epub-codec 1.1.0 → 1.2.0

package/dist/xhtml/inline.js

@@ -1,6 +1,8 @@
 import { EpubDiagnosticCodes } from "../diagnostics.js";
+import { decodeEntities, decodeTextLikeNode } from "../xml/entities.js";
+import { isTextLikeNode } from "../xml/node.js";
 import { attrValue } from "../xml/query.js";
-import { decodeEntities } from "../xml/entities.js";
+import { isInertElement, reportInertElementSkip } from "./context.js";
 import { isFootnoteReferenceAnchor, sameDocumentFragment } from "./footnote.js";
 import { MONOSPACE_FONT_FAMILY } from "./style-constants.js";
 //#region src/xhtml/inline.ts
@@ -27,8 +29,8 @@ function buildInlineRuns(nodes, style, context) {
 	const runs = [];
 	const constructs = [];
 	for (const node of nodes) {
-		if (node.type === "text") {
-			const text = normalizeWhitespace(decodeEntities(node.value));
+		if (isTextLikeNode(node)) {
+			const text = normalizeWhitespace(decodeTextLikeNode(node));
 			if (text.length > 0) runs.push(styledRun(text, style));
 			continue;
 		}
@@ -41,6 +43,10 @@ function buildInlineRuns(nodes, style, context) {
 	};
 }
 function appendElement(element, style, context, runs, constructs) {
+	if (isInertElement(element.tag)) {
+		reportInertElementSkip(element.tag, context);
+		return;
+	}
 	switch (element.tag) {
 		case "strong":
 		case "b":

package/dist/xhtml/read.cjs

@@ -1,11 +1,13 @@
 Object.defineProperty(exports, Symbol.toStringTag, { value: "Module" });
 const require_diagnostics = require("../diagnostics.cjs");
+const require_xml_entities = require("../xml/entities.cjs");
+const require_xml_node = require("../xml/node.cjs");
 const require_xml_query = require("../xml/query.cjs");
 const require_xml_parse = require("../xml/parse.cjs");
-const require_xml_entities = require("../xml/entities.cjs");
 const require_image_dimensions = require("../image/dimensions.cjs");
 const require_util_base64 = require("../util/base64.cjs");
 const require_xml_build = require("../xml/build.cjs");
+const require_xhtml_context = require("./context.cjs");
 const require_xhtml_footnote = require("./footnote.cjs");
 const require_xhtml_style_constants = require("./style-constants.cjs");
 const require_xhtml_inline = require("./inline.cjs");
@@ -62,18 +64,32 @@ function decorateParagraph(paragraph, state) {
 	}
 	return decorated;
 }
+function constructsField(inline) {
+	return inline.constructs.length > 0 ? { constructs: inline.constructs } : {};
+}
 function buildIdElementMap(nodes) {
 	const map = /* @__PURE__ */ new Map();
 	const stack = [...nodes];
 	while (stack.length > 0) {
 		const node = stack.pop();
-		if (node?.type !== "element") continue;
+		if (node?.type !== "element" || require_xhtml_context.isInertElement(node.tag)) continue;
 		const id = require_xml_query.attrValue(node, "id");
 		if (id !== void 0 && !map.has(id)) map.set(id, node);
 		stack.push(...node.children);
 	}
 	return map;
 }
+function elementsWithTagSkippingInert(nodes, tag) {
+	const out = [];
+	const stack = [...nodes];
+	while (stack.length > 0) {
+		const node = stack.pop();
+		if (node?.type !== "element" || require_xhtml_context.isInertElement(node.tag)) continue;
+		if (node.tag === tag) out.push(node);
+		stack.push(...node.children);
+	}
+	return out;
+}
 function readStyleResidue(html, context) {
 	const head = require_xml_query.findChildElement(html.children, "head");
 	if (head === void 0) return;
@@ -100,7 +116,7 @@ function readXhtmlBody(xml, options) {
 	};
 	const idElements = buildIdElementMap(body.children);
 	const footnoteTargetIds = /* @__PURE__ */ new Set();
-	for (const anchor of require_xml_query.elementsWithTag(body.children, "a")) {
+	for (const anchor of elementsWithTagSkippingInert(body.children, "a")) {
 		const name = require_xhtml_footnote.isFootnoteReferenceAnchor(anchor, idElements);
