js.documents 4.0.20 → 4.4.5

package/dist/layout/cell-typing.d.ts

@@ -1,22 +1,2 @@
-import { ContentCellValue } from "document-schema.js";
-//#region src/layout/cell-typing.d.ts
-type CellTypeRule = 'boolean-literal' | 'iso-date' | 'named-month-date' | 'plain-number' | 'grouped-number' | 'percentage' | 'currency';
-type CellTypeDeclineReason = 'ambiguous-boolean-word' | 'ambiguous-date-order' | 'ambiguous-grouping-separator' | 'leading-zero-digits' | 'precision-loss';
-type CellTypeInferenceResult = {
-  readonly outcome: 'retyped';
-  readonly value: ContentCellValue;
-  readonly rule: CellTypeRule;
-} | {
-  readonly outcome: 'declined';
-  readonly reason: CellTypeDeclineReason;
-};
-type CellTypeInference = {
-  readonly sheetIndex: number;
-  readonly row: number;
-  readonly column: number;
-  readonly displayText: string;
-} & CellTypeInferenceResult;
-type CellTypeInferenceSink = (inference: CellTypeInference) => void;
-declare function inferCellValue(displayText: string): CellTypeInferenceResult | undefined;
-//#endregion
+import { a as CellTypeRule, i as CellTypeInferenceSink, n as CellTypeInference, o as inferCellValue, r as CellTypeInferenceResult, t as CellTypeDeclineReason } from "../cell-typing-CEWlywIc.js";
 export { CellTypeDeclineReason, CellTypeInference, CellTypeInferenceResult, CellTypeInferenceSink, CellTypeRule, inferCellValue };

