kepano/defuddle 0.9.0

on GitHub

Improvements

• Async extraction support e.g. X URLs
• Generic footnote detection fallback and backref cleanup (#138, #120)
• Substack app support
• Better Wikidot support
• Better heading/code/pre preservation
• Shiki language detection for code blocks
• Improved scoring around code blocks and bios
• Fixed nested list indentation

Fixes

• Fix HTML element with id="menu" breaking content extraction (#106)
• Fix page content not being able to start with a divider (#114)
• Fix invalid CSS selector span.leading-tight,, img (#128)
• Fix [href*="/category"] exact selector removing legitimate page content (#131)
• Fix .hero exact selector removing primary content on documentation landing pages (#132)
• Fix content of <time> element being removed (#136)
• Fix DOMParser is not defined when running via defuddle/node (#137)
• Fix content sanitization bypass via schema.org text fallback (#139)

Security

• Fix XSS via attribute injection in image handling
• Sanitize HTML to prevent unsafe elements in schema text fallback (#139)

Other

• New website (#133), playground updates, README updates