 		if (name !== void 0) footnoteTargetIds.add(name);
 	}
@@ -156,11 +172,11 @@ function readContainerChildren(nodes, state) {
 		if (segment.length === 0) return;
 		const inline = require_xhtml_inline.buildInlineRuns(segment, {}, state.context);
 		segment = [];
-		if (inline.runs.every((run) => run.text.trim().length === 0)) return;
+		if (inline.runs.every((run) => run.text.trim().length === 0) && inline.constructs.length === 0) return;
 		const paragraph = {
 			kind: "paragraph",
 			runs: inline.runs,
-			...inline.constructs.length > 0 ? { constructs: inline.constructs } : {}
+			...constructsField(inline)
 		};
 		blocks.push(decorateParagraph(paragraph, state));
 	};
@@ -170,7 +186,7 @@ function readContainerChildren(nodes, state) {
 			blocks.push(...readBlockElement(node, state));
 			continue;
 		}
-		if (node.type === "element" || node.type === "text") segment.push(node);
+		if (node.type === "element" || require_xml_node.isTextLikeNode(node)) segment.push(node);
 	}
 	flush();
 	return blocks;
@@ -205,7 +221,7 @@ function readBlockElementInner(element, state) {
 			kind: "paragraph",
 			headingLevel,
 			runs: inline.runs,
-			...inline.constructs.length > 0 ? { constructs: inline.constructs } : {}
+			...constructsField(inline)
 		}, state)];
 	}
 	switch (element.tag) {
@@ -220,12 +236,16 @@ function readBlockElementInner(element, state) {
 		case "ul":
 		case "ol": return readList(element, state);
 		case "dl": return readDefinitionList(element, state);
-		case "table": return [readTable(element, state)];
+		case "table": return readTable(element, state);
 		case "figure": return readContainerChildren(element.children, state);
-		case "figcaption": return [decorateParagraph({
-			kind: "paragraph",
-			runs: require_xhtml_inline.buildInlineRuns(element.children, {}, state.context).runs
-		}, state)];
+		case "figcaption": {
+			const inline = require_xhtml_inline.buildInlineRuns(element.children, {}, state.context);
+			return [decorateParagraph({
+				kind: "paragraph",
+				runs: inline.runs,
+				...constructsField(inline)
+			}, state)];
+		}
 		case "img": {
 			const block = readImage(element, state);
 			return block === void 0 ? [] : [block];
@@ -237,23 +257,123 @@ function readBlockElementInner(element, state) {
 function readPre(element, state) {
 	const languageSource = require_xml_query.findChildElement(element.children, "code") ?? element;
 	const codeLanguage = languageFromClass(require_xml_query.attrValue(languageSource, "class"));
-	const text = require_xml_entities.decodeEntities(require_xml_query.textContent(element.children));
+	const inline = containsFootnoteReference(element.children, state.context) ? readPreRuns(element.children, state.context) : {
+		runs: readPreFlatRuns(element.children, state.context),
+		constructs: []
+	};
 	return decorateParagraph({
 		kind: "paragraph",
-		runs: text.length > 0 ? [{
-			text,
-			fontFamily: require_xhtml_style_constants.MONOSPACE_FONT_FAMILY
-		}] : [],
-		...codeLanguage !== void 0 ? { codeLanguage } : {}
+		runs: inline.runs,
+		preformatted: true,
+		...codeLanguage !== void 0 ? { codeLanguage } : {},
+		...constructsField(inline)
 	}, state);
 }
+function readPreFlatRuns(nodes, context) {
+	const text = readPreText(nodes, context);
+	return text.length > 0 ? [{
+		text,
+		fontFamily: require_xhtml_style_constants.MONOSPACE_FONT_FAMILY
+	}] : [];
+}
+function containsFootnoteReference(nodes, context) {
+	for (const node of nodes) {
+		if (node.type !== "element" || require_xhtml_context.isInertElement(node.tag)) continue;
+		if (node.tag === "a" && require_xhtml_footnote.isFootnoteReferenceAnchor(node, context.idElements) !== void 0) return true;
