Changelog

Unreleased

0.1.6 (2026-11-02)

• Raise the same exceptions as Python's open() for file problems: PermissionError and IsADirectoryError instead of ValueError, plus NotADirectoryError and OSError. They include errno and filename.
• Raise ValueError for a path that contains a null byte, instead of FileNotFoundError.
• Change the FileNotFoundError message to match Python's: [Errno 2] No such file or directory: 'report.pdf'.

0.1.5 (2026-11-02)

• Keep the text of the other pages when one page can't be read, and emit an ozaru.PageWarning with the page number and the reason. The page used to be dropped silently, and a crash in one page lost the whole document.

0.1.4 (2026-11-01)

• Raise ValueError in extract_text for PDFs that need a password, instead of silently returning an empty string.
• Create a GitHub release for each new version.

0.1.3 (2026-11-01)

• Release the GIL while reading a PDF, so other Python threads keep running and several PDFs can be read in parallel.

0.1.2 (2026-11-01)

• Fix a bug where a damaged PDF could make later PDFs silently lose text in the same process.
• Raise ValueError for damaged PDFs that used to raise PanicException, which except Exception can't catch.

0.1.1 (2026-11-01)

• Accept pathlib.Path and other path-like objects in extract_text and page_count.
• Translate error messages to English.
• Rewrite the README in English.
• Add this documentation site.

0.1.0 (2026-11-01)

First release.

• Add extract_text to get the text of a PDF.
• Add page_count to get the number of pages of a PDF.