Switched from the stdlib minidom parser to defusedxml.

[afourney]

This pull request introduces changes to improve XML handling in the markitdown package by replacing the use of the standard xml.dom.minidom module with the more secure defusedxml.minidom. The changes also update type annotations to use Document and Element directly from xml.dom.minidom. These updates enhance security and maintain consistency in type annotations.

Security Improvements:

• Added defusedxml as a dependency in pyproject.toml to replace the standard xml.dom.minidom for safer XML parsing. (packages/markitdown/pyproject.toml, packages/markitdown/pyproject.tomlR32)
• Replaced imports of xml.dom.minidom with defusedxml.minidom in _epub_converter.py and _rss_converter.py for improved security. (packages/markitdown/src/markitdown/converters/_epub_converter.py, [1]; packages/markitdown/src/markitdown/converters/_rss_converter.py, [2]

Code Consistency:

• Updated type annotations in _epub_converter.py to use Document from xml.dom.minidom instead of minidom.Document. (packages/markitdown/src/markitdown/converters/_epub_converter.py, packages/markitdown/src/markitdown/converters/_epub_converter.pyL131-R140)
• Updated type annotations in _rss_converter.py to use Document and Element directly, ensuring consistent usage across methods. (packages/markitdown/src/markitdown/converters/_rss_converter.py, [1] [2] [3]