js.documents 14.4.56 → 14.4.57

package/dist/latex/temml.d.ts

@@ -1,25 +1,2 @@
-import { MathMlNode } from "document-schema.js";
-//#region src/latex/temml.d.ts
-interface TemmlNode {
-  readonly type: string;
-  readonly [field: string]: unknown;
-}
-declare function isTemmlNode(value: unknown): value is TemmlNode;
-declare function isTemmlNodeArray(value: unknown): value is TemmlNode[];
-interface TemmlSourceSpan {
-  readonly start: number;
-  readonly end: number;
-}
-declare function sourceSpanOf(node: TemmlNode): TemmlSourceSpan | undefined;
-declare function sourceSlice(input: string, span: TemmlSourceSpan | undefined): string;
-type LatexParseResult = {
-  readonly status: "parsed";
-  readonly nodes: readonly TemmlNode[];
-  readonly mathml: readonly MathMlNode[];
-} | {
-  readonly status: "unparseable";
-  readonly message: string;
-};
-declare function parseLatex(latex: string): LatexParseResult;
-//#endregion
+import { a as isTemmlNodeArray, c as sourceSpanOf, i as isTemmlNode, n as TemmlNode, o as parseLatex, r as TemmlSourceSpan, s as sourceSlice, t as LatexParseResult } from "../temml-1-I9N9DD.js";
 export { LatexParseResult, TemmlNode, TemmlSourceSpan, isTemmlNode, isTemmlNodeArray, parseLatex, sourceSlice, sourceSpanOf };

package/dist/layout/engine.cjs

@@ -1,7 +1,7 @@
 Object.defineProperty(exports, Symbol.toStringTag, { value: "Module" });
 const require_model_geometry = require("../model/geometry.cjs");
 const require_model_formula = require("../model/formula.cjs");
-const require_mathml_layout = require("../mathml/layout.cjs");
+const require_layout = require("../layout-iuiyl_Fk.cjs");
 const require_layout_text_layout = require("./text-layout.cjs");
 const require_layout_shared = require("./shared.cjs");
 let document_schema_js = require("document-schema.js");
@@ -229,7 +229,7 @@ function layoutFormulaFlow(block, section, pages, state, contentLeftXDown, conte
 	}
 	const sizePt = require_layout_shared.formulaSizePtForFrame(formula.mathml, block.frame, mathMetricsAt);
 	const metrics = mathMetricsAt(sizePt);
-	const { box } = require_mathml_layout.layoutFormula(formula.mathml, {
+	const { box } = require_layout.layoutFormula(formula.mathml, {
 		metrics,
 		sizePt,
 		color: document_schema_js.COLOR_BLACK

package/dist/layout/engine.js

@@ -1,6 +1,6 @@
 import { flipY } from "../model/geometry.js";
 import { formulaOfBlock, formulaPlaceholderText } from "../model/formula.js";
-import { layoutFormula } from "../mathml/layout.js";
+import { t as layoutFormula } from "../layout-DIJzJk4x.js";
 import { wrapRunsToWidth } from "./text-layout.js";
 import { alignmentOffsetPt, effectiveStyledRuns, estimateRowHeightPt, formulaSizePtForFrame, headingStyleFor, justifyLineGapsPt, layoutDocumentOf, lineNaturalHeightPt, packagePagesOf, pushCellBorderLines, registerImage, stampFragmentFrame, stampFrame, tableAnchorBoxes, textBoxForFragment } from "./shared.js";
 import { COLOR_BLACK, resolveCellFillColor } from "document-schema.js";