package/dist/layout/reconstruct.cjs

@@ -5,7 +5,8 @@ const require_layout_cell_typing = require("./cell-typing.cjs");
 const require_layout_shared = require("./shared.cjs");
 const require_ports_abort = require("../ports/abort.cjs");
 const require_layout_lattice = require("./lattice.cjs");
-let pdf_codec = require("pdf-codec");
+let pdf_codec_fonts = require("pdf-codec/fonts");
+let pdf_codec_afm_widths = require("pdf-codec/afm-widths");
 //#region src/layout/reconstruct.ts
 const ZERO_MARGINS = {
 	topPt: 0,
@@ -31,8 +32,8 @@ function clusterIntoLines(items) {
 	return working;
 }
 function textItemVerticalExtent(item) {
-	const { standardName } = (0, pdf_codec.resolveStandardFont)(item.font.family, item.font.weight === "bold", item.font.style === "italic");
-	const metrics = pdf_codec.STANDARD_METRICS[standardName];
+	const { standardName } = (0, pdf_codec_fonts.resolveStandardFont)(item.font.family, item.font.weight === "bold", item.font.style === "italic");
+	const metrics = pdf_codec_afm_widths.STANDARD_METRICS[standardName];
 	return {
 		ascentPt: metrics.ascender / 1e3 * item.sizePt,
 		descentPt: Math.abs(metrics.descender) / 1e3 * item.sizePt
@@ -41,8 +42,8 @@ function textItemVerticalExtent(item) {
 function textItemToContentRun(item) {
 	return {
 		text: item.text,
-		bold: item.font.weight === "bold" ? true : void 0,
-		italic: item.font.style === "italic" ? true : void 0,
+		...item.font.weight === "bold" ? { bold: true } : {},
+		...item.font.style === "italic" ? { italic: true } : {},
 		fontFamily: item.font.family,
 		sizePt: item.sizePt,
 		color: item.color
@@ -133,33 +134,70 @@ function estimateModalLineSpacing(lines) {
 }
 function reconstructWordprocessing(doc, options) {
 	const signal = options?.signal;
+	doc = dropHiddenLayerContent(doc);
+	const headingLevels = headingSizeLevels(doc);
 	const sections = [];
 	let currentGroup = [];
 	let groupStartPageIndex = 0;
 	for (const page of doc.pages) {
 		require_ports_abort.throwIfAborted(signal);
 		if (currentGroup.length > 0 && !samePageSize(currentGroup[0], page)) {
-			sections.push(buildSection(currentGroup, groupStartPageIndex, doc.images));
+			sections.push(buildSection(currentGroup, groupStartPageIndex, doc.images, headingLevels, doc.form));
 			groupStartPageIndex += currentGroup.length;
 			currentGroup = [];
 		}
 		currentGroup.push(page);
 	}
-	if (currentGroup.length > 0) sections.push(buildSection(currentGroup, groupStartPageIndex, doc.images));
+	if (currentGroup.length > 0) sections.push(buildSection(currentGroup, groupStartPageIndex, doc.images, headingLevels, doc.form));
 	return {
 		kind: "wordprocessing",
 		metadata: doc.metadata,
 		sections
 	};
 }
+function layerOf(item) {
+	return item.kind === "link" || item.kind === "internalLink" ? void 0 : item.layer;
+}
+function dropHiddenLayerContent(doc) {
+	const hidden = new Set((doc.layers ?? []).filter((layer) => layer.visible !== true).map((layer) => layer.name));
+	if (hidden.size === 0) return doc;
+	return {
+		...doc,
+		pages: doc.pages.map((page) => ({
+			...page,
+			items: page.items.filter((item) => layerOf(item) === void 0 || !hidden.has(layerOf(item)))
+		}))
+	};
+}
+const HEADING_MIN_SIZE_DELTA_PT = 2;
+const MAX_HEADING_LEVEL = 6;
+function headingSizeLevels(doc) {
+	const sizes = [];
+	for (const page of doc.pages) for (const item of page.items) if (item.kind === "text") sizes.push(item.sizePt);
+	if (sizes.length === 0) return /* @__PURE__ */ new Map();
+	const bodySizePt = modeOf(sizes, .5);
+	const headingBuckets = /* @__PURE__ */ new Set();
+	for (const size of sizes) {
+		const bucket = Math.round(size / .5) * .5;
+		if (bucket - bodySizePt >= HEADING_MIN_SIZE_DELTA_PT) headingBuckets.add(bucket);
+	}
+	const levels = /* @__PURE__ */ new Map();
+	[...headingBuckets].sort((a, b) => b - a).forEach((bucket, index) => {
