/** * Heading anchor (`id` slug) generation and deduplication extension. */ import {GrowableArray, i32ToString} from 'zena:core'; import {StringBuilder, StringReader, asciiLowerByte, isAsciiAlpha, isAsciiDigit, isAsciiWhitespace} from 'zena:string'; import {HashMap} from 'zena:collections'; import {Block, Document, Heading, BlockQuote, ListBlock, ListItem, inlinesToPlainText} from '../ast.zena'; import {AstTransformer, BlockParserFactory, DelimiterProcessor, InlineContentParser, InlineParserRunner, MarkdownExtension, NodeRenderer} from '../plugin.zena'; final class HeadingAnchorsTransformer implements AstTransformer { #prefix: String; new(prefix: String) : #prefix = prefix; transform(doc: Document, inlineParser: InlineParserRunner): void { let counts = new HashMap(); this.#visitBlocks(doc.blocks, counts); } #visitBlocks(blocks: GrowableArray, counts: HashMap): void { for (let block in blocks) { match (block) { case Heading as h: { if (h.id == null) { let text = inlinesToPlainText(h.inlines); let baseSlug = slugifyHeading(text); if (!baseSlug.isEmpty) { if (counts.has(baseSlug)) { let n = counts[baseSlug]; counts[baseSlug] = n + 1; h.id = `${this.#prefix}${baseSlug}-${i32ToString(n)}`; } else { counts[baseSlug] = 1; h.id = `${this.#prefix}${baseSlug}`; }; }; }; } case BlockQuote {blocks: let children}: { this.#visitBlocks(children, counts); } case ListBlock {items}: { for (let item in items) { this.#visitBlocks(item.blocks, counts); } } case ListItem {blocks: let children}: { this.#visitBlocks(children, counts); } case _: {} }; } } } /** * Converts heading text to a lowercase URL slug, turning whitespace into `-` * and keeping ASCII letters, digits, `-`, `_`, and non-ASCII Unicode scalars. */ export function slugifyHeading(text: String): String { let trimmed = text.trim(); let r = new StringReader(trimmed); let sb = new StringBuilder(); var lastWrittenWasDash = false; var lastWasStrippedAfterDash = false; while (!r.isAtEnd) { let cp = r.advance(); if (isAsciiAlpha(cp) || isAsciiDigit(cp) || cp == 0x5F || cp == 0x2D || cp >= 0x80) { sb.appendCodePoint(asciiLowerByte(cp)); lastWrittenWasDash = cp == 0x2D; lastWasStrippedAfterDash = false; } else if (isAsciiWhitespace(cp)) { if (!lastWasStrippedAfterDash) { sb.appendByte(0x2D); lastWrittenWasDash = true; } lastWasStrippedAfterDash = false; } else if (lastWrittenWasDash) { lastWasStrippedAfterDash = true; } } return sb.toString(); } /** * Returns the `MarkdownExtension` that populates deduplicated `id` slugs on * `Heading` AST nodes, each after `prefix`, which keeps them apart from the * IDs of the page the document is shown in (GitHub uses `user-content-`). */ export function headingAnchorsExtension(prefix: String = ''): MarkdownExtension { let transformers = new GrowableArray(1); transformers.push(new HeadingAnchorsTransformer(prefix)); return new MarkdownExtension( new GrowableArray(0), new GrowableArray(0), new GrowableArray(0), transformers, new GrowableArray(0) ); }