Expand description
Dependency-free HTML → Markdown / plain-text conversion.
A small tag tokenizer feeds a state-machine renderer. The goal is clean,
readable content for an LLM, not a faithful DOM: noise elements (script,
style, nav chrome) are dropped, block structure becomes Markdown headings /
lists / paragraphs, links become [text](href), and <pre> becomes fenced
code. Implemented without an HTML crate to stay in line with the project’s
zero-heavy-dependency stance.
Structs§
- HtmlDoc
- Parsed document: optional
<title>, rendered Markdown, and extracted links. - Link
- A hyperlink extracted from the document.
Functions§
- decode_
entities - Decode HTML/XML character entities (
&,',’, …). - markdown_
to_ text - Strip Markdown decorations to obtain flowing plain text.
- parse
- Convert an HTML document into Markdown plus extracted metadata.
- title
- Extract just the document
<title>without rendering the body.