+		levels.set(bucket, Math.min(index + 1, MAX_HEADING_LEVEL));
+	});
+	return levels;
+}
+function headingLevelOf(paragraph, levels) {
+	return levels.get(modeOf(paragraph.lines.flatMap((line) => line.items.map((item) => item.sizePt)), .5));
+}
 function samePageSize(a, b) {
 	return a.widthPt === b.widthPt && a.heightPt === b.heightPt;
 }
-function buildSection(pages, startPageIndex, images) {
+function buildSection(pages, startPageIndex, images, headingLevels, form) {
 	const blocks = [];
 	pages.forEach((page, i) => {
 		if (i > 0) blocks.push({ kind: "pageBreak" });
-		blocks.push(...reconstructPageBlocks(page, startPageIndex + i, images));
+		blocks.push(...reconstructPageBlocks(page, startPageIndex + i, images, headingLevels, form));
 	});
 	return {
 		pageSize: {
@@ -170,7 +208,7 @@ function buildSection(pages, startPageIndex, images) {
 		blocks
 	};
 }
-function reconstructPageBlocks(page, pageIndex, images) {
+function reconstructPageBlocks(page, pageIndex, images, headingLevels, form) {
 	const recoveredTable = recoverTable(page, pageIndex);
 	const consumedText = recoveredTable?.consumedText;
 	const textItems = page.items.filter((i) => i.kind === "text" && consumedText?.has(i) !== true);
@@ -179,7 +217,7 @@ function reconstructPageBlocks(page, pageIndex, images) {
 	const positioned = [];
 	for (const paragraph of paragraphs) positioned.push({
 		yPt: paragraph.lines[0].baselineY,
-		block: paragraphToContentParagraph(paragraph, pageIndex)
+		block: paragraphToContentParagraph(paragraph, pageIndex, headingLevelOf(paragraph, headingLevels))
 	});
 	for (const img of imageItems) {
 		const asset = images[img.imageId];
@@ -213,7 +251,121 @@ function reconstructPageBlocks(page, pageIndex, images) {
 		block: recoveredVectors.block
 	});
 	positioned.sort((a, b) => b.yPt - a.yPt);
-	return positioned.map((p) => p.block);
+	const blocks = positioned.map((p) => p.block);
+	reconcileLinks(blocks, page, pageIndex);
+	appendAnnotationConstructs(blocks, page, pageIndex);
+	appendFormFieldConstructs(blocks, form, pageIndex);
+	return blocks;
+}
+function framesIntersect(a, b) {
+	return a.xPt < b.xPt + b.widthPt && b.xPt < a.xPt + a.widthPt && a.yPt < b.yPt + b.heightPt && b.yPt < a.yPt + a.heightPt;
+}
+function intersectionArea(a, b) {
+	const width = Math.min(a.xPt + a.widthPt, b.xPt + b.widthPt) - Math.max(a.xPt, b.xPt);
+	const height = Math.min(a.yPt + a.heightPt, b.yPt + b.heightPt) - Math.max(a.yPt, b.yPt);
+	return width > 0 && height > 0 ? width * height : 0;
+}
+function reconcileLinks(blocks, page, pageIndex) {
+	for (const item of page.items) {
+		if (item.kind !== "link" && item.kind !== "internalLink") continue;
+		if (item.kind === "link") {
+			let matchedRun = false;
+			for (const block of blocks) {
+				if (block.kind !== "paragraph") continue;
+				for (const run of block.runs) if ((run.frames ?? []).some((frame) => frame.pageIndex === pageIndex && framesIntersect(item, frame))) {
+					run.hyperlink = item.uri;
+					matchedRun = true;
+				}
+			}
+			if (matchedRun) continue;
+		}
+		wrapBestBlock(blocks, item, pageIndex, linkDescriptor(item));
+	}
+}
+function linkDescriptor(item) {
+	if (item.kind === "link") return {
+		kind: "link",
+		target: {
+			kind: "external",
+			uri: item.uri
+		},
+		...item.title !== void 0 ? { title: item.title } : {}
+	};
+	return {
+		kind: "link",
+		target: {
+			kind: "internal",
+			anchor: item.destination
+		},
+		...item.title !== void 0 ? { title: item.title } : {}
+	};
+}
+function blockFramesOn(block, pageIndex) {
+	return block.kind === "constructStart" || block.kind === "constructEnd" ? [] : (block.frames ?? []).filter((frame) => frame.pageIndex === pageIndex);
+}
+function wrapBestBlock(blocks, rect, pageIndex, descriptor) {
+	let bestIndex = -1;
+	let bestArea = 0;
+	blocks.forEach((block, index) => {
