markdown-codec 4.0.15 → 4.1.1

package/dist/lower/inline.js

@@ -11,7 +11,7 @@ function buildRun(text, style, fontFamily) {
 		...fontFamily !== void 0 ? { fontFamily } : {}
 	};
 }
-function lowerNestedEmphasisLike(kind, node, style, context) {
+function lowerNestedEmphasisLike(kind, node, style, context, runs, extents) {
 	if (style[kind] === true) context.sink({
 		code: MarkdownDiagnosticCodes.NESTED_EMPHASIS_FLATTENED,
 		severity: "info",
@@ -21,15 +21,25 @@ function lowerNestedEmphasisLike(kind, node, style, context) {
 		...style,
 		[kind]: true
 	};
-	return node.children.flatMap((child) => lowerInlineNode(child, childStyle, context));
+	lowerNodesInto(node.children, childStyle, context, runs, extents);
 }
-function lowerInlineNode(node, style, context) {
+function lowerInlineNodeInto(node, style, context, runs, extents) {
 	switch (node.type) {
-		case "text": return node.value.length === 0 ? [] : [buildRun(node.value, style)];
-		case "entity": return node.value.length === 0 ? [] : [buildRun(node.value, style)];
-		case "softBreak": return [buildRun(" ", style)];
-		case "hardBreak": return [buildRun("\n", style)];
-		case "codeSpan": return [buildRun(node.literal, style, MONOSPACE_FONT_FAMILY)];
+		case "text":
+			if (node.value.length > 0) runs.push(buildRun(node.value, style));
+			return;
+		case "entity":
+			if (node.value.length > 0) runs.push(buildRun(node.value, style));
+			return;
+		case "softBreak":
+			runs.push(buildRun(" ", style));
+			return;
+		case "hardBreak":
+			runs.push(buildRun("\n", style));
+			return;
+		case "codeSpan":
+			runs.push(buildRun(node.literal, style, MONOSPACE_FONT_FAMILY));
+			return;
 		case "rawHtml":
 			if (context.rawHtml === "drop") {
 				context.sink({
@@ -37,47 +47,66 @@ function lowerInlineNode(node, style, context) {
 					severity: "info",
 					message: "inline raw HTML was dropped per the rawHtml: \"drop\" option"
 				});
-				return [];
+				return;
 			}
 			context.sink({
 				code: MarkdownDiagnosticCodes.RAW_HTML_PRESERVED_AS_TEXT,
 				severity: "info",
-				message: "inline raw HTML was preserved as literal text; it will not be rendered as HTML by any consumer of the resulting ContentDocument"
+				message: "inline raw HTML was preserved as literal text; it will not be rendered as HTML by any consumer of the resulting ContentDocument, and its verbatim original rides the run's own markdown residue for this package's writer to re-emit as-is"
 			});
-			return node.literal.length === 0 ? [] : [buildRun(node.literal, style)];
+			if (node.literal.length > 0) runs.push({
+				...buildRun(node.literal, style),
+				source: {
+					format: "markdown",
+					xml: node.literal
+				}
+			});
+			return;
 		case "mathInline":
 			context.sink({
 				code: MarkdownDiagnosticCodes.MATH_INLINE_PRESERVED_AS_TEXT,
 				severity: "info",
 				message: "inline math (\\( \\)) was preserved as literal raw LaTeX text; it is not parsed as LaTeX or converted to MathML by this package"
 			});
-			return [buildRun(node.literal, style, MATH_INLINE_FONT_MARKER)];
+			runs.push(buildRun(node.literal, style, MATH_INLINE_FONT_MARKER));
+			return;
 		case "footnoteReference":
 			context.sink({
 				code: MarkdownDiagnosticCodes.FOOTNOTE_REFERENCE_PRESERVED_AS_TEXT,
 				severity: "info",
 				message: `footnote reference "[^${node.label}]" is preserved as a marked text run rather than an anchor construct: a construct's extent is block-scoped, and a reference site sits between two runs inside a paragraph, which no block-level boundary marker can bracket`