+		if (containsFootnoteReference(node.children, context)) return true;
+	}
+	return false;
+}
+function readPreRuns(nodes, context) {
+	const runs = [];
+	const constructs = [];
+	let buffer = "";
+	const flush = () => {
+		if (buffer.length === 0) return;
+		runs.push({
+			text: buffer,
+			fontFamily: require_xhtml_style_constants.MONOSPACE_FONT_FAMILY
+		});
+		buffer = "";
+	};
+	for (const node of nodes) {
+		if (require_xml_node.isTextLikeNode(node)) {
+			buffer += require_xml_entities.decodeTextLikeNode(node);
+			continue;
+		}
+		if (node.type !== "element") continue;
+		if (require_xhtml_context.isInertElement(node.tag)) {
+			require_xhtml_context.reportInertElementSkip(node.tag, context);
+			continue;
+		}
+		if (node.tag === "img") {
+			buffer += require_xml_entities.decodeEntities(readPreImageFallbackText(node, context));
+			continue;
+		}
+		if (node.tag === "br") {
+			buffer += "\n";
+			continue;
+		}
+		const footnoteName = node.tag === "a" ? require_xhtml_footnote.isFootnoteReferenceAnchor(node, context.idElements) : void 0;
+		if (footnoteName === void 0 && !containsFootnoteReference([node], context)) {
+			buffer += readPreText(node.children, context);
+			continue;
+		}
+		flush();
+		const startRun = runs.length;
+		const nested = readPreRuns(node.children, context);
+		runs.push(...nested.runs);
+		for (const construct of nested.constructs) constructs.push({
+			...construct,
+			startRun: construct.startRun + startRun,
+			endRun: construct.endRun + startRun
+		});
+		if (footnoteName !== void 0) constructs.push({
+			descriptor: {
+				kind: "anchor",
+				anchorType: "footnote",
+				name: footnoteName
+			},
+			startRun,
+			endRun: runs.length
+		});
+	}
+	flush();
+	return {
+		runs,
+		constructs
+	};
+}
+function readPreText(nodes, context) {
+	let out = "";
+	for (const node of nodes) if (require_xml_node.isTextLikeNode(node)) out += require_xml_entities.decodeTextLikeNode(node);
+	else if (node.type === "element" && node.tag === "br") out += "\n";
+	else if (node.type === "element" && node.tag === "img") out += require_xml_entities.decodeEntities(readPreImageFallbackText(node, context));
+	else if (node.type === "element" && require_xhtml_context.isInertElement(node.tag)) {
+		require_xhtml_context.reportInertElementSkip(node.tag, context);
+		continue;
+	} else if (node.type === "element") out += readPreText(node.children, context);
+	return out;
+}
+function readPreImageFallbackText(element, context) {
+	const src = require_xml_query.attrValue(element, "src");
+	const alt = require_xml_query.attrValue(element, "alt");
+	const label = src === void 0 ? "<img>" : `<img src="${src}">`;
+	context.sink({
+		code: require_diagnostics.EpubDiagnosticCodes.IMAGE_PRE_UNSUPPORTED,
+		severity: "warning",
+		message: `${label} inside a <pre>/<code> block cannot become a real image block (this package reads a <pre> as a single text-content paragraph, so there is no block list to insert an image block into); degraded to its alt text`,
+		href: context.sourceHref
+	});
+	return alt ?? "";
+}
 function languageFromClass(className) {
 	if (className === void 0) return;
 	return /(?:^|\s)language-(\S+)/u.exec(className)?.[1];
 }
 function containsHeading(nodes) {
 	for (const node of nodes) {
-		if (node.type !== "element") continue;
+		if (node.type !== "element" || require_xhtml_context.isInertElement(node.tag)) continue;
 		if (headingLevelOf(node.tag) !== void 0) return true;
 		if (containsHeading(node.children)) return true;
 	}
@@ -279,74 +399,237 @@ function readList(element, state) {
 	});
 	const level = state.list === void 0 ? 0 : state.list.level + 1;
 	const blocks = [];
+	let previousItem;
+	let strayNodes = [];
 	for (const child of element.children) {
-		if (child.type !== "element" || child.tag !== "li") continue;