+		const area = blockFramesOn(block, pageIndex).reduce((sum, frame) => sum + intersectionArea(rect, frame), 0);
+		if (area > bestArea) {
+			bestArea = area;
+			bestIndex = index;
+		}
+	});
+	const start = {
+		kind: "constructStart",
+		descriptor
+	};
+	const end = { kind: "constructEnd" };
+	if (bestIndex >= 0) {
+		blocks.splice(bestIndex, 0, start);
+		blocks.splice(bestIndex + 2, 0, end);
+	} else blocks.push(start, end);
+}
+function appendAnnotationConstructs(blocks, page, pageIndex) {
+	(page.annotations ?? []).forEach((annotation, annotIndex) => {
+		const key = `pdf-annot-${String(pageIndex)}-${String(annotIndex)}`;
+		blocks.push({
+			kind: "constructStart",
+			descriptor: {
+				kind: "anchor",
+				anchorType: "comment",
+				name: key,
+				definition: key,
+				...annotation.source !== void 0 ? { source: annotation.source } : {}
+			}
+		}, { kind: "constructEnd" });
+	});
+}
+function controlTypeOf(field) {
+	if (field.fieldType === "text") return "plainText";
+	if (field.fieldType === "button") return "button";
+	if (field.fieldType === "combobox") return "comboBox";
+	if (field.fieldType === "listbox") return "dropDown";
+	return "checkbox";
+}
+function appendFormFieldConstructs(blocks, form, pageIndex) {
+	const visit = (fields) => {
+		for (const field of fields) {
+			visit(field.children);
+			if (field.fieldType === "group" || field.fieldType === "signature") continue;
+			for (const widget of field.widgets) {
+				if (widget.pageIndex !== pageIndex) continue;
+				wrapBestBlock(blocks, widget, pageIndex, {
+					kind: "contentControl",
+					controlType: controlTypeOf(field),
+					tag: field.name,
+					...field.alias !== void 0 ? { alias: field.alias } : {},
+					...field.readOnly === true ? { lock: "content" } : {},
+					...field.value !== void 0 ? { value: field.value } : {},
+					...field.checked !== void 0 ? { checked: field.checked } : {},
+					...field.options !== void 0 ? { options: [...field.options] } : {}
+				});
+			}
+		}
+	};
+	visit(form ?? []);
 }
 function clusterIntoParagraphs(lines) {
 	if (lines.length === 0) return [];
@@ -230,6 +382,7 @@ function clusterIntoParagraphs(lines) {
 }
 function startsNewParagraph(prev, next, modalSpacing, dominantLeftX) {
 	if (prev.baselineY - next.baselineY > PARAGRAPH_GAP_MULTIPLIER * modalSpacing) return true;
+	if (!fontSizesClose(prev.items[0].sizePt, next.items[0].sizePt)) return true;
 	const nextLeft = next.items[0].xPt;
 	const prevLeft = prev.items[0].xPt;
 	const emPt = next.items[0].sizePt;
@@ -238,12 +391,16 @@ function startsNewParagraph(prev, next, modalSpacing, dominantLeftX) {
 	return nextIndented && prevAtMargin;
 }
 const LEFT_ALIGN_TOLERANCE_PT = 2;
-function paragraphToContentParagraph(paragraph, pageIndex) {
+function paragraphToContentParagraph(paragraph, pageIndex, headingLevel) {
 	const dominantLeftX = modeOf(paragraph.lines.map((l) => l.items[0].xPt), 1);
 	const result = {
 		kind: "paragraph",
 		runs: [],
-		alignment: paragraph.lines.every((l) => Math.abs(l.items[0].xPt - dominantLeftX) <= LEFT_ALIGN_TOLERANCE_PT) ? "left" : void 0
+		alignment: paragraph.lines.every((l) => Math.abs(l.items[0].xPt - dominantLeftX) <= LEFT_ALIGN_TOLERANCE_PT) ? "left" : void 0,
+		...headingLevel !== void 0 ? {
+			styleId: `Heading${String(headingLevel)}`,
+			headingLevel
+		} : {}
 	};
 	paragraph.lines.forEach((line, lineIndex) => {
 		require_layout_shared.stampFrame(result, pageIndex, lineBox(line, pageIndex));
@@ -253,6 +410,12 @@ function paragraphToContentParagraph(paragraph, pageIndex) {
 		}
 		pushRunsForLine(result.runs, line, pageIndex);
 	});
+	if (headingLevel !== void 0) result.runs = result.runs.map((run) => {
+		if (run.bold !== true) return run;
+		const withoutBold = { ...run };
+		delete withoutBold.bold;
+		return withoutBold;
+	});
 	return result;
 }
 function reconstructPresentation(doc, options) {

