firecrawl__anydoc/.gitattributes at feat/python-hosted-ocr - firecrawl__anydoc - GetSkill.work

skills/firecrawl__anydoc

mirror of https://github.com/firecrawl/anydoc.git synced 2026-09-14 14:18:33 +08:00

Files

T

tomsideguide 5f587c34e4 release 0.3.0: content-based format detection, crates.io deps

- Format::from_bytes detects the format from container signatures and
  package identity (PDF header, RTF open group, OLE stream names, OPC
  officeDocument content type with root-element and conventional-path
  fallbacks, ODF/EPUB mimetype with manifest/container fallbacks);
  to_markdown prefers content detection over the extension
- calamine and pdf-inspector move from git forks to crates.io releases
  (0.36.1 contains both fork fixes; 0.1.7 matches the pinned branch),
  unblocking cargo publish; migrate off deprecated merged-region APIs
- clean_text treats a CRLF pair as one line break instead of two spaces
- fixtures are checked out byte-exact via .gitattributes (EOL conversion
  was silently editing text fixtures on Windows)
- CI runs on pull requests and manual dispatch, not pushes to main
- README trimmed to user-facing behavior; package metadata for crates.io

2026-08-01 01:06:35 +01:00

3 lines

91 B

Plaintext

Raw Permalink Blame History

	`# Fixture bytes are snapshot ground truth: never EOL-convert them.`
	`tests/fixtures/** -text`