docrev 0.9.18 → 0.10.1

package/lib/protect-restore.ts

@@ -1,520 +1,520 @@
-/**
- * Protection and restoration utilities for markdown elements during Word import
- *
- * These functions protect special markdown syntax (anchors, cross-refs, math, citations,
- * images, tables) by replacing them with placeholders before diffing, then restore them after.
- */
-
-// =============================================================================
-// Interfaces
-// =============================================================================
-
-interface MarkdownPrefix {
-  prefix: string;
-  content: string;
-}
-
-interface ProtectedItem {
-  original: string;
-  placeholder: string;
-}
-
-interface ProtectedMath extends ProtectedItem {
-  type: 'inline' | 'display';
-  simplified: string;
-}
-
-interface ProtectedImage extends ProtectedItem {
-  label: string | null;
-  caption: string;
-  path: string;
-  figureNumber: string | null;
-}
-
-interface ProtectedTable extends ProtectedItem {
-  cellCount: number;
-}
-
-interface ProtectAnchorsResult {
-  text: string;
-  anchors: ProtectedItem[];
-}
-
-interface ProtectCrossrefsResult {
-  text: string;
-  crossrefs: ProtectedItem[];
-}
-
-interface ProtectMathResult {
-  text: string;
-  mathBlocks: ProtectedMath[];
-}
-
-interface ProtectCitationsResult {
-  text: string;
-  citations: string[];
-}
-
-interface ProtectImagesResult {
-  text: string;
-  images: ProtectedImage[];
-}
-
-interface ProtectTablesResult {
-  text: string;
-  tables: ProtectedTable[];
-}
-
-interface ImageRegistry {
-  byNumber?: Map<string, { label: string }>;
-}
-
-// =============================================================================
-// Shared Helpers
-// =============================================================================
-
-/**
- * Replace regex matches with indexed placeholders and collect originals
- */
-function collectAndReplace(
-  text: string,
-  pattern: RegExp,
-  prefix: string,
-  suffix: string,
-): { text: string; items: ProtectedItem[] } {
-  const items: ProtectedItem[] = [];
-  const result = text.replace(pattern, (match) => {
-    const idx = items.length;
-    const placeholder = `${prefix}${idx}${suffix}`;
-    items.push({ original: match, placeholder });
-    return placeholder;
-  });
-  return { text: result, items };
-}
-
-/**
- * Restore protected items from placeholders, handling annotation wrappers
- * (deletion {--...--} and insertion {++...++} wrappers are unwrapped)
- */
-function restoreProtectedItems(text: string, items: ProtectedItem[]): string {
-  for (const item of items) {
-    const deletionPattern = new RegExp(`\\{--[^}]*?${item.placeholder}[^}]*?--\\}`, 'g');
-    text = text.replace(deletionPattern, item.original);
-
-    const insertionPattern = new RegExp(`\\{\\+\\+[^}]*?${item.placeholder}[^}]*?\\+\\+\\}`, 'g');
-    text = text.replace(insertionPattern, item.original);
-
-    text = text.split(item.placeholder).join(item.original);
-  }
-  return text;
-}
-
-// =============================================================================
-// Public Functions
-// =============================================================================
-
-/**
- * Extract markdown prefix (headers, list markers) from a line
- */
-export function extractMarkdownPrefix(line: string): MarkdownPrefix {
-  // Headers
-  const headerMatch = line.match(/^(#{1,6}\s+)/);
-  if (headerMatch && headerMatch[1]) {
-    return { prefix: headerMatch[1], content: line.slice(headerMatch[1].length) };
-  }
-
-  // List items
-  const listMatch = line.match(/^(\s*[-*+]\s+|\s*\d+\.\s+)/);
-  if (listMatch && listMatch[1]) {
-    return { prefix: listMatch[1], content: line.slice(listMatch[1].length) };
-  }
-
-  // Blockquotes
-  const quoteMatch = line.match(/^(>\s*)/);
-  if (quoteMatch && quoteMatch[1]) {
-    return { prefix: quoteMatch[1], content: line.slice(quoteMatch[1].length) };
-  }
-
-  return { prefix: '', content: line };
-}
-
-/**
- * Protect figure/table anchors before diffing
- * Anchors like {#fig:heatmap} and {#tbl:results} should never be deleted
- */
-export function protectAnchors(md: string): ProtectAnchorsResult {
-  // Match {#fig:label}, {#tbl:label}, {#eq:label}, {#sec:label} etc.