 			});
-			return [buildRun(`[^${node.label}]`, style, FOOTNOTE_REFERENCE_FONT_MARKER)];
+			runs.push(buildRun(`[^${node.label}]`, style, FOOTNOTE_REFERENCE_FONT_MARKER));
+			return;
 		case "autolink": {
 			const destination = node.email ? `mailto:${node.destination}` : node.destination;
-			return [buildRun(node.destination, {
+			runs.push(buildRun(node.destination, {
 				...style,
 				hyperlink: destination
-			})];
+			}));
+			return;
 		}
 		case "link": {
-			if (node.title !== void 0) context.sink({
-				code: MarkdownDiagnosticCodes.LINK_TITLE_DROPPED,
-				severity: "info",
-				message: `link title "${node.title}" has no ContentRun equivalent and was dropped`
-			});
 			const childStyle = {
 				...style,
 				hyperlink: node.destination
 			};
-			const runs = node.children.flatMap((child) => lowerInlineNode(child, childStyle, context));
-			return runs.length > 0 ? runs : [buildRun("", childStyle)];
+			const startRun = runs.length;
+			lowerNodesInto(node.children, childStyle, context, runs, extents);
+			if (runs.length === startRun) runs.push(buildRun("", childStyle));
+			if (node.title !== void 0) extents.push({
+				descriptor: {
+					kind: "link",
+					target: {
+						kind: "external",
+						uri: node.destination
+					},
+					title: node.title
+				},
+				startRun,
+				endRun: runs.length
+			});
+			return;
 		}
 		case "image":
 			if (node.title !== void 0) context.sink({
@@ -85,17 +114,33 @@ function lowerInlineNode(node, style, context) {
 				severity: "info",
 				message: `image title "${node.title}" has no ContentRun equivalent and was dropped`
 			});
-			return [buildRun(node.alt, {
+			runs.push(buildRun(node.alt, {
 				...style,
 				hyperlink: node.destination
-			})];
-		case "emphasis": return lowerNestedEmphasisLike("italic", node, style, context);
-		case "strong": return lowerNestedEmphasisLike("bold", node, style, context);
-		case "strikethrough": return lowerNestedEmphasisLike("strike", node, style, context);
+			}));
+			return;
+		case "emphasis":
+			lowerNestedEmphasisLike("italic", node, style, context, runs, extents);
+			return;
+		case "strong":
+			lowerNestedEmphasisLike("bold", node, style, context, runs, extents);
+			return;
+		case "strikethrough":
+			lowerNestedEmphasisLike("strike", node, style, context, runs, extents);
+			return;
 	}
 }
+function lowerNodesInto(nodes, style, context, runs, extents) {
+	for (const node of nodes) lowerInlineNodeInto(node, style, context, runs, extents);
+}
 function lowerInlineNodes(nodes, context) {
-	return nodes.flatMap((node) => lowerInlineNode(node, {}, context));
+	const runs = [];
+	const extents = [];
+	lowerNodesInto(nodes, {}, context, runs, extents);
+	return {
+		runs,
+		linkTitleExtents: extents
+	};
 }
 function lowerCodeBlockRun(literal) {
 	return {
@@ -104,4 +149,4 @@ function lowerCodeBlockRun(literal) {
 	};
 }
 //#endregion
-export { lowerCodeBlockRun, lowerInlineNode, lowerInlineNodes };
+export { lowerCodeBlockRun, lowerInlineNodes };

package/dist/lower/lower.cjs

@@ -25,17 +25,21 @@ function decorateParagraph(paragraph, context) {
 	};
 	if (context.list !== void 0) result = {
 		...result,
-		list: {
-			numId: context.list.numId,
-			level: context.list.level
-		}
+		list: { ...context.list }
 	};
 	return result;
 }
+function paragraphWithConstructs(runs, linkTitleExtents) {
