Skip to content

v1.0.2 — Token Efficiency Optimizations

Choose a tag to compare

View all tags
@albertje8118 albertje8118 released this 15 Feb 12:59
· 42 commits to main since this release
v1.0.2
8ae95d9

What's Changed

Phase 1 — Core Optimizations

  • Shared helpers module (_helpers.py) — eliminates code duplication across tool modules
  • 3-tier caching — HTML page cache + parsed WikiPage cache + search response cache (2-min TTL)
  • Topic previews — \list_code_wiki_topics\ returns titles + 200-char summaries instead of full page dump
  • Pagination —
    ead_wiki_contents\ supports \offset/\limit\ parameters (default: 5 sections per call)
  • Input validation — new \ContentsInput\ Pydantic model with \�alidate_contents_input\

Phase 2 — Documentation

  • Updated tool descriptions on docs site reflecting new parameters and behaviors
  • Rewritten Agents page with token-efficient workflow guide and anti-patterns
  • Added Model Selection & Credit Optimization section (GitHub Copilot multiplier table)
  • Added comprehensive v1.0.2 release notes
  • Version badges updated across all documentation pages

Phase 3 — Session Pool

  • LRU-ordered persistent browser context pool for search sessions (avoids opening fresh browser for each query)
  • Search automatically reuses warm sessions with fallback to fresh context on failure
  • Pool cleanup integrated into server shutdown

Code Quality

  • Pylint: 9.93/10
  • Ruff: all checks passed
  • Black: fully formatted
  • Tests: 101 passed, 0 failures

Full Changelog: v1.0.1...v1.0.2

Assets 2
Loading