-  // Also match with additional attributes like {#fig:label width=50%}
-  const { text, items: anchors } = collectAndReplace(
-    md, /\{#(fig|tbl|eq|sec|lst):[^}]+\}/g, 'ANCHORBLOCK', 'ENDANCHOR',
-  );
-  return { text, anchors };
-}
-
-/**
- * Restore anchors from placeholders
- */
-export function restoreAnchors(text: string, anchors: ProtectedItem[]): string {
-  for (const anchor of anchors) {
-    // Handle case where anchor is inside a deletion annotation
-    // {--...ANCHORBLOCK0ENDANCHOR--} should become {--...--}{#fig:label}
-    const deletionPattern = new RegExp(`\\{--([^}]*?)${anchor.placeholder}([^}]*?)--\\}`, 'g');
-    text = text.replace(deletionPattern, (match, before, after) => {
-      const cleanBefore = before.trim();
-      const cleanAfter = after.trim();
-      let result = '';
-      if (cleanBefore) result += `{--${cleanBefore}--}`;
-      result += anchor.original;
-      if (cleanAfter) result += `{--${cleanAfter}--}`;
-      return result;
-    });
-
-    // Handle case where anchor is inside a substitution
-    // {~~old ANCHORBLOCK0ENDANCHOR~>new~~} -> {~~old~>new~~}{#fig:label}
-    const substitutionPattern = new RegExp(`\\{~~([^~]*?)${anchor.placeholder}([^~]*?)~>([^~]*)~~\\}`, 'g');
-    text = text.replace(substitutionPattern, (match: string, oldBefore: string, oldAfter: string, newText: string) => {
-      const cleanOldBefore = (oldBefore ?? '').trim();
-      const cleanOldAfter = (oldAfter ?? '').trim();
-      const cleanNew = (newText ?? '').trim();
-      const oldText = (cleanOldBefore + ' ' + cleanOldAfter).trim();
-      let result = '';
-      if (oldText !== cleanNew) {
-        result += `{~~${oldText}~>${cleanNew}~~}`;
-      } else {
-        result += cleanNew;
-      }
-      result += anchor.original;
-      return result;
-    });
-
-    // Normal replacement
-    text = text.split(anchor.placeholder).join(anchor.original);
-  }
-  return text;
-}
-
-/**
- * Protect cross-references before diffing
- * References like @fig:label, @tbl:label should be preserved
- */
-export function protectCrossrefs(md: string): ProtectCrossrefsResult {
-  // Match @fig:label, @tbl:label, @eq:label, @sec:label
-  // Can appear as @fig:label or (@fig:label) or [@fig:label]
-  const { text, items: crossrefs } = collectAndReplace(
-    md, /@(fig|tbl|eq|sec|lst):[a-zA-Z0-9_-]+/g, 'XREFBLOCK', 'ENDXREF',
-  );
-  return { text, crossrefs };
-}
-
-/**
- * Restore cross-references from placeholders
- */
-export function restoreCrossrefs(text: string, crossrefs: ProtectedItem[]): string {
-  for (const xref of crossrefs) {
-    // Handle deletions - restore the reference even if marked deleted
-    const deletionPattern = new RegExp(`\\{--([^}]*?)${xref.placeholder}([^}]*?)--\\}`, 'g');
-    text = text.replace(deletionPattern, (match, before, after) => {
-      const cleanBefore = before.trim();
-      const cleanAfter = after.trim();
-      let result = '';
-      if (cleanBefore) result += `{--${cleanBefore}--}`;
-      result += xref.original;
-      if (cleanAfter) result += `{--${cleanAfter}--}`;
-      return result;
-    });
-
-    // Handle substitutions where rendered form (Figure 1) replaced the reference
-    // {~~XREFBLOCK0ENDXREF~>Figure 1~~} -> @fig:label
-    const substitutionPattern = new RegExp(`\\{~~${xref.placeholder}~>[^~]+~~\\}`, 'g');
-    text = text.replace(substitutionPattern, xref.original);
-
-    // Normal replacement
-    text = text.split(xref.placeholder).join(xref.original);
-  }
-  return text;
-}
-
-/**
- * Simplify LaTeX math for fuzzy matching against Word text
- * Word renders math as text, so we need to match the rendered form
- */