-		const itemState = withListItem(state, {
-			numId,
-			level,
-			itemId: state.minter.mintItemId()
-		});
-		blocks.push(...readContainerChildren(child.children, itemState));
+		if (child.type === "element" && child.tag === "li") {
+			blocks.push(...flushListStrayContent(strayNodes, previousItem, element.tag, state));
+			strayNodes = [];
+			previousItem = {
+				numId,
+				level,
+				itemId: state.minter.mintItemId()
+			};
+			blocks.push(...readContainerChildren(child.children, withListItem(state, previousItem)));
+			continue;
+		}
+		if (child.type === "element" && require_xhtml_context.isInertElement(child.tag)) {
+			require_xhtml_context.reportInertElementSkip(child.tag, state.context);
+			continue;
+		}
+		if (child.type === "element" || require_xml_node.isTextLikeNode(child)) strayNodes.push(child);
 	}
+	blocks.push(...flushListStrayContent(strayNodes, previousItem, element.tag, state));
+	return blocks;
+}
+function flushListStrayContent(nodes, previousItem, tag, state) {
+	if (nodes.length === 0) return [];
+	const blocks = readContainerChildren(nodes, previousItem === void 0 ? state : withListItem(state, previousItem));
+	if (blocks.length === 0) return [];
+	state.context.sink({
+		code: require_diagnostics.EpubDiagnosticCodes.LIST_CONTENT_OUTSIDE_ITEM,
+		severity: "info",
+		message: previousItem === void 0 ? `content sits directly inside a <${tag}> before its first <li> (not valid HTML5); recovered as ordinary content immediately before the list, inheriting whatever list membership its own enclosing context already carries (none, unless this <${tag}> is itself nested inside another list's <li>)` : `content sits directly inside a <${tag}> rather than inside an <li> (not valid HTML5); recovered as a continuation of the preceding <li>'s own content`,
+		href: state.context.sourceHref
+	});
 	return blocks;
 }
 function startAttr(element) {
 	return positiveIntAttr(element, "start");
 }
 function readDefinitionList(element, state) {
+	return readDefinitionListEntries(element.children, state);
+}
+function readDefinitionListEntries(nodes, state) {
 	const blocks = [];
-	for (const child of element.children) {
-		if (child.type !== "element") continue;
-		if (child.tag === "dt") {
+	let strayNodes = [];
+	const flushStray = () => {
+		blocks.push(...flushDefinitionListStrayContent(strayNodes, state));
+		strayNodes = [];
+	};
+	for (const child of nodes) {
+		if (child.type === "element" && child.tag === "dt") {
+			flushStray();
 			const inline = require_xhtml_inline.buildInlineRuns(child.children, {}, state.context);
 			blocks.push(decorateParagraph({
 				kind: "paragraph",
-				runs: inline.runs
+				runs: inline.runs,
+				...constructsField(inline)
 			}, state));
-		} else if (child.tag === "dd") {
+			continue;
+		}
+		if (child.type === "element" && child.tag === "dd") {
+			flushStray();
 			const inline = require_xhtml_inline.buildInlineRuns(child.children, {}, state.context);
 			blocks.push(decorateParagraph({
 				kind: "paragraph",
 				runs: inline.runs,
+				...constructsField(inline),
 				indentLeftPt: 36 + state.quoteDepth * 36
 			}, state));
+			continue;
+		}
+		if (child.type === "element" && child.tag === "div") {
+			flushStray();
+			blocks.push(...readDefinitionListEntries(child.children, state));
+			continue;
 		}
+		if (child.type === "element" && require_xhtml_context.isInertElement(child.tag)) {
+			require_xhtml_context.reportInertElementSkip(child.tag, state.context);
+			continue;
+		}
+		if (child.type === "element" || require_xml_node.isTextLikeNode(child)) strayNodes.push(child);
 	}
+	flushStray();
+	return blocks;
+}
+function flushDefinitionListStrayContent(nodes, state) {
+	if (nodes.length === 0) return [];