package/dist/layout/reconstruct.d.cts

@@ -1,4 +1,4 @@
-import { CellTypeInferenceSink } from "./cell-typing.cjs";
+import { i as CellTypeInferenceSink } from "../cell-typing-CEWlywIc.cjs";
 import { ContentDocument } from "document-schema.js";
 import { LayoutDocument } from "pdf-codec";
 //#region src/layout/reconstruct.d.ts

package/dist/layout/reconstruct.d.ts

@@ -1,4 +1,4 @@
-import { CellTypeInferenceSink } from "./cell-typing.js";
+import { i as CellTypeInferenceSink } from "../cell-typing-CEWlywIc.js";
 import { ContentDocument } from "document-schema.js";
 import { LayoutDocument } from "pdf-codec";
 //#region src/layout/reconstruct.d.ts

package/dist/layout/reconstruct.js

@@ -4,7 +4,8 @@ import { inferCellValue } from "./cell-typing.js";
 import { stampFrame } from "./shared.js";
 import { throwIfAborted } from "../ports/abort.js";
 import { detectGridLattice, findColumnIndex, findRowIndex } from "./lattice.js";
-import { STANDARD_METRICS, resolveStandardFont } from "pdf-codec";
+import { resolveStandardFont } from "pdf-codec/fonts";
+import { STANDARD_METRICS } from "pdf-codec/afm-widths";
 //#region src/layout/reconstruct.ts
 const ZERO_MARGINS = {
 	topPt: 0,
@@ -40,8 +41,8 @@ function textItemVerticalExtent(item) {
 function textItemToContentRun(item) {
 	return {
 		text: item.text,
-		bold: item.font.weight === "bold" ? true : void 0,
-		italic: item.font.style === "italic" ? true : void 0,
+		...item.font.weight === "bold" ? { bold: true } : {},
+		...item.font.style === "italic" ? { italic: true } : {},
 		fontFamily: item.font.family,
 		sizePt: item.sizePt,
 		color: item.color
@@ -132,33 +133,70 @@ function estimateModalLineSpacing(lines) {
 }
 function reconstructWordprocessing(doc, options) {
 	const signal = options?.signal;
+	doc = dropHiddenLayerContent(doc);
+	const headingLevels = headingSizeLevels(doc);
 	const sections = [];
 	let currentGroup = [];
 	let groupStartPageIndex = 0;
 	for (const page of doc.pages) {
 		throwIfAborted(signal);
 		if (currentGroup.length > 0 && !samePageSize(currentGroup[0], page)) {
-			sections.push(buildSection(currentGroup, groupStartPageIndex, doc.images));
+			sections.push(buildSection(currentGroup, groupStartPageIndex, doc.images, headingLevels, doc.form));
 			groupStartPageIndex += currentGroup.length;
 			currentGroup = [];
 		}
 		currentGroup.push(page);
 	}
-	if (currentGroup.length > 0) sections.push(buildSection(currentGroup, groupStartPageIndex, doc.images));
+	if (currentGroup.length > 0) sections.push(buildSection(currentGroup, groupStartPageIndex, doc.images, headingLevels, doc.form));
 	return {
 		kind: "wordprocessing",
 		metadata: doc.metadata,
 		sections
 	};
 }
+function layerOf(item) {
+	return item.kind === "link" || item.kind === "internalLink" ? void 0 : item.layer;
+}
+function dropHiddenLayerContent(doc) {
+	const hidden = new Set((doc.layers ?? []).filter((layer) => layer.visible !== true).map((layer) => layer.name));
+	if (hidden.size === 0) return doc;
+	return {
+		...doc,
+		pages: doc.pages.map((page) => ({
+			...page,
+			items: page.items.filter((item) => layerOf(item) === void 0 || !hidden.has(layerOf(item)))
+		}))
+	};
+}
+const HEADING_MIN_SIZE_DELTA_PT = 2;
+const MAX_HEADING_LEVEL = 6;
+function headingSizeLevels(doc) {
+	const sizes = [];
+	for (const page of doc.pages) for (const item of page.items) if (item.kind === "text") sizes.push(item.sizePt);
+	if (sizes.length === 0) return /* @__PURE__ */ new Map();
+	const bodySizePt = modeOf(sizes, .5);
+	const headingBuckets = /* @__PURE__ */ new Set();
+	for (const size of sizes) {
+		const bucket = Math.round(size / .5) * .5;
+		if (bucket - bodySizePt >= HEADING_MIN_SIZE_DELTA_PT) headingBuckets.add(bucket);
+	}
+	const levels = /* @__PURE__ */ new Map();
+	[...headingBuckets].sort((a, b) => b - a).forEach((bucket, index) => {