+	return {
+		runs,
+		...linkTitleExtents.length > 0 ? { constructs: [...linkTitleExtents] } : {}
+	};
+}
 function lowerHeading(node, context) {
+	const inline = require_lower_inline.lowerInlineNodes(node.children, inlineContext(context));
 	return [decorateParagraph({
 		kind: "paragraph",
-		runs: require_lower_inline.lowerInlineNodes(node.children, inlineContext(context)),
+		...paragraphWithConstructs(inline.runs, inline.linkTitleExtents),
 		styleId: require_shared_style_constants.headingStyleId(node.level),
 		headingLevel: node.level
 	}, context)];
@@ -46,9 +50,10 @@ function lowerParagraph(node, context) {
 	let segment = [];
 	const flushSegment = (force) => {
 		if (segment.length === 0 && !force) return;
+		const inline = require_lower_inline.lowerInlineNodes(segment, inlineCtx);
 		blocks.push(decorateParagraph({
 			kind: "paragraph",
-			runs: require_lower_inline.lowerInlineNodes(segment, inlineCtx)
+			...paragraphWithConstructs(inline.runs, inline.linkTitleExtents)
 		}, context));
 		segment = [];
 	};
@@ -70,39 +75,73 @@ function lowerParagraph(node, context) {
 			segment.push(child);
 			continue;
 		}
-		if (child.title !== void 0) context.sink({
-			code: require_diagnostics_diagnostics.MarkdownDiagnosticCodes.LINK_TITLE_DROPPED,
-			severity: "info",
-			message: `image title "${child.title}" has no ContentImageBlock equivalent and was dropped`
-		});
 		flushSegment(false);
 		if (context.list !== void 0) context.sink({
 			code: require_diagnostics_diagnostics.MarkdownDiagnosticCodes.LIST_ITEM_BLOCK_UNLISTED,
 			severity: "info",
 			message: "a resolved image block directly inside a list item has no ContentListMembership field of its own -- only ContentParagraph carries .list -- so its association with the enclosing list item is lost"
 		});
-		blocks.push({
+		const image = {
 			kind: "image",
 			format: resolved.format,
 			base64: resolved.base64,
 			widthPt: resolved.widthPt,
 			heightPt: resolved.heightPt,
 			...child.alt.length > 0 ? { altText: child.alt } : {}
-		});
+		};
+		if (child.title === void 0) {
+			blocks.push(image);
+			continue;
+		}
+		blocks.push({
+			kind: "constructStart",
+			descriptor: {
+				kind: "link",
+				target: {
+					kind: "external",
+					uri: child.destination
+				},
+				title: child.title
+			}
+		}, image, { kind: "constructEnd" });
 	}
 	flushSegment(blocks.length === 0);
 	return blocks;
 }
+function splitInfoString(infoString) {
+	const firstWhitespace = infoString.search(/\s/);
+	if (firstWhitespace === -1) return infoString.startsWith("{") ? {
+		language: void 0,
+		remainder: infoString
+	} : {
+		language: infoString,
+		remainder: void 0
+	};
+	const firstWord = infoString.slice(0, firstWhitespace);
+	const remainder = infoString.slice(firstWhitespace).trim();
+	if (firstWord.startsWith("{")) return {
+		language: void 0,
+		remainder: infoString.trim()
+	};
+	return {
+		language: firstWord,
+		remainder: remainder.length > 0 ? remainder : void 0
+	};
+}
 function lowerCodeBlock(node, context) {
-	if (node.fenced && node.infoString !== void 0 && node.infoString.length > 0) context.sink({
-		code: require_diagnostics_diagnostics.MarkdownDiagnosticCodes.CODE_BLOCK_INFO_STRING_DROPPED,
-		severity: "info",
-		message: `fenced code block's own info string "${node.infoString}" has no ContentParagraph equivalent and was dropped`
-	});
+	const info = node.fenced && node.infoString !== void 0 && node.infoString.length > 0 ? splitInfoString(node.infoString) : {
