@knight-owl-llc/pandoc-parser

Pandoc 3.11's Markdown reader, ported to JavaScript: a document's AST as Pandoc reads it, each node with the span of source it was read from.

The AST is Pandoc's JSON with a start and end on each node: JSON.stringify(withoutSpans(doc)) equals pandoc -f markdown -t json on the same text. The reader runs with Pandoc's default Markdown extensions and reads raw TeX with a port of Pandoc's LaTeX reader.

Use

import { readMarkdown } from "@knight-owl-llc/pandoc-parser";

const doc = readMarkdown(text, { tabStop: 4 });
doc.blocks[0].start; // where the first block starts in `text`

| Export | What it is | | ----------------------------------- | --------------------------------------------------------------------------------- | | readMarkdown(text, {tabStop}) | The document as Pandoc's Markdown reader reads it | | readLaTeX(text, options) | The document as Pandoc's LaTeX reader reads it | | withoutSpans(value) | A read without its spans: Pandoc's JSON | | toSources(files) | Files joined as Pandoc's CLI joins them, with a lookup from an offset to its file | | continuesParagraph(next, options) | Whether Pandoc reads a paragraph on into the line next | | DEFAULT_TAB_STOP | Pandoc's default --tab-stop, 4 |

A Markdown read also carries four properties outside its JSON:

| Property | What it holds | | ---------------- | ----------------------------------------------------------------------------------- | | definitions | Each reference definition, with the span of each of its parts | | log | The messages Pandoc logs reading it, in order, each with Pandoc's fields and a span | | unresolved | Each reference and note looked up and not found: kind, form, label and span | | metadataBlocks | Each YAML metadata block and title block read for metadata, with its span |

Offsets are UTF-16 code units, as JavaScript indexes a string.

License

GPL-2.0-or-later, as Pandoc's: LICENSE. NOTICE.md gives the port's provenance and the notices of the libraries ported with it.