+		levels.set(bucket, Math.min(index + 1, MAX_HEADING_LEVEL));
+	});
+	return levels;
+}
+function headingLevelOf(paragraph, levels) {
+	return levels.get(modeOf(paragraph.lines.flatMap((line) => line.items.map((item) => item.sizePt)), .5));
+}
 function samePageSize(a, b) {
 	return a.widthPt === b.widthPt && a.heightPt === b.heightPt;
 }
-function buildSection(pages, startPageIndex, images) {
+function buildSection(pages, startPageIndex, images, headingLevels, form) {
 	const blocks = [];
 	pages.forEach((page, i) => {
 		if (i > 0) blocks.push({ kind: "pageBreak" });
-		blocks.push(...reconstructPageBlocks(page, startPageIndex + i, images));
+		blocks.push(...reconstructPageBlocks(page, startPageIndex + i, images, headingLevels, form));
 	});
 	return {
 		pageSize: {
@@ -169,7 +207,7 @@ function buildSection(pages, startPageIndex, images) {
 		blocks
 	};
 }
-function reconstructPageBlocks(page, pageIndex, images) {
+function reconstructPageBlocks(page, pageIndex, images, headingLevels, form) {
 	const recoveredTable = recoverTable(page, pageIndex);
 	const consumedText = recoveredTable?.consumedText;
 	const textItems = page.items.filter((i) => i.kind === "text" && consumedText?.has(i) !== true);
@@ -178,7 +216,7 @@ function reconstructPageBlocks(page, pageIndex, images) {
 	const positioned = [];
 	for (const paragraph of paragraphs) positioned.push({
 		yPt: paragraph.lines[0].baselineY,
-		block: paragraphToContentParagraph(paragraph, pageIndex)
+		block: paragraphToContentParagraph(paragraph, pageIndex, headingLevelOf(paragraph, headingLevels))
 	});
 	for (const img of imageItems) {
 		const asset = images[img.imageId];
@@ -212,7 +250,121 @@ function reconstructPageBlocks(page, pageIndex, images) {
 		block: recoveredVectors.block
 	});
 	positioned.sort((a, b) => b.yPt - a.yPt);
-	return positioned.map((p) => p.block);
+	const blocks = positioned.map((p) => p.block);
+	reconcileLinks(blocks, page, pageIndex);
+	appendAnnotationConstructs(blocks, page, pageIndex);
+	appendFormFieldConstructs(blocks, form, pageIndex);
+	return blocks;
+}
+function framesIntersect(a, b) {
+	return a.xPt < b.xPt + b.widthPt && b.xPt < a.xPt + a.widthPt && a.yPt < b.yPt + b.heightPt && b.yPt < a.yPt + a.heightPt;
+}
+function intersectionArea(a, b) {
+	const width = Math.min(a.xPt + a.widthPt, b.xPt + b.widthPt) - Math.max(a.xPt, b.xPt);
+	const height = Math.min(a.yPt + a.heightPt, b.yPt + b.heightPt) - Math.max(a.yPt, b.yPt);
+	return width > 0 && height > 0 ? width * height : 0;
+}
+function reconcileLinks(blocks, page, pageIndex) {
+	for (const item of page.items) {
+		if (item.kind !== "link" && item.kind !== "internalLink") continue;
+		if (item.kind === "link") {
+			let matchedRun = false;
+			for (const block of blocks) {
+				if (block.kind !== "paragraph") continue;
+				for (const run of block.runs) if ((run.frames ?? []).some((frame) => frame.pageIndex === pageIndex && framesIntersect(item, frame))) {
+					run.hyperlink = item.uri;
+					matchedRun = true;
+				}
+			}
+			if (matchedRun) continue;
+		}
+		wrapBestBlock(blocks, item, pageIndex, linkDescriptor(item));
+	}
+}
+function linkDescriptor(item) {
+	if (item.kind === "link") return {
+		kind: "link",
+		target: {
+			kind: "external",
+			uri: item.uri
+		},
+		...item.title !== void 0 ? { title: item.title } : {}
+	};
+	return {
+		kind: "link",
+		target: {
+			kind: "internal",
+			anchor: item.destination
+		},
+		...item.title !== void 0 ? { title: item.title } : {}
+	};
+}
+function blockFramesOn(block, pageIndex) {
+	return block.kind === "constructStart" || block.kind === "constructEnd" ? [] : (block.frames ?? []).filter((frame) => frame.pageIndex === pageIndex);
+}
+function wrapBestBlock(blocks, rect, pageIndex, descriptor) {
+	let bestIndex = -1;
+	let bestArea = 0;
+	blocks.forEach((block, index) => {