package/dist/layout/reconstruct-lines.cjs

@@ -0,0 +1,315 @@
+Object.defineProperty(exports, Symbol.toStringTag, { value: "Module" });
+const require_layout_shared = require("./shared.cjs");
+let pdf_codec_fonts = require("pdf-codec/fonts");
+let pdf_codec_afm_widths = require("pdf-codec/afm-widths");
+let pdf_codec_text_group = require("pdf-codec/text-group");
+//#region src/layout/reconstruct-lines.ts
+const LINE_BASELINE_TOLERANCE_FACTOR = .5;
+const AFM_UNITS_PER_EM = 1e3;
+const HALF_POINT_BUCKET_PT = .5;
+function sharesBaseline(anchor, item) {
+	return (0, pdf_codec_text_group.runsShareBaseline)(anchor, item, { baselineToleranceEm: LINE_BASELINE_TOLERANCE_FACTOR });
+}
+const DUPLICATE_PAINT_BUCKET_PT = .1;
+function duplicatePaintKey(item) {
+	const xBucket = Math.round(item.xPt / DUPLICATE_PAINT_BUCKET_PT);
+	const yBucket = Math.round(item.yPt / DUPLICATE_PAINT_BUCKET_PT);
+	return `${item.text.trim()}|${String(xBucket)}|${String(yBucket)}`;
+}
+function dropDuplicatePaints(items) {
+	const seen = /* @__PURE__ */ new Set();
+	const kept = [];
+	for (const item of items) {
+		const key = duplicatePaintKey(item);
+		if (seen.has(key)) continue;
+		seen.add(key);
+		kept.push(item);
+	}
+	return kept;
+}
+function dropOverlappingRepeatsWithinLine(items) {
+	const rightEdgeByText = /* @__PURE__ */ new Map();
+	const kept = [];
+	for (const item of items) {
+		const textKey = item.text.trim();
+		const priorRightEdge = rightEdgeByText.get(textKey);
+		if (priorRightEdge !== void 0 && item.xPt < priorRightEdge) continue;
+		rightEdgeByText.set(textKey, item.xPt + (item.widthPt ?? 0));
+		kept.push(item);
+	}
+	return kept;
+}
+const REDRAW_PASS_REWIND_TOLERANCE_PT = 2;
+const MIN_REDRAW_PASS_TEXT_LENGTH = 12;
+const REDRAW_PASS_SIMILARITY_THRESHOLD = .6;
+function groupByBaselineInDocumentOrder(items) {
+	const groups = [];
+	for (const item of items) {
+		const group = groups.find((g) => sharesBaseline(g.anchor, item));
+		if (group === void 0) groups.push({
+			anchor: item,
+			items: [item]
+		});
+		else group.items.push(item);
+	}
+	return groups.map((g) => g.items);
+}
+function splitIntoPasses(group) {
+	const passes = [];
+	for (const item of group) {
+		const current = passes.at(-1);
+		const itemRightPt = item.xPt + (item.widthPt ?? 0);
+		if (current !== void 0 && item.xPt >= current.endXPt - REDRAW_PASS_REWIND_TOLERANCE_PT) {
+			current.items.push(item);
+			current.endXPt = Math.max(current.endXPt, itemRightPt);
+		} else passes.push({
+			items: [item],
+			startXPt: item.xPt,
+			endXPt: itemRightPt
+		});
+	}
+	return passes;
+}
+function passText(pass) {
+	return pass.items.map((item) => item.text).join("").trim();
+}
+function passesOverlap(a, b) {
+	return a.startXPt < b.endXPt && b.startXPt < a.endXPt;
+}
+function levenshteinDistance(a, b) {
+	const memo = /* @__PURE__ */ new Map();
+	function distance(i, j) {
+		if (i === 0) return j;
+		if (j === 0) return i;
+		const key = `${String(i)}:${String(j)}`;
+		const cached = memo.get(key);
+		if (cached !== void 0) return cached;
+		const cost = a.charAt(i - 1) === b.charAt(j - 1) ? 0 : 1;
+		const result = Math.min(distance(i - 1, j) + 1, distance(i, j - 1) + 1, distance(i - 1, j - 1) + cost);
+		memo.set(key, result);
+		return result;
+	}
+	return distance(a.length, b.length);
+}
+function textSimilarity(a, b) {
+	const maxLength = Math.max(a.length, b.length);
+	return 1 - levenshteinDistance(a, b) / maxLength;
+}
+function dropFuzzyRedrawnPasses(items) {
+	const toDrop = /* @__PURE__ */ new Set();
+	for (const group of groupByBaselineInDocumentOrder(items)) {