+	const blocks = readContainerChildren(nodes, state);
+	if (blocks.length === 0) return [];
+	state.context.sink({
+		code: require_diagnostics.EpubDiagnosticCodes.DEFINITION_LIST_CONTENT_OUTSIDE_ENTRY,
+		severity: "info",
+		message: "content sits directly inside a <dl> (or one of its <div> wrappers) outside any dt/dd (not valid HTML5); recovered as ordinary content -- a non-conformant wrapper's own dt/dd children, if any, lose their distinct term/definition treatment and degrade to plain concatenated text",
+		href: state.context.sourceHref
+	});
 	return blocks;
 }
 function readTable(element, state) {
+	const captionElements = element.children.filter((c) => c.type === "element" && c.tag === "caption");
 	const rows = [];
 	let columnCount = 0;
+	const strayNodes = [];
 	for (const section of element.children) {
-		if (section.type !== "element") continue;
-		const rowContainers = section.tag === "tr" ? [section] : section.tag === "thead" || section.tag === "tbody" || section.tag === "tfoot" ? section.children.filter((c) => c.type === "element" && c.tag === "tr") : [];
+		if (section.type === "element" && section.tag === "caption") continue;
+		if (section.type !== "element") {
+			if (require_xml_node.isTextLikeNode(section)) strayNodes.push(section);
+			continue;
+		}
+		if (section.tag === "colgroup") {
+			collectColgroupStrayContent(section, strayNodes, state.context);
+			continue;
+		}
+		if (require_xhtml_context.isInertElement(section.tag)) {
+			require_xhtml_context.reportInertElementSkip(section.tag, state.context);
+			continue;
+		}
+		const rowContainers = section.tag === "tr" ? [section] : section.tag === "thead" || section.tag === "tbody" || section.tag === "tfoot" ? collectRowGroupRows(section, strayNodes, state.context) : void 0;
+		if (rowContainers === void 0) {
+			strayNodes.push(section);
+			continue;
+		}
 		for (const tr of rowContainers) {
 			const cells = [];
-			for (const cellNode of tr.children) {
-				if (cellNode.type !== "element" || cellNode.tag !== "td" && cellNode.tag !== "th") continue;
-				const cellStyle = cellNode.tag === "th" ? { bold: true } : {};
-				const paragraph = {
-					kind: "paragraph",
-					runs: require_xhtml_inline.buildInlineRuns(cellNode.children, cellStyle, state.context).runs
-				};
-				const colSpan = positiveIntAttr(cellNode, "colspan");
-				const rowSpan = positiveIntAttr(cellNode, "rowspan");
-				cells.push({
-					blocks: [paragraph],
-					...colSpan !== void 0 ? { colSpan } : {},
-					...rowSpan !== void 0 ? { rowSpan } : {}
+			let strayCellNodes = [];
+			const flushStrayCell = () => {
+				if (strayCellNodes.length === 0) return;
+				const recovered = readContainerChildren(strayCellNodes, state);
+				strayCellNodes = [];
+				if (recovered.length === 0) return;
+				state.context.sink({
+					code: require_diagnostics.EpubDiagnosticCodes.TABLE_ROW_CONTENT_OUTSIDE_CELL,
+					severity: "info",
+					message: "content sits directly inside a <tr> rather than inside a <td>/<th> (not valid HTML5); recovered as its own cell in the row's own column sequence",
+					href: state.context.sourceHref
 				});
+				cells.push({ blocks: recovered });
+			};
+			for (const cellNode of tr.children) {
+				if (cellNode.type === "element" && (cellNode.tag === "td" || cellNode.tag === "th")) {
+					flushStrayCell();
+					const cellStyle = cellNode.tag === "th" ? { bold: true } : {};
+					const inline = require_xhtml_inline.buildInlineRuns(cellNode.children, cellStyle, state.context);
+					const paragraph = {
+						kind: "paragraph",
+						runs: inline.runs,
+						...constructsField(inline)
+					};
+					const colSpan = positiveIntAttr(cellNode, "colspan");
+					const rowSpan = positiveIntAttr(cellNode, "rowspan");
+					cells.push({