+		const area = blockFramesOn(block, pageIndex).reduce((sum, frame) => sum + intersectionArea(rect, frame), 0);
+		if (area > bestArea) {
+			bestArea = area;
+			bestIndex = index;
+		}
+	});
+	const start = {
+		kind: "constructStart",
+		descriptor
+	};
+	const end = { kind: "constructEnd" };
+	if (bestIndex >= 0) {
+		blocks.splice(bestIndex, 0, start);
+		blocks.splice(bestIndex + 2, 0, end);
+	} else blocks.push(start, end);
+}
+function appendAnnotationConstructs(blocks, page, pageIndex) {
+	(page.annotations ?? []).forEach((annotation, annotIndex) => {
+		const key = `pdf-annot-${String(pageIndex)}-${String(annotIndex)}`;
+		blocks.push({
+			kind: "constructStart",
+			descriptor: {
+				kind: "anchor",
+				anchorType: "comment",
+				name: key,
+				definition: key,
+				...annotation.source !== void 0 ? { source: annotation.source } : {}
+			}
+		}, { kind: "constructEnd" });
+	});
+}
+function controlTypeOf(field) {
+	if (field.fieldType === "text") return "plainText";
+	if (field.fieldType === "button") return "button";
+	if (field.fieldType === "combobox") return "comboBox";
+	if (field.fieldType === "listbox") return "dropDown";
+	return "checkbox";
+}
+function appendFormFieldConstructs(blocks, form, pageIndex) {
+	const visit = (fields) => {
+		for (const field of fields) {
+			visit(field.children);
+			if (field.fieldType === "group" || field.fieldType === "signature") continue;
+			for (const widget of field.widgets) {
+				if (widget.pageIndex !== pageIndex) continue;
+				wrapBestBlock(blocks, widget, pageIndex, {
+					kind: "contentControl",
+					controlType: controlTypeOf(field),
+					tag: field.name,
+					...field.alias !== void 0 ? { alias: field.alias } : {},
+					...field.readOnly === true ? { lock: "content" } : {},
+					...field.value !== void 0 ? { value: field.value } : {},
+					...field.checked !== void 0 ? { checked: field.checked } : {},
+					...field.options !== void 0 ? { options: [...field.options] } : {}
+				});
+			}
+		}
+	};
+	visit(form ?? []);
 }
 function clusterIntoParagraphs(lines) {
 	if (lines.length === 0) return [];
@@ -229,6 +381,7 @@ function clusterIntoParagraphs(lines) {
 }
 function startsNewParagraph(prev, next, modalSpacing, dominantLeftX) {
 	if (prev.baselineY - next.baselineY > PARAGRAPH_GAP_MULTIPLIER * modalSpacing) return true;
+	if (!fontSizesClose(prev.items[0].sizePt, next.items[0].sizePt)) return true;
 	const nextLeft = next.items[0].xPt;
 	const prevLeft = prev.items[0].xPt;
 	const emPt = next.items[0].sizePt;
@@ -237,12 +390,16 @@ function startsNewParagraph(prev, next, modalSpacing, dominantLeftX) {
 	return nextIndented && prevAtMargin;
 }
 const LEFT_ALIGN_TOLERANCE_PT = 2;
-function paragraphToContentParagraph(paragraph, pageIndex) {
+function paragraphToContentParagraph(paragraph, pageIndex, headingLevel) {
 	const dominantLeftX = modeOf(paragraph.lines.map((l) => l.items[0].xPt), 1);
 	const result = {
 		kind: "paragraph",
 		runs: [],
-		alignment: paragraph.lines.every((l) => Math.abs(l.items[0].xPt - dominantLeftX) <= LEFT_ALIGN_TOLERANCE_PT) ? "left" : void 0
+		alignment: paragraph.lines.every((l) => Math.abs(l.items[0].xPt - dominantLeftX) <= LEFT_ALIGN_TOLERANCE_PT) ? "left" : void 0,
+		...headingLevel !== void 0 ? {
+			styleId: `Heading${String(headingLevel)}`,
+			headingLevel
+		} : {}
 	};
 	paragraph.lines.forEach((line, lineIndex) => {
 		stampFrame(result, pageIndex, lineBox(line, pageIndex));
@@ -252,6 +409,12 @@ function paragraphToContentParagraph(paragraph, pageIndex) {
 		}
 		pushRunsForLine(result.runs, line, pageIndex);
 	});
+	if (headingLevel !== void 0) result.runs = result.runs.map((run) => {
+		if (run.bold !== true) return run;
+		const withoutBold = { ...run };
+		delete withoutBold.bold;
+		return withoutBold;
+	});
 	return result;
 }
 function reconstructPresentation(doc, options) {

