Skip to content

v0.4.1

Choose a tag to compare

View all tags
@agamm agamm released this 30 Jul 19:01
· 13 commits to main since this release
v0.4.1
0d5ee6f

PDF Cost Estimation Fix & Dry Run Feature

Major Improvements

  • Fixed PDF cost estimation - Previously overestimated by ~50x (e.g. $18.70 → $0.32)
  • Added dry run feature - Preview costs before execution with batch.run(dry_run=True)
  • Intelligent PDF processing - Uses text extraction for accurate token counting

New Features

  • PDF text extraction with pypdf library
  • Configurable token coefficient (1.5x default for PDF overhead)
  • Smart classification of textual vs image-based PDFs
  • Comprehensive test coverage for PDF features

Improvements

  • Reduced logging verbosity for cleaner output
  • Better error handling for PDF processing
  • Fallback to page-based estimation for image PDFs

No breaking changes

All existing code continues to work unchanged.

Assets 2
Loading