v1.8.0

1.8.0 - 2026-02-05

Bug Fixes

• (leann) Sanitize extracted text for utf-8 surrogates
• (paper) Add debug logging to silent exception handlers
• (regenerate) Update pdf_path after rename, respect --overwrite semantics
• Address PR review comments (P1, P2)
• Catch RuntimeError for PyMuPDF exceptions
• Address PR review issues for pdf-extraction-and-dedup

Features

• (cli) Add early duplicate detection by source URL
• (paper) Add source_url provenance tracking for local PDFs
• (paper) Improve LLM prompts to reduce hallucination
• (paper) Extract text from PDF when no LaTeX source available

Miscellaneous

• Bump version to 1.8.0

Testing

• (integration) Add local PDF workflow integration tests
• (leann) Isolate env vars in gemini default test
• (paper) Add integration tests for PDF text extraction fallback