-export function simplifyMathForMatching(latex: string): string {
-  return latex
-    // Remove common LaTeX commands
-    .replace(/\\text\{([^}]+)\}/g, '$1')
-    .replace(/\\hat\{([^}]+)\}/g, '$1')
-    .replace(/\\bar\{([^}]+)\}/g, '$1')
-    .replace(/\\frac\{([^}]+)\}\{([^}]+)\}/g, '$1/$2')
-    .replace(/\\sum_([a-z])/g, 'Σ')
-    .replace(/\\sum/g, 'Σ')
-    .replace(/\\cdot/g, '·')
-    .replace(/\\quad/g, ' ')
-    .replace(/\\,/g, ' ')
-    .replace(/\\_/g, '_')
-    .replace(/\\{/g, '{')
-    .replace(/\\}/g, '}')
-    .replace(/\\/g, '')  // Remove remaining backslashes
-    .replace(/[{}]/g, '')  // Remove braces
-    .replace(/\s+/g, ' ')
-    .trim();
-}
-
-/**
- * Protect mathematical notation before diffing by replacing with placeholders
- * Handles both inline $...$ and display $$...$$ math
- */
-export function protectMath(md: string): ProtectMathResult {
-  const mathBlocks: ProtectedMath[] = [];
-
-  // First protect display math ($$...$$) - must be done before inline math
-  let text = md.replace(/\$\$([^$]+)\$\$/g, (match, content) => {
-    const idx = mathBlocks.length;
-    const placeholder = `MATHBLOCK${idx}ENDMATH`;
-    // Create simplified version for matching in Word text
-    const simplified = simplifyMathForMatching(content);
-    mathBlocks.push({ original: match, placeholder, type: 'display', simplified });
-    return placeholder;
-  });
-
-  // Then protect inline math ($...$)
-  text = text.replace(/\$([^$\n]+)\$/g, (match, content) => {
-    const idx = mathBlocks.length;
-    const placeholder = `MATHBLOCK${idx}ENDMATH`;
-    const simplified = simplifyMathForMatching(content);
-    mathBlocks.push({ original: match, placeholder, type: 'inline', simplified });
-    return placeholder;
-  });
-
-  return { text, mathBlocks };
-}
-
-/**
- * Restore math from placeholders
- */
-export function restoreMath(text: string, mathBlocks: ProtectedMath[]): string {
-  for (const block of mathBlocks) {
-    text = text.split(block.placeholder).join(block.original);
-  }
-  return text;
-}
-
-/**
- * Replace rendered math in Word text with matching placeholders
- * This is heuristic-based since Word can render math in various ways
- */
-export function replaceRenderedMath(wordText: string, mathBlocks: ProtectedMath[]): string {
-  let result = wordText;
-
-  for (const block of mathBlocks) {
-    // For inline math, try to find the simplified form in Word text
-    if (block.simplified.length >= 2) {
-      // Try exact match first
-      if (result.includes(block.simplified)) {
-        result = result.replace(block.simplified, block.placeholder);
-      }
-    }
-  }
-
-  return result;
-}
-
-/**
- * Protect citations before diffing by replacing with placeholders
- */
-export function protectCitations(md: string): ProtectCitationsResult {
-  const citations: string[] = [];
-  const text = md.replace(/\[@[^\]]+\]/g, (match) => {
-    const idx = citations.length;
-    citations.push(match);
-    return `CITEREF${idx}ENDCITE`;
-  });
-  return { text, citations };
-}
-
-/**
- * Restore citations from placeholders
- */
-export function restoreCitations(text: string, citations: string[]): string {
-  for (let i = 0; i < citations.length; i++) {
-    // Handle cases where placeholder might be inside annotations
-    const placeholder = `CITEREF${i}ENDCITE`;
-    text = text.split(placeholder).join(citations[i]);
-  }
-  return text;
-}
-
-/**
- * Remove rendered citations from Word text (replace with matching placeholders)
- */
-export function replaceRenderedCitations(wordText: string, count: number): string {
-  // Match rendered citation patterns: (Author 2021), (Author et al. 2021), etc.