+		language: void 0,
+		remainder: void 0
+	};
 	return [decorateParagraph({
 		kind: "paragraph",
 		runs: [require_lower_inline.lowerCodeBlockRun(node.literal.replace(/\n$/, ""))],
-		styleId: require_shared_style_constants.CODE_BLOCK_STYLE_ID
+		styleId: require_shared_style_constants.CODE_BLOCK_STYLE_ID,
+		...info.language !== void 0 ? { codeLanguage: info.language } : {},
+		...info.remainder !== void 0 ? { source: {
+			format: "markdown",
+			xml: info.remainder
+		} } : {}
 	}, context)];
 }
 function lowerThematicBreak(context) {
@@ -124,70 +163,95 @@ function lowerHtmlBlock(node, context) {
 	context.sink({
 		code: require_diagnostics_diagnostics.MarkdownDiagnosticCodes.RAW_HTML_PRESERVED_AS_TEXT,
 		severity: "info",
-		message: "block-level raw HTML was preserved as literal text (styleId \"HTMLPreformatted\"); it will not be rendered as HTML by any consumer of the resulting ContentDocument"
+		message: "block-level raw HTML was preserved as literal text (styleId \"HTMLPreformatted\"); it will not be rendered as HTML by any consumer of the resulting ContentDocument, and its verbatim original rides the paragraph's own markdown residue for this package's writer to re-emit as-is"
 	});
 	const literal = node.literal.replace(/\n+$/, "");
 	return [decorateParagraph({
 		kind: "paragraph",
 		runs: literal.length === 0 ? [] : [{ text: literal }],
-		styleId: require_shared_style_constants.HTML_PREFORMATTED_STYLE_ID
+		styleId: require_shared_style_constants.HTML_PREFORMATTED_STYLE_ID,
+		source: {
+			format: "markdown",
+			xml: node.literal
+		}
 	}, context)];
 }
 function lowerMathBlock(node, context) {
-	context.sink({
-		code: require_diagnostics_diagnostics.MarkdownDiagnosticCodes.MATH_BLOCK_PRESERVED_AS_TEXT,
+	if (context.list !== void 0) context.sink({
+		code: require_diagnostics_diagnostics.MarkdownDiagnosticCodes.LIST_ITEM_BLOCK_UNLISTED,
 		severity: "info",
-		message: "block math ($$...$$) was preserved as literal raw LaTeX text (styleId \"MathBlock\"); it is not parsed as LaTeX or converted to MathML by this package"
+		message: "a display-math block directly inside a list item has no ContentListMembership field of its own -- only ContentParagraph carries .list -- so its association with the enclosing list item is lost"
 	});
-	const literal = node.literal.replace(/\n$/, "");
-	return [decorateParagraph({
-		kind: "paragraph",
-		runs: literal.length === 0 ? [] : [{ text: literal }],
-		styleId: require_shared_style_constants.MATH_BLOCK_STYLE_ID
-	}, context)];
+	return [{
+		kind: "embeddedObject",
+		objectKind: "formula",
+		document: {
+			kind: "formula",
+			metadata: {},
+			formula: {
+				mathml: [],
+				presentation: { latex: node.literal.replace(/\n$/, "") }
+			}
+		},
+		frame: {
+			xPt: 0,
+			yPt: 0,
+			widthPt: 0,
+			heightPt: 0
+		}
+	}];
+}
+function blockquoteSubtreeContainsHeading(node) {
+	const walk = (block) => {
+		switch (block.type) {
+			case "heading": return true;
+			case "blockquote":
+			case "list":
+			case "listItem": return block.children.some(walk);
+			default: return false;
+		}
+	};
+	return node.children.some(walk);
 }
 function lowerBlockquote(node, context, contentWidthPt) {
-	if (context.quoteDepth >= 1) context.sink({
-		code: require_diagnostics_diagnostics.MarkdownDiagnosticCodes.BLOCKQUOTE_NESTED_DEPTH,
-		severity: "info",