package/dist/layout/shared.cjs

@@ -3,8 +3,8 @@ const require_mathml_layout = require("../mathml/layout.cjs");
 const require_layout_text_layout = require("./text-layout.cjs");
 let ooxml_js = require("ooxml.js");
 let document_schema_js = require("document-schema.js");
-let pdf_codec = require("pdf-codec");
 let byte_codec = require("byte-codec");
+let pdf_codec_layout = require("pdf-codec/layout");
 //#region src/layout/shared.ts
 function stampFrame(node, pageIndex, box) {
 	const frame = {
@@ -39,7 +39,7 @@ function packagePagesOf(pages) {
 }
 function layoutDocumentOf(metadata, pages, images) {
 	return {
-		formatVersion: pdf_codec.LAYOUT_FORMAT_VERSION,
+		formatVersion: pdf_codec_layout.LAYOUT_FORMAT_VERSION,
 		metadata,
 		pages,
 		images

package/dist/layout/shared.js

@@ -2,8 +2,8 @@ import { layoutFormula } from "../mathml/layout.js";
 import { wrapRunsToWidth } from "./text-layout.js";
 import { base64ToBytes } from "ooxml.js";
 import { COLOR_BLACK, DEFAULT_LAYOUT_FONT } from "document-schema.js";
-import { LAYOUT_FORMAT_VERSION } from "pdf-codec";
 import { crc32, decodePng, readJpegInfo } from "byte-codec";
+import { LAYOUT_FORMAT_VERSION } from "pdf-codec/layout";
 //#region src/layout/shared.ts
 function stampFrame(node, pageIndex, box) {
 	const frame = {

package/dist/markdown/math.cjs

@@ -3,7 +3,6 @@ const require_model_formula = require("../model/formula.cjs");
 const require_latex_symbols = require("../latex/symbols.cjs");
 const require_latex_lower = require("../latex/lower.cjs");
 //#region src/markdown/math.ts
-const MATH_BLOCK_STYLE_ID = "MathBlock";
 const MATH_INLINE_FONT_MARKER = "Cambria Math";
 const STAND_IN_FRAME = {
 	xPt: 0,
@@ -62,14 +61,13 @@ function lowerMarkdownMath(document, options) {
 				});
 				continue;
 			}
-			if (block.kind !== "paragraph") {
-				out.push(block);
+			if (block.kind === "embeddedObject" && block.objectKind === "formula" && block.document.kind === "formula" && block.document.formula.presentation !== void 0) {
+				const formulaBlock = lowerOne(block.document.formula.presentation.latex, MATH_BLOCK_SOURCE);
+				out.push(formulaBlock ?? block);
 				continue;
 			}
-			if (block.styleId === MATH_BLOCK_STYLE_ID) {
-				const latex = block.runs.map((run) => run.text).join("");
-				const formulaBlock = lowerOne(latex, MATH_BLOCK_SOURCE);
-				out.push(formulaBlock ?? block);
+			if (block.kind !== "paragraph") {
+				out.push(block);
 				continue;
 			}
 			const extraction = extractInlineMath(block.runs);