+						blocks: [paragraph],
+						...colSpan !== void 0 ? { colSpan } : {},
+						...rowSpan !== void 0 ? { rowSpan } : {}
+					});
+					continue;
+				}
+				if (cellNode.type === "element" && require_xhtml_context.isInertElement(cellNode.tag)) {
+					require_xhtml_context.reportInertElementSkip(cellNode.tag, state.context);
+					continue;
+				}
+				if (cellNode.type === "element" || require_xml_node.isTextLikeNode(cellNode)) strayCellNodes.push(cellNode);
 			}
+			flushStrayCell();
 			columnCount = Math.max(columnCount, cells.length);
 			rows.push({ cells });
 		}
 	}
+	const strayBlocks = strayNodes.length === 0 ? [] : readContainerChildren(strayNodes, state);
+	if (strayBlocks.length > 0) state.context.sink({
+		code: require_diagnostics.EpubDiagnosticCodes.TABLE_CONTENT_UNRECOGNIZED,
+		severity: "info",
+		message: "content sits directly inside a <table> (or one of its <thead>/<tbody>/<tfoot> row groups) outside any row, caption, or colgroup, or inside a <colgroup> itself (not valid HTML5); recovered as ordinary content immediately before the table",
+		href: state.context.sourceHref
+	});
 	const width = columnCount > 0 ? state.contentWidthPt / columnCount : state.contentWidthPt;
-	return {
+	const table = {
 		kind: "table",
 		rows,
 		columnWidthsPt: new Array(Math.max(columnCount, 1)).fill(width)
 	};
+	const captionBlocks = captionElements.flatMap((captionElement, index) => readTableCaption(captionElement, index > 0, state));
+	return [
+		...strayBlocks,
+		...captionBlocks,
+		table
+	];
+}
+function collectRowGroupRows(section, strayNodes, context) {
+	const trs = [];
+	for (const child of section.children) {
+		if (child.type === "element" && child.tag === "tr") {
+			trs.push(child);
+			continue;
+		}
+		if (child.type === "element" && require_xhtml_context.isInertElement(child.tag)) {
+			require_xhtml_context.reportInertElementSkip(child.tag, context);
+			continue;
+		}
+		if (child.type === "element" || require_xml_node.isTextLikeNode(child)) strayNodes.push(child);
+	}
+	return trs;
+}
+function collectColgroupStrayContent(section, strayNodes, context) {
+	for (const child of section.children) {
+		if (child.type === "element" && child.tag === "col") continue;
+		if (child.type === "element" && require_xhtml_context.isInertElement(child.tag)) {
+			require_xhtml_context.reportInertElementSkip(child.tag, context);
+			continue;
+		}
+		if (child.type === "element" || require_xml_node.isTextLikeNode(child)) strayNodes.push(child);
+	}
+}
+function readTableCaption(captionElement, isDuplicate, state) {
+	const captionInline = require_xhtml_inline.buildInlineRuns(captionElement.children, {}, state.context);
+	if (captionInline.runs.every((run) => run.text.trim().length === 0) && captionInline.constructs.length === 0) return [];
+	if (isDuplicate) state.context.sink({
+		code: require_diagnostics.EpubDiagnosticCodes.TABLE_DUPLICATE_CAPTION,
+		severity: "info",
+		message: "<table> carries more than one <caption> (HTML5 permits at most one); every caption beyond the first is still read as its own ordinary paragraph immediately before the table, rather than being silently discarded",
+		href: state.context.sourceHref
+	});
+	state.context.sink({
+		code: require_diagnostics.EpubDiagnosticCodes.TABLE_CAPTION_UNSUPPORTED,
+		severity: "info",
+		message: "<caption> has no document-schema.js table-caption field to carry its own distinct tag; read as an ordinary paragraph immediately before the table",
+		href: state.context.sourceHref
+	});
+	return [decorateParagraph({
+		kind: "paragraph",
+		runs: captionInline.runs,
+		...constructsField(captionInline)
+	}, state)];
 }
 function positiveIntAttr(element, name) {
 	const raw = require_xml_query.attrValue(element, name);