-		message: `blockquote nesting beyond level 1 is represented only as a larger indentLeftPt (${String((context.quoteDepth + 1) * 36)}pt); recovering the exact nesting depth back out is an approximation, not an exact inverse`
-	});
 	const nested = {
 		...context,
 		quoteDepth: context.quoteDepth + 1
 	};
 	const blocks = node.children.flatMap((child) => lowerBlock(child, nested, contentWidthPt));
-	if (blocks.length === 0) return [decorateParagraph({
+	const inner = blocks.length === 0 ? [decorateParagraph({
 		kind: "paragraph",
 		runs: []
-	}, nested)];
-	return blocks;
-}
-function applyTaskCheckbox(blocks, checked) {
-	const first = blocks[0];
-	if (first?.kind !== "paragraph") return false;
-	const checkboxRun = { text: `${checked ? "☒" : "☐"} ` };
-	blocks[0] = {
-		...first,
-		runs: [checkboxRun, ...first.runs]
-	};
-	return true;
+	}, nested)] : blocks;
+	if (blockquoteSubtreeContainsHeading(node)) {
+		context.sink({
+			code: require_diagnostics_diagnostics.MarkdownDiagnosticCodes.BLOCKQUOTE_CONTAINER_SKIPPED,
+			severity: "info",
+			message: "a blockquote containing a heading cannot carry its division construct -- a marker extent may not open a heading scope, and the last heading inside an extent always leaves one standing -- so this quote degrades to indent-only structure while the heading keeps its heading fidelity"
+		});
+		return inner;
+	}
+	return [
+		{
+			kind: "constructStart",
+			descriptor: { kind: "division" }
+		},
+		...inner,
+		{ kind: "constructEnd" }
+	];
 }
 function lowerListItem(item, numId, level, context, contentWidthPt) {
-	if (item.children.filter((child) => child.type !== "list").length > 1) context.sink({
-		code: require_diagnostics_diagnostics.MarkdownDiagnosticCodes.LIST_ITEM_MULTI_BLOCK_FLATTENED,
-		severity: "info",
-		message: "a list item directly containing more than one block loses its own item boundary once lowered -- ContentListMembership carries only numId/level, with no field distinguishing \"one item, several blocks\" from \"several items sharing this numId/level\""
-	});
+	const membership = {
+		numId,
+		level,
+		...item.checked !== void 0 ? { checked: item.checked } : {},
+		itemId: require_shared_list_id.mintListItemId(context.numIdState)
+	};
 	const itemContext = {
 		...context,
-		list: {
-			numId,
-			level
-		}
+		list: membership
 	};
 	const blocks = [];
-	let checkboxApplied = item.checked === void 0;
 	let ownLevelBlockCount = 0;
 	for (const child of item.children) {
 		if (child.type === "list") {
@@ -196,17 +260,12 @@ function lowerListItem(item, numId, level, context, contentWidthPt) {
 		}
 		const childBlocks = lowerBlock(child, itemContext, contentWidthPt);
 		ownLevelBlockCount += childBlocks.length;
-		if (!checkboxApplied) checkboxApplied = applyTaskCheckbox(childBlocks, item.checked === true);
 		blocks.push(...childBlocks);
 	}
-	if (ownLevelBlockCount === 0) {
-		const placeholder = [decorateParagraph({
-			kind: "paragraph",
-			runs: []
-		}, itemContext)];
-		if (!checkboxApplied) applyTaskCheckbox(placeholder, item.checked === true);
-		blocks.unshift(...placeholder);
-	}
+	if (ownLevelBlockCount === 0) blocks.unshift(decorateParagraph({
+		kind: "paragraph",
+		runs: []
+	}, itemContext));
 	return blocks;
 }
 function lowerList(node, ancestorNumId, level, context, contentWidthPt) {
@@ -331,15 +390,16 @@ function lowerParsedMarkdown(parsed, options = {}, metadata = {}) {
 		}]
 	};
 }
-function lowerMarkdown(source, options = {}) {
+function lowerMarkdownDetailed(source, options = {}) {
