Skip to content

v0.2.1 - Chat Mode Enhancements & Execution Tracing

Choose a tag to compare

@hidai25 hidai25 released this 11 Jan 22:10
· 423 commits to main since this release

What's New

Chat Mode Enhancements

  • New slash commands: /run, /test, /adapters, /compare
  • Natural language execution: LLM suggests commands and prompts user to run them
  • /compare command: Side-by-side regression detection between test runs

Execution Tracing

  • OpenTelemetry-style spans across all adapters
  • LLM call tracking with token usage and costs
  • Tool execution spans with timing

Documentation

  • Expanded Chat Mode section with slash commands table
  • Added Natural Language Execution examples
  • Updated Architecture section with tracing components

Bug Fixes

  • Fixed 20 mypy type errors in chat.py
  • Corrected Evaluations attribute access
  • Fixed variable shadowing issues

Installation

pip install evalview==0.2.1

Full Changelog

v0.2.0...v0.2.1