+		const passes = splitIntoPasses(group);
+		if (passes.length < 2) continue;
+		const [first, ...rest] = passes;
+		if (first === void 0) continue;
+		const firstText = passText(first);
+		if (firstText.length < MIN_REDRAW_PASS_TEXT_LENGTH) continue;
+		for (const candidate of rest) {
+			const candidateText = passText(candidate);
+			if (candidateText.length < MIN_REDRAW_PASS_TEXT_LENGTH) continue;
+			if (!passesOverlap(first, candidate)) continue;
+			if (textSimilarity(firstText, candidateText) < REDRAW_PASS_SIMILARITY_THRESHOLD) continue;
+			for (const item of candidate.items) toDrop.add(item);
+		}
+	}
+	if (toDrop.size === 0) return items;
+	return items.filter((item) => !toDrop.has(item));
+}
+function clusterIntoLines(items) {
+	const sorted = [...dropDuplicatePaints(dropFuzzyRedrawnPasses(items))].sort((a, b) => b.yPt - a.yPt || a.xPt - b.xPt);
+	const working = [];
+	for (const item of sorted) {
+		const line = working.find((l) => sharesBaseline(l.anchor, item));
+		if (line === void 0) working.push({
+			items: [item],
+			baselineY: item.yPt,
+			anchor: item
+		});
+		else line.items.push(item);
+	}
+	for (const line of working) {
+		line.items.sort((a, b) => a.xPt - b.xPt);
+		line.items = dropOverlappingRepeatsWithinLine(line.items);
+	}
+	working.sort((a, b) => b.baselineY - a.baselineY);
+	return working;
+}
+function textItemVerticalExtent(item) {
+	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 / AFM_UNITS_PER_EM * item.sizePt,
+		descentPt: Math.abs(metrics.descender) / AFM_UNITS_PER_EM * item.sizePt
+	};
+}
+function textItemToContentRun(item) {
+	return {
+		text: item.text,
+		...item.font.weight === "bold" ? { bold: true } : {},
+		...item.font.style === "italic" ? { italic: true } : {},
+		fontFamily: item.font.family,
+		sizePt: item.sizePt,
+		color: item.color
+	};
+}
+const MIN_WORD_GAP_PT = .5;
+function gapExceeds(previous, next, thresholdPt) {
+	const gap = (0, pdf_codec_text_group.runGapPt)(previous, next);
+	return gap !== void 0 && gap > thresholdPt;
+}
+function textBoxOfItem(item, pageIndex) {
+	const { ascentPt, descentPt } = textItemVerticalExtent(item);
+	return {
+		pageIndex,
+		xPt: item.xPt,
+		yPt: item.yPt - descentPt,
+		widthPt: item.widthPt ?? 0,
+		heightPt: ascentPt + descentPt
+	};
+}
+function lineBox(line, pageIndex) {
+	let minX = Number.POSITIVE_INFINITY;
+	let maxX = Number.NEGATIVE_INFINITY;
+	let minY = Number.POSITIVE_INFINITY;
+	let maxY = Number.NEGATIVE_INFINITY;
+	for (const item of line.items) {
+		const { ascentPt, descentPt } = textItemVerticalExtent(item);
+		minX = Math.min(minX, item.xPt);
+		maxX = Math.max(maxX, item.xPt + (item.widthPt ?? 0));
+		minY = Math.min(minY, item.yPt - descentPt);
+		maxY = Math.max(maxY, item.yPt + ascentPt);
+	}
+	return {
+		pageIndex,
+		xPt: minX,
+		yPt: minY,
+		widthPt: maxX - minX,
+		heightPt: maxY - minY
+	};
+}
+function pushRunsForLine(runs, line, pageIndex) {
+	line.items.forEach((item, itemIndex) => {
+		if (itemIndex > 0) {
+			const prevItem = line.items[itemIndex - 1];
+			if (gapExceeds(prevItem, item, 2 * item.sizePt)) runs.push({ text: "	" });
+			else if (gapExceeds(prevItem, item, .5)) runs[runs.length - 1].text += " ";
+		}
+		const run = textItemToContentRun(item);
+		require_layout_shared.stampFrame(run, pageIndex, textBoxOfItem(item, pageIndex));
+		runs.push(run);
+	});
+}
+function bucketCounts(values, bucketSize) {
+	const counts = /* @__PURE__ */ new Map();
+	for (const v of values) {
+		const bucket = Math.round(v / bucketSize) * bucketSize;
+		counts.set(bucket, (counts.get(bucket) ?? 0) + 1);
+	}
+	return counts;
+}
+function modeOf(values, bucketSize) {
+	if (values.length === 0) return 0;
+	const counts = bucketCounts(values, bucketSize);
+	let bestBucket = values[0];
+	let bestCount = 0;
+	for (const [bucket, count] of counts) if (count > bestCount) {
+		bestBucket = bucket;
+		bestCount = count;
+	}
+	return bestBucket;
+}
+function modalLineGap(gaps) {
+	const counts = bucketCounts(gaps, HALF_POINT_BUCKET_PT);
+	let bestBucket = gaps[0];
+	let bestCount = 0;
+	for (const [bucket, count] of counts) if (count > bestCount) {
+		bestBucket = bucket;
+		bestCount = count;
+	}
+	return bestCount > 1 ? bestBucket : Math.min(...gaps);
+}
+const LARGE_GAP_EM_MULTIPLIER = 2;
+const FONT_SIZE_CLOSE_TOLERANCE_PT = 1;
+const FONT_SIZE_CLOSE_TOLERANCE_RATIO = .15;
+function fontSizesClose(a, b) {
+	return Math.abs(a - b) <= FONT_SIZE_CLOSE_TOLERANCE_PT || Math.abs(a - b) / Math.max(a, b) <= FONT_SIZE_CLOSE_TOLERANCE_RATIO;
+}
+const PARAGRAPH_GAP_MULTIPLIER = 1.25;
+const NOMINAL_LINE_SPACING_RATIO = 1.2;
+const MIN_LINES_FOR_GAP_MODE = 3;
+function estimateModalLineSpacing(lines) {
+	if (lines.length < MIN_LINES_FOR_GAP_MODE) return modeOf(lines.map((l) => l.items[0].sizePt), HALF_POINT_BUCKET_PT) * NOMINAL_LINE_SPACING_RATIO;
+	const gaps = [];
+	for (let i = 1; i < lines.length; i++) gaps.push(lines[i - 1].baselineY - lines[i].baselineY);
+	return modalLineGap(gaps);
+}
+function clusterIntoParagraphs(lines) {
+	if (lines.length === 0) return [];
+	const modalSpacing = estimateModalLineSpacing(lines);
+	const dominantLeftX = modeOf(lines.map((l) => l.items[0].xPt), 1);
+	const paragraphs = [{ lines: [lines[0]] }];
+	for (let i = 1; i < lines.length; i++) {
+		const prev = lines[i - 1];
+		const next = lines[i];
+		if (startsNewParagraph(prev, next, modalSpacing, dominantLeftX)) paragraphs.push({ lines: [next] });
+		else paragraphs[paragraphs.length - 1].lines.push(next);
+	}
+	return paragraphs;
+}
+function startsNewParagraph(prev, next, modalSpacing, dominantLeftX) {
+	if (prev.baselineY - next.baselineY > 1.25 * 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;
+	const nextIndented = Math.abs(nextLeft - dominantLeftX) > emPt;
+	const prevAtMargin = Math.abs(prevLeft - dominantLeftX) <= emPt;
+	return nextIndented && prevAtMargin;
+}
+const LEFT_ALIGN_TOLERANCE_PT = 2;
+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) <= 2) ? "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));
+		if (lineIndex > 0) {
+			const lastRun = result.runs[result.runs.length - 1];
+			if (lastRun !== void 0) lastRun.text += " ";
+		}
+		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;
+}
+//#endregion
+exports.HALF_POINT_BUCKET_PT = HALF_POINT_BUCKET_PT;
+exports.LARGE_GAP_EM_MULTIPLIER = LARGE_GAP_EM_MULTIPLIER;
+exports.LEFT_ALIGN_TOLERANCE_PT = LEFT_ALIGN_TOLERANCE_PT;
+exports.MIN_WORD_GAP_PT = MIN_WORD_GAP_PT;
+exports.PARAGRAPH_GAP_MULTIPLIER = PARAGRAPH_GAP_MULTIPLIER;
+exports.bucketCounts = bucketCounts;
+exports.clusterIntoLines = clusterIntoLines;
+exports.clusterIntoParagraphs = clusterIntoParagraphs;
+exports.estimateModalLineSpacing = estimateModalLineSpacing;
+exports.fontSizesClose = fontSizesClose;
+exports.gapExceeds = gapExceeds;
+exports.lineBox = lineBox;
+exports.modeOf = modeOf;
+exports.paragraphToContentParagraph = paragraphToContentParagraph;
+exports.pushRunsForLine = pushRunsForLine;
+exports.textBoxOfItem = textBoxOfItem;
+exports.textItemToContentRun = textItemToContentRun;
+exports.textItemVerticalExtent = textItemVerticalExtent;

