Added DOC file support to MarkItDown

[dizzydroid]

Summary

Adds support for legacy Microsoft Word DOC files (.doc) to MarkItDown.

Changes

• New DocConverter: Handles .doc files and application/msword mimetype
• Pure Python implementation: Uses existing olefile dependency (no new system requirements)
• Lightweight approach: No external tools like LibreOffice required
• Test integration: Adds test vector and verifies converter registration
• Documentation: Updates pyproject.toml with new doc optional dependency group

Implementation Details

• Uses olefile to parse the OLE structure of DOC files
• Implements text extraction with multiple fallback methods
• Handles binary format complexities gracefully
• Follows existing converter patterns and error handling

Testing

• All existing tests pass
• DocConverter properly registered and accepts DOC files
• Handles conversion without crashing
• Extracts readable content

Installation

Users can install with: pip install markitdown[doc] or pip install markitdown[all]

Addresses issue #23

[dizzydroid]

@microsoft-github-policy-service agree

[BetterAndBetterII]

really need it

[dizzydroid]

Updated the doc converter to be more reliable. I could not find an out-of-the-box library to do doc to md conversion, so I went with a 2-step approach, converting the doc to docx then converting the docx using the converter module to md. The minor issue here is the dependencies, all libraries require some sort of dependency (usually Libreoffice), I implemented an OS-specific approach that checks if the user is on Linux, it uses the Libreoffice cli tool, but, on Windows it would use MS Word's COM interface, this is to eliminate the need to install external dependencies as much as possible.