pub fn parse_djot(djot: &str, options: &DjotImportOptions) -> Vec<ParsedElement>Expand description
Parse djot source into the shared ParsedElement intermediate, mirroring
parse_markdown. Uses the jotdown pull parser.
Constructs the document model cannot represent are dropped, and their text
content is discarded so it never leaks into the document: footnotes, math,
fenced divs, raw blocks/inline, thematic breaks, description lists,
captions, symbols, link-reference definitions, and highlight/mark. Inline
images keep their alt text as plain text (the image itself is not modelled),
matching the Markdown importer. Smart-punctuation events are normalised to
their canonical Unicode characters so the model→djot→model round-trip is a
fixpoint.
Standalone paragraphs and headings additionally carry the optional
block-style attributes selected by options — paragraph alignment, line
height, text direction, non-breakable lines and background color — read from
djot {key=value} block attributes (see DjotImportOptions). List items,
code blocks and table cells normalise their block styling away.
Known model limitations (normalised, not preserved on round-trip): ordered-list start number, table column alignment, and list tight/loose.