, and run summarization tooling
  • update benchmark/docs plumbing and API/adapter paths for namespace-aware filtering plus refreshed benchmark documentation/spec details

Test plan

  • npm run typecheck --workspace=packages/api-core
  • npm run typecheck --workspace=packages/db
  • npm run bench:quick:norerank
  • Verify CI passes on PR (lint/type/tests)

Made with Cursor

@outbounder
feat: add human-readable benchmark markdown summarizer
cdf6597 
Add a Node.js/TypeScript benchmark summary script that generates a narrative markdown report with explicit regression callouts to support debugging and optimization workflows.
@outbounder
feat: add adaptive graph retrieval and write-time confidence signals
9ef7c2d 
Introduce production-grade retrieval and ingest upgrades with adaptive graph expansion budgets, retrieval trace observability, namespace-aware filtering, and first-wave relation stitching with confidence metadata while keeping the external facts API simple.

Made-with: Cursor
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

1 participant

@outbounder