-  const pattern = /\((?:[A-Z][a-zé]+(?:\s+et\s+al\.?)?(?:\s*[&,;]\s*[A-Z][a-zé]+(?:\s+et\s+al\.?)?)*\s+\d{4}(?:[a-z])?(?:\s*[,;]\s*(?:[A-Z][a-zé]+(?:\s+et\s+al\.?)?\s+)?\d{4}(?:[a-z])?)*)\)/g;
-
-  let idx = 0;
-  return wordText.replace(pattern, (match) => {
-    if (idx < count) {
-      const placeholder = `CITEREF${idx}ENDCITE`;
-      idx++;
-      return placeholder;
-    }
-    return match;
-  });
-}
-
-/**
- * Protect markdown images before diffing by replacing with placeholders
- * Images are treated as atomic blocks to prevent corruption during diff
- *
- * Matches: ![caption](path){#fig:label} or ![caption](path)
- * Also matches Word-style: ![Figure N: caption](media/path)
- */
-export function protectImages(md: string, registry: ImageRegistry | null = null): ProtectImagesResult {
-  const images: ProtectedImage[] = [];
-
-  // Match markdown images: ![caption](path){#anchor} or ![caption](path)
-  // The anchor is optional and can have additional attributes
-  const imagePattern = /!\[([^\]]*)\]\(([^)]+)\)(?:\{([^}]+)\})?/g;
-
-  const text = md.replace(imagePattern, (match, caption, path, anchor) => {
-    const idx = images.length;
-    const placeholder = `IMAGEBLOCK${idx}ENDIMAGE`;
-
-    // Extract label from anchor if present (e.g., "#fig:map" -> "map")
-    let label: string | null = null;
-    if (anchor) {
-      const labelMatch = anchor.match(/#(fig|tbl):([a-zA-Z0-9_-]+)/);
-      if (labelMatch) {
-        label = labelMatch[2];
-      }
-    }
-
-    // Try to extract figure number from Word-style caption "Figure N: ..."
-    let figureNumber: string | null = null;
-    const figNumMatch = caption.match(/^(?:Figure|Fig\.?|Table|Tbl\.?)\s+(\d+|S\d+)[:\.]?\s*/i);
-    if (figNumMatch) {
-      figureNumber = figNumMatch[1];
-    }
-
-    images.push({
-      original: match,
-      placeholder,
-      label,
-      caption: caption.trim(),
-      path,
-      figureNumber,
-    });
-
-    return placeholder;
-  });
-
-  return { text, images };
-}
-
-/**
- * Restore images from placeholders
- */
-export function restoreImages(text: string, images: ProtectedImage[]): string {
-  return restoreProtectedItems(text, images);
-}
-
-/**
- * Match Word-extracted images to original images using registry
- * Returns a mapping of Word image placeholders to original image placeholders
- */
-export function matchWordImagesToOriginal(
-  originalImages: ProtectedImage[],
-  wordImages: ProtectedImage[],
-  registry: ImageRegistry | null = null
-): Map<string, string> {
-  const mapping = new Map<string, string>();
-  const usedOriginals = new Set<string>();
-
-  for (const wordImg of wordImages) {
-    let bestMatch: ProtectedImage | null = null;
-    let bestScore = 0;
-
-    for (const origImg of originalImages) {
-      if (usedOriginals.has(origImg.placeholder)) continue;
-
-      let score = 0;
-
-      // Match by label (most reliable)
-      if (wordImg.label && origImg.label && wordImg.label === origImg.label) {
-        score += 100;
-      }
-
-      // Match by figure number via registry
-      if (wordImg.figureNumber && registry) {
-        const entry = registry.byNumber?.get(`fig:${wordImg.figureNumber}`);
-        if (entry && entry.label === origImg.label) {
-          score += 90;
-        }
-      }
-
-      // Match by caption similarity (first 50 chars, normalized)
-      const wordCaption = wordImg.caption.replace(/^(?:Figure|Fig\.?|Table|Tbl\.?)\s+\d+[:\.]?\s*/i, '').toLowerCase().slice(0, 50);
-      const origCaption = origImg.caption.toLowerCase().slice(0, 50);
-      if (wordCaption && origCaption && wordCaption === origCaption) {
-        score += 80;