package/dist/layout/reconstruct-lines.d.cts

@@ -0,0 +1,2 @@
+import { _ as pushRunsForLine, a as PARAGRAPH_GAP_MULTIPLIER, b as textItemVerticalExtent, c as bucketCounts, d as estimateModalLineSpacing, f as fontSizesClose, g as paragraphToContentParagraph, h as modeOf, i as MIN_WORD_GAP_PT, l as clusterIntoLines, m as lineBox, n as LARGE_GAP_EM_MULTIPLIER, o as TextLine, p as gapExceeds, r as LEFT_ALIGN_TOLERANCE_PT, s as TextParagraph, t as HALF_POINT_BUCKET_PT, u as clusterIntoParagraphs, v as textBoxOfItem, y as textItemToContentRun } from "../reconstruct-lines-B7t66ICm.cjs";
+export { HALF_POINT_BUCKET_PT, LARGE_GAP_EM_MULTIPLIER, LEFT_ALIGN_TOLERANCE_PT, MIN_WORD_GAP_PT, PARAGRAPH_GAP_MULTIPLIER, TextLine, TextParagraph, bucketCounts, clusterIntoLines, clusterIntoParagraphs, estimateModalLineSpacing, fontSizesClose, gapExceeds, lineBox, modeOf, paragraphToContentParagraph, pushRunsForLine, textBoxOfItem, textItemToContentRun, textItemVerticalExtent };

package/dist/layout/reconstruct-lines.d.ts

@@ -0,0 +1,2 @@
+import { _ as pushRunsForLine, a as PARAGRAPH_GAP_MULTIPLIER, b as textItemVerticalExtent, c as bucketCounts, d as estimateModalLineSpacing, f as fontSizesClose, g as paragraphToContentParagraph, h as modeOf, i as MIN_WORD_GAP_PT, l as clusterIntoLines, m as lineBox, n as LARGE_GAP_EM_MULTIPLIER, o as TextLine, p as gapExceeds, r as LEFT_ALIGN_TOLERANCE_PT, s as TextParagraph, t as HALF_POINT_BUCKET_PT, u as clusterIntoParagraphs, v as textBoxOfItem, y as textItemToContentRun } from "../reconstruct-lines-B7t66ICm.js";
+export { HALF_POINT_BUCKET_PT, LARGE_GAP_EM_MULTIPLIER, LEFT_ALIGN_TOLERANCE_PT, MIN_WORD_GAP_PT, PARAGRAPH_GAP_MULTIPLIER, TextLine, TextParagraph, bucketCounts, clusterIntoLines, clusterIntoParagraphs, estimateModalLineSpacing, fontSizesClose, gapExceeds, lineBox, modeOf, paragraphToContentParagraph, pushRunsForLine, textBoxOfItem, textItemToContentRun, textItemVerticalExtent };

package/dist/layout/reconstruct-lines.js

@@ -0,0 +1,297 @@
+import { stampFrame } from "./shared.js";
+import { resolveStandardFont } from "pdf-codec/fonts";
+import { STANDARD_METRICS } from "pdf-codec/afm-widths";
+import { runGapPt, runsShareBaseline } from "pdf-codec/text-group";
+//#region src/layout/reconstruct-lines.ts
+const LINE_BASELINE_TOLERANCE_FACTOR = .5;
+const AFM_UNITS_PER_EM = 1e3;
+const HALF_POINT_BUCKET_PT = .5;
+function sharesBaseline(anchor, item) {
+	return runsShareBaseline(anchor, item, { baselineToleranceEm: LINE_BASELINE_TOLERANCE_FACTOR });
+}
+const DUPLICATE_PAINT_BUCKET_PT = .1;
+function duplicatePaintKey(item) {
+	const xBucket = Math.round(item.xPt / DUPLICATE_PAINT_BUCKET_PT);
+	const yBucket = Math.round(item.yPt / DUPLICATE_PAINT_BUCKET_PT);
+	return `${item.text.trim()}|${String(xBucket)}|${String(yBucket)}`;
+}
+function dropDuplicatePaints(items) {
+	const seen = /* @__PURE__ */ new Set();
+	const kept = [];
+	for (const item of items) {
+		const key = duplicatePaintKey(item);
+		if (seen.has(key)) continue;
+		seen.add(key);
+		kept.push(item);
+	}
+	return kept;
+}
+function dropOverlappingRepeatsWithinLine(items) {
+	const rightEdgeByText = /* @__PURE__ */ new Map();
+	const kept = [];
+	for (const item of items) {
+		const textKey = item.text.trim();
+		const priorRightEdge = rightEdgeByText.get(textKey);
+		if (priorRightEdge !== void 0 && item.xPt < priorRightEdge) continue;
+		rightEdgeByText.set(textKey, item.xPt + (item.widthPt ?? 0));
+		kept.push(item);
+	}
+	return kept;
+}
+const REDRAW_PASS_REWIND_TOLERANCE_PT = 2;
+const MIN_REDRAW_PASS_TEXT_LENGTH = 12;
+const REDRAW_PASS_SIMILARITY_THRESHOLD = .6;
+function groupByBaselineInDocumentOrder(items) {
+	const groups = [];
+	for (const item of items) {
+		const group = groups.find((g) => sharesBaseline(g.anchor, item));
+		if (group === void 0) groups.push({
+			anchor: item,
+			items: [item]
+		});
+		else group.items.push(item);
+	}
+	return groups.map((g) => g.items);
+}
+function splitIntoPasses(group) {
+	const passes = [];
+	for (const item of group) {
+		const current = passes.at(-1);
+		const itemRightPt = item.xPt + (item.widthPt ?? 0);
+		if (current !== void 0 && item.xPt >= current.endXPt - REDRAW_PASS_REWIND_TOLERANCE_PT) {
+			current.items.push(item);
+			current.endXPt = Math.max(current.endXPt, itemRightPt);
+		} else passes.push({
+			items: [item],
+			startXPt: item.xPt,
+			endXPt: itemRightPt
+		});
+	}
+	return passes;
+}
+function passText(pass) {
+	return pass.items.map((item) => item.text).join("").trim();
+}
+function passesOverlap(a, b) {
+	return a.startXPt < b.endXPt && b.startXPt < a.endXPt;
+}
+function levenshteinDistance(a, b) {
+	const memo = /* @__PURE__ */ new Map();
+	function distance(i, j) {
+		if (i === 0) return j;
+		if (j === 0) return i;
+		const key = `${String(i)}:${String(j)}`;
+		const cached = memo.get(key);
+		if (cached !== void 0) return cached;
+		const cost = a.charAt(i - 1) === b.charAt(j - 1) ? 0 : 1;
+		const result = Math.min(distance(i - 1, j) + 1, distance(i, j - 1) + 1, distance(i - 1, j - 1) + cost);
+		memo.set(key, result);
+		return result;
+	}
+	return distance(a.length, b.length);
+}
+function textSimilarity(a, b) {
+	const maxLength = Math.max(a.length, b.length);
+	return 1 - levenshteinDistance(a, b) / maxLength;
+}
+function dropFuzzyRedrawnPasses(items) {
+	const toDrop = /* @__PURE__ */ new Set();
+	for (const group of groupByBaselineInDocumentOrder(items)) {
+		const passes = splitIntoPasses(group);
+		if (passes.length < 2) continue;
+		const [first, ...rest] = passes;
+		if (first === void 0) continue;
+		const firstText = passText(first);
+		if (firstText.length < MIN_REDRAW_PASS_TEXT_LENGTH) continue;
+		for (const candidate of rest) {
+			const candidateText = passText(candidate);
+			if (candidateText.length < MIN_REDRAW_PASS_TEXT_LENGTH) continue;
+			if (!passesOverlap(first, candidate)) continue;
+			if (textSimilarity(firstText, candidateText) < REDRAW_PASS_SIMILARITY_THRESHOLD) continue;
+			for (const item of candidate.items) toDrop.add(item);
+		}
+	}
+	if (toDrop.size === 0) return items;
+	return items.filter((item) => !toDrop.has(item));
+}
+function clusterIntoLines(items) {
+	const sorted = [...dropDuplicatePaints(dropFuzzyRedrawnPasses(items))].sort((a, b) => b.yPt - a.yPt || a.xPt - b.xPt);
+	const working = [];
+	for (const item of sorted) {
+		const line = working.find((l) => sharesBaseline(l.anchor, item));
+		if (line === void 0) working.push({
+			items: [item],
+			baselineY: item.yPt,
+			anchor: item
+		});
+		else line.items.push(item);
+	}
+	for (const line of working) {
+		line.items.sort((a, b) => a.xPt - b.xPt);
+		line.items = dropOverlappingRepeatsWithinLine(line.items);
+	}
+	working.sort((a, b) => b.baselineY - a.baselineY);
+	return working;
+}
+function textItemVerticalExtent(item) {
+	const { standardName } = resolveStandardFont(item.font.family, item.font.weight === "bold", item.font.style === "italic");
+	const metrics = STANDARD_METRICS[standardName];
+	return {
+		ascentPt: metrics.ascender / AFM_UNITS_PER_EM * item.sizePt,
+		descentPt: Math.abs(metrics.descender) / AFM_UNITS_PER_EM * item.sizePt
+	};
+}
+function textItemToContentRun(item) {
+	return {
+		text: item.text,
+		...item.font.weight === "bold" ? { bold: true } : {},
+		...item.font.style === "italic" ? { italic: true } : {},
+		fontFamily: item.font.family,
+		sizePt: item.sizePt,
+		color: item.color
+	};
+}
+const MIN_WORD_GAP_PT = .5;
+function gapExceeds(previous, next, thresholdPt) {
+	const gap = runGapPt(previous, next);
+	return gap !== void 0 && gap > thresholdPt;
+}
+function textBoxOfItem(item, pageIndex) {
+	const { ascentPt, descentPt } = textItemVerticalExtent(item);
+	return {
+		pageIndex,
+		xPt: item.xPt,
+		yPt: item.yPt - descentPt,
+		widthPt: item.widthPt ?? 0,
+		heightPt: ascentPt + descentPt
+	};
+}
+function lineBox(line, pageIndex) {
+	let minX = Number.POSITIVE_INFINITY;
+	let maxX = Number.NEGATIVE_INFINITY;
+	let minY = Number.POSITIVE_INFINITY;
+	let maxY = Number.NEGATIVE_INFINITY;
+	for (const item of line.items) {
+		const { ascentPt, descentPt } = textItemVerticalExtent(item);
+		minX = Math.min(minX, item.xPt);
+		maxX = Math.max(maxX, item.xPt + (item.widthPt ?? 0));
+		minY = Math.min(minY, item.yPt - descentPt);
+		maxY = Math.max(maxY, item.yPt + ascentPt);
+	}
+	return {
+		pageIndex,
+		xPt: minX,
+		yPt: minY,
+		widthPt: maxX - minX,
+		heightPt: maxY - minY
+	};
+}
+function pushRunsForLine(runs, line, pageIndex) {
+	line.items.forEach((item, itemIndex) => {
+		if (itemIndex > 0) {
+			const prevItem = line.items[itemIndex - 1];
+			if (gapExceeds(prevItem, item, 2 * item.sizePt)) runs.push({ text: "	" });
+			else if (gapExceeds(prevItem, item, .5)) runs[runs.length - 1].text += " ";
+		}
+		const run = textItemToContentRun(item);
+		stampFrame(run, pageIndex, textBoxOfItem(item, pageIndex));
+		runs.push(run);
+	});
+}
+function bucketCounts(values, bucketSize) {
+	const counts = /* @__PURE__ */ new Map();
+	for (const v of values) {
+		const bucket = Math.round(v / bucketSize) * bucketSize;
+		counts.set(bucket, (counts.get(bucket) ?? 0) + 1);
+	}
+	return counts;
+}
+function modeOf(values, bucketSize) {
+	if (values.length === 0) return 0;
+	const counts = bucketCounts(values, bucketSize);
+	let bestBucket = values[0];
+	let bestCount = 0;
+	for (const [bucket, count] of counts) if (count > bestCount) {
+		bestBucket = bucket;
+		bestCount = count;