 	if (options.maxInputBytes !== void 0) {
 		const actualBytes = new TextEncoder().encode(source).length;
 		if (actualBytes > options.maxInputBytes) throw new require_diagnostics_diagnostics.MarkdownInputTooLargeError(options.maxInputBytes, actualBytes);
 	}
 	const sink = options.sink ?? require_diagnostics_diagnostics.NOOP_MARKDOWN_DIAGNOSTIC_SINK;
-	const { metadata, rest } = options.frontMatter ?? false ? require_lower_front_matter.extractFrontMatter(source, sink) : {
+	const extracted = options.frontMatter ?? false ? require_lower_front_matter.extractFrontMatter(source, sink) : {
 		metadata: {},
-		rest: source
+		rest: source,
+		source: void 0
 	};
 	const parseOptions = {
 		gfmTables: options.gfmTables,
@@ -350,8 +410,17 @@ function lowerMarkdown(source, options = {}) {
 		maxNesting: options.maxBlockNesting,
 		sink
 	};
-	return lowerParsedMarkdown(require_block_block.parseMarkdown(rest, parseOptions), options, metadata);
+	const parsed = require_block_block.parseMarkdown(extracted.rest, parseOptions);
+	return {
+		document: lowerParsedMarkdown(parsed, options, extracted.metadata),
+		references: parsed.references,
+		frontMatterSource: extracted.source
+	};
+}
+function lowerMarkdown(source, options = {}) {
+	return lowerMarkdownDetailed(source, options).document;
 }
 //#endregion
 exports.lowerMarkdown = lowerMarkdown;
+exports.lowerMarkdownDetailed = lowerMarkdownDetailed;
 exports.lowerParsedMarkdown = lowerParsedMarkdown;

package/dist/lower/lower.d.cts

@@ -1,8 +1,15 @@
+import { LinkReferenceMap } from "../inline/link.cjs";
 import { ParsedMarkdown } from "../block/block.cjs";
 import { ReadMarkdownOptions } from "../options/options.cjs";
 import { ContentDocument, LayoutMetadata } from "document-schema.js";
 //#region src/lower/lower.d.ts
 declare function lowerParsedMarkdown(parsed: ParsedMarkdown, options?: ReadMarkdownOptions, metadata?: LayoutMetadata): ContentDocument;
+interface LoweredMarkdownDetail {
+  readonly document: ContentDocument;
+  readonly references: LinkReferenceMap;
+  readonly frontMatterSource: string | undefined;
+}
+declare function lowerMarkdownDetailed(source: string, options?: ReadMarkdownOptions): LoweredMarkdownDetail;
 declare function lowerMarkdown(source: string, options?: ReadMarkdownOptions): ContentDocument;
 //#endregion
-export { lowerMarkdown, lowerParsedMarkdown };
+export { LoweredMarkdownDetail, lowerMarkdown, lowerMarkdownDetailed, lowerParsedMarkdown };

package/dist/lower/lower.d.ts

@@ -1,8 +1,15 @@
+import { LinkReferenceMap } from "../inline/link.js";
 import { ParsedMarkdown } from "../block/block.js";
 import { ReadMarkdownOptions } from "../options/options.js";
 import { ContentDocument, LayoutMetadata } from "document-schema.js";
 //#region src/lower/lower.d.ts
 declare function lowerParsedMarkdown(parsed: ParsedMarkdown, options?: ReadMarkdownOptions, metadata?: LayoutMetadata): ContentDocument;
+interface LoweredMarkdownDetail {
+  readonly document: ContentDocument;
+  readonly references: LinkReferenceMap;
+  readonly frontMatterSource: string | undefined;
+}
+declare function lowerMarkdownDetailed(source: string, options?: ReadMarkdownOptions): LoweredMarkdownDetail;
 declare function lowerMarkdown(source: string, options?: ReadMarkdownOptions): ContentDocument;
 //#endregion
-export { lowerMarkdown, lowerParsedMarkdown };
+export { LoweredMarkdownDetail, lowerMarkdown, lowerMarkdownDetailed, lowerParsedMarkdown };