-      } else if (wordCaption && origCaption && (wordCaption.includes(origCaption.slice(0, 30)) || origCaption.includes(wordCaption.slice(0, 30)))) {
-        score += 40;
-      }
-
-      // Match by path similarity (filename)
-      const wordFile = wordImg.path.split('/').pop()?.toLowerCase() || '';
-      const origFile = origImg.path.split('/').pop()?.toLowerCase() || '';
-      if (wordFile === origFile) {
-        score += 30;
-      }
-
-      if (score > bestScore) {
-        bestScore = score;
-        bestMatch = origImg;
-      }
-    }
-
-    if (bestMatch && bestScore >= 40) {
-      mapping.set(wordImg.placeholder, bestMatch.placeholder);
-      usedOriginals.add(bestMatch.placeholder);
-    }
-  }
-
-  return mapping;
-}
-
-/**
- * Protect markdown tables before diffing by replacing with placeholders
- * Tables are treated as atomic blocks to prevent corruption during diff
- */
-export function protectTables(md: string): ProtectTablesResult {
-  const tables: ProtectedTable[] = [];
-
-  // Match markdown tables: lines starting with | and containing |
-  // A table is: optional caption, header row, separator row (|---|), data rows
-  const tablePattern = /(?:^(?:\*\*)?Table[^\n]*\n\n?)?(?:^\|[^\n]+\|\n)+/gm;
-
-  const text = md.replace(tablePattern, (match) => {
-    // Verify it's actually a table (has separator row with dashes)
-    if (!match.includes('|---') && !match.includes('| ---') && !match.includes('|:--')) {
-      return match; // Not a real table, just lines with pipes
-    }
-
-    const idx = tables.length;
-    const placeholder = `\n\nTABLEBLOCK${idx}ENDTABLE\n\n`;
-
-    // Count cells for matching in Word (approximate)
-    const cellCount = (match.match(/\|/g) || []).length;
-
-    tables.push({ original: match.trim(), placeholder: placeholder.trim(), cellCount });
-    return placeholder;
-  });
-
-  return { text, tables };
-}
-
-/**
- * Restore tables from placeholders
- */
-export function restoreTables(text: string, tables: ProtectedTable[]): string {
-  return restoreProtectedItems(text, tables);
-}
+/**
+ * Protection and restoration utilities for markdown elements during Word import
+ *
+ * These functions protect special markdown syntax (anchors, cross-refs, math, citations,
+ * images, tables) by replacing them with placeholders before diffing, then restore them after.
+ */
+
+// =============================================================================
+// Interfaces
+// =============================================================================
+
+interface MarkdownPrefix {
+  prefix: string;
+  content: string;
+}
+
+interface ProtectedItem {
+  original: string;
+  placeholder: string;
+}
+
+interface ProtectedMath extends ProtectedItem {
+  type: 'inline' | 'display';
+  simplified: string;
+}
+
+interface ProtectedImage extends ProtectedItem {
+  label: string | null;
+  caption: string;
+  path: string;
+  figureNumber: string | null;
+}
+
+interface ProtectedTable extends ProtectedItem {
+  cellCount: number;
+}
+
+interface ProtectAnchorsResult {
+  text: string;
+  anchors: ProtectedItem[];
+}
+
+interface ProtectCrossrefsResult {
+  text: string;
+  crossrefs: ProtectedItem[];
+}
+
+interface ProtectMathResult {
+  text: string;
+  mathBlocks: ProtectedMath[];
+}
+
+interface ProtectCitationsResult {
+  text: string;
+  citations: string[];
+}
+
+interface ProtectImagesResult {
+  text: string;
+  images: ProtectedImage[];
+}
+
+interface ProtectTablesResult {
+  text: string;
+  tables: ProtectedTable[];
+}
+
+interface ImageRegistry {
+  byNumber?: Map<string, { label: string }>;
+}
+
+// =============================================================================
+// Shared Helpers
+// =============================================================================