+	}
+	return bestBucket;
+}
+function modalLineGap(gaps) {
+	const counts = bucketCounts(gaps, HALF_POINT_BUCKET_PT);
+	let bestBucket = gaps[0];
+	let bestCount = 0;
+	for (const [bucket, count] of counts) if (count > bestCount) {
+		bestBucket = bucket;
+		bestCount = count;
+	}
+	return bestCount > 1 ? bestBucket : Math.min(...gaps);
+}
+const LARGE_GAP_EM_MULTIPLIER = 2;
+const FONT_SIZE_CLOSE_TOLERANCE_PT = 1;
+const FONT_SIZE_CLOSE_TOLERANCE_RATIO = .15;
+function fontSizesClose(a, b) {
+	return Math.abs(a - b) <= FONT_SIZE_CLOSE_TOLERANCE_PT || Math.abs(a - b) / Math.max(a, b) <= FONT_SIZE_CLOSE_TOLERANCE_RATIO;
+}
+const PARAGRAPH_GAP_MULTIPLIER = 1.25;
+const NOMINAL_LINE_SPACING_RATIO = 1.2;
+const MIN_LINES_FOR_GAP_MODE = 3;
+function estimateModalLineSpacing(lines) {
+	if (lines.length < MIN_LINES_FOR_GAP_MODE) return modeOf(lines.map((l) => l.items[0].sizePt), HALF_POINT_BUCKET_PT) * NOMINAL_LINE_SPACING_RATIO;
+	const gaps = [];
+	for (let i = 1; i < lines.length; i++) gaps.push(lines[i - 1].baselineY - lines[i].baselineY);
+	return modalLineGap(gaps);
+}
+function clusterIntoParagraphs(lines) {
+	if (lines.length === 0) return [];
+	const modalSpacing = estimateModalLineSpacing(lines);
+	const dominantLeftX = modeOf(lines.map((l) => l.items[0].xPt), 1);
+	const paragraphs = [{ lines: [lines[0]] }];
+	for (let i = 1; i < lines.length; i++) {
+		const prev = lines[i - 1];
+		const next = lines[i];
+		if (startsNewParagraph(prev, next, modalSpacing, dominantLeftX)) paragraphs.push({ lines: [next] });
+		else paragraphs[paragraphs.length - 1].lines.push(next);
+	}
+	return paragraphs;
+}
+function startsNewParagraph(prev, next, modalSpacing, dominantLeftX) {
+	if (prev.baselineY - next.baselineY > 1.25 * 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;
+	const nextIndented = Math.abs(nextLeft - dominantLeftX) > emPt;
+	const prevAtMargin = Math.abs(prevLeft - dominantLeftX) <= emPt;
+	return nextIndented && prevAtMargin;
+}
+const LEFT_ALIGN_TOLERANCE_PT = 2;
+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) <= 2) ? "left" : void 0,
+		...headingLevel !== void 0 ? {
+			styleId: `Heading${String(headingLevel)}`,
+			headingLevel
+		} : {}
+	};
+	paragraph.lines.forEach((line, lineIndex) => {
+		stampFrame(result, pageIndex, lineBox(line, pageIndex));
+		if (lineIndex > 0) {
+			const lastRun = result.runs[result.runs.length - 1];
+			if (lastRun !== void 0) lastRun.text += " ";
+		}
+		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;
+}
+//#endregion
+export { HALF_POINT_BUCKET_PT, LARGE_GAP_EM_MULTIPLIER, LEFT_ALIGN_TOLERANCE_PT, MIN_WORD_GAP_PT, PARAGRAPH_GAP_MULTIPLIER, bucketCounts, clusterIntoLines, clusterIntoParagraphs, estimateModalLineSpacing, fontSizesClose, gapExceeds, lineBox, modeOf, paragraphToContentParagraph, pushRunsForLine, textBoxOfItem, textItemToContentRun, textItemVerticalExtent };

package/dist/layout/reconstruct-presentations-prod.cjs

@@ -0,0 +1,9 @@
+Object.defineProperty(exports, Symbol.toStringTag, { value: "Module" });
+const require_reconstruct_presentations_prod = require("../reconstruct-presentations-prod-X315aBmI.cjs");
+exports.layoutItemToVector = require_reconstruct_presentations_prod.layoutItemToVector;
+exports.lineToParagraph = require_reconstruct_presentations_prod.lineToParagraph;
+exports.reconstructDrawing = require_reconstruct_presentations_prod.reconstructDrawing;
+exports.reconstructPresentation = require_reconstruct_presentations_prod.reconstructPresentation;
+exports.reconstructSheet = require_reconstruct_presentations_prod.reconstructSheet;
+exports.recoverPageVectors = require_reconstruct_presentations_prod.recoverPageVectors;
+exports.splitLineByLargeGaps = require_reconstruct_presentations_prod.splitLineByLargeGaps;