+
+/**
+ * Replace regex matches with indexed placeholders and collect originals
+ */
+function collectAndReplace(
+  text: string,
+  pattern: RegExp,
+  prefix: string,
+  suffix: string,
+): { text: string; items: ProtectedItem[] } {
+  const items: ProtectedItem[] = [];
+  const result = text.replace(pattern, (match) => {
+    const idx = items.length;
+    const placeholder = `${prefix}${idx}${suffix}`;
+    items.push({ original: match, placeholder });
+    return placeholder;
+  });
+  return { text: result, items };
+}
+
+/**
+ * Restore protected items from placeholders, handling annotation wrappers
+ * (deletion {--...--} and insertion {++...++} wrappers are unwrapped)
+ */
+function restoreProtectedItems(text: string, items: ProtectedItem[]): string {
+  for (const item of items) {
+    const deletionPattern = new RegExp(`\\{--[^}]*?${item.placeholder}[^}]*?--\\}`, 'g');
+    text = text.replace(deletionPattern, item.original);
+
+    const insertionPattern = new RegExp(`\\{\\+\\+[^}]*?${item.placeholder}[^}]*?\\+\\+\\}`, 'g');
+    text = text.replace(insertionPattern, item.original);
+
+    text = text.split(item.placeholder).join(item.original);
+  }
+  return text;
+}
+
+// =============================================================================
+// Public Functions
+// =============================================================================
+
+/**
+ * Extract markdown prefix (headers, list markers) from a line
+ */
+export function extractMarkdownPrefix(line: string): MarkdownPrefix {
+  // Headers
+  const headerMatch = line.match(/^(#{1,6}\s+)/);
+  if (headerMatch && headerMatch[1]) {
+    return { prefix: headerMatch[1], content: line.slice(headerMatch[1].length) };
+  }
+
+  // List items
+  const listMatch = line.match(/^(\s*[-*+]\s+|\s*\d+\.\s+)/);
+  if (listMatch && listMatch[1]) {
+    return { prefix: listMatch[1], content: line.slice(listMatch[1].length) };
+  }
+
+  // Blockquotes
+  const quoteMatch = line.match(/^(>\s*)/);
+  if (quoteMatch && quoteMatch[1]) {
+    return { prefix: quoteMatch[1], content: line.slice(quoteMatch[1].length) };
+  }
+
+  return { prefix: '', content: line };
+}
+
+/**
+ * Protect figure/table anchors before diffing
+ * Anchors like {#fig:heatmap} and {#tbl:results} should never be deleted
+ */
+export function protectAnchors(md: string): ProtectAnchorsResult {
+  // Match {#fig:label}, {#tbl:label}, {#eq:label}, {#sec:label} etc.
+  // Also match with additional attributes like {#fig:label width=50%}
+  const { text, items: anchors } = collectAndReplace(
+    md, /\{#(fig|tbl|eq|sec|lst):[^}]+\}/g, 'ANCHORBLOCK', 'ENDANCHOR',
+  );
+  return { text, anchors };
+}
+
+/**
+ * Restore anchors from placeholders
+ */
+export function restoreAnchors(text: string, anchors: ProtectedItem[]): string {
+  for (const anchor of anchors) {
+    // Handle case where anchor is inside a deletion annotation
+    // {--...ANCHORBLOCK0ENDANCHOR--} should become {--...--}{#fig:label}
+    const deletionPattern = new RegExp(`\\{--([^}]*?)${anchor.placeholder}([^}]*?)--\\}`, 'g');
+    text = text.replace(deletionPattern, (match, before, after) => {
+      const cleanBefore = before.trim();
+      const cleanAfter = after.trim();
+      let result = '';
+      if (cleanBefore) result += `{--${cleanBefore}--}`;
+      result += anchor.original;
+      if (cleanAfter) result += `{--${cleanAfter}--}`;
+      return result;
+    });
+
+    // Handle case where anchor is inside a substitution
+    // {~~old ANCHORBLOCK0ENDANCHOR~>new~~} -> {~~old~>new~~}{#fig:label}
+    const substitutionPattern = new RegExp(`\\{~~([^~]*?)${anchor.placeholder}([^~]*?)~>([^~]*)~~\\}`, 'g');
+    text = text.replace(substitutionPattern, (match: string, oldBefore: string, oldAfter: string, newText: string) => {
+      const cleanOldBefore = (oldBefore ?? '').trim();
+      const cleanOldAfter = (oldAfter ?? '').trim();
+      const cleanNew = (newText ?? '').trim();
+      const oldText = (cleanOldBefore + ' ' + cleanOldAfter).trim();
+      let result = '';
+      if (oldText !== cleanNew) {
+        result += `{~~${oldText}~>${cleanNew}~~}`;
+      } else {
+        result += cleanNew;
+      }
+      result += anchor.original;
+      return result;
+    });
+
+    // Normal replacement
+    text = text.split(anchor.placeholder).join(anchor.original);
+  }
+  return text;
+}
+
+/**
+ * Protect cross-references before diffing
+ * References like @fig:label, @tbl:label should be preserved
+ */
+export function protectCrossrefs(md: string): ProtectCrossrefsResult {
+  // Match @fig:label, @tbl:label, @eq:label, @sec:label
+  // Can appear as @fig:label or (@fig:label) or [@fig:label]
+  const { text, items: crossrefs } = collectAndReplace(
+    md, /@(fig|tbl|eq|sec|lst):[a-zA-Z0-9_-]+/g, 'XREFBLOCK', 'ENDXREF',
+  );
+  return { text, crossrefs };
+}
+
+/**
+ * Restore cross-references from placeholders
+ */
+export function restoreCrossrefs(text: string, crossrefs: ProtectedItem[]): string {
+  for (const xref of crossrefs) {
+    // Handle deletions - restore the reference even if marked deleted
+    const deletionPattern = new RegExp(`\\{--([^}]*?)${xref.placeholder}([^}]*?)--\\}`, 'g');
+    text = text.replace(deletionPattern, (match, before, after) => {
+      const cleanBefore = before.trim();
+      const cleanAfter = after.trim();
+      let result = '';
+      if (cleanBefore) result += `{--${cleanBefore}--}`;
+      result += xref.original;
+      if (cleanAfter) result += `{--${cleanAfter}--}`;
+      return result;
+    });
+
+    // Handle substitutions where rendered form (Figure 1) replaced the reference
+    // {~~XREFBLOCK0ENDXREF~>Figure 1~~} -> @fig:label
+    const substitutionPattern = new RegExp(`\\{~~${xref.placeholder}~>[^~]+~~\\}`, 'g');
+    text = text.replace(substitutionPattern, xref.original);
+
+    // Normal replacement
+    text = text.split(xref.placeholder).join(xref.original);
+  }
+  return text;
+}
+
+/**
+ * Simplify LaTeX math for fuzzy matching against Word text
+ * Word renders math as text, so we need to match the rendered form
+ */
+export function simplifyMathForMatching(latex: string): string {
+  return latex
+    // Remove common LaTeX commands
+    .replace(/\\text\{([^}]+)\}/g, '$1')
+    .replace(/\\hat\{([^}]+)\}/g, '$1')
+    .replace(/\\bar\{([^}]+)\}/g, '$1')
+    .replace(/\\frac\{([^}]+)\}\{([^}]+)\}/g, '$1/$2')
+    .replace(/\\sum_([a-z])/g, 'Σ')
+    .replace(/\\sum/g, 'Σ')
+    .replace(/\\cdot/g, '·')
+    .replace(/\\quad/g, ' ')
+    .replace(/\\,/g, ' ')
+    .replace(/\\_/g, '_')
+    .replace(/\\{/g, '{')
+    .replace(/\\}/g, '}')
+    .replace(/\\/g, '')  // Remove remaining backslashes
+    .replace(/[{}]/g, '')  // Remove braces
+    .replace(/\s+/g, ' ')
+    .trim();
+}
+
+/**
+ * Protect mathematical notation before diffing by replacing with placeholders
+ * Handles both inline $...$ and display $$...$$ math
+ */
+export function protectMath(md: string): ProtectMathResult {
+  const mathBlocks: ProtectedMath[] = [];
+
+  // First protect display math ($$...$$) - must be done before inline math
+  let text = md.replace(/\$\$([^$]+)\$\$/g, (match, content) => {
+    const idx = mathBlocks.length;
+    const placeholder = `MATHBLOCK${idx}ENDMATH`;
+    // Create simplified version for matching in Word text
+    const simplified = simplifyMathForMatching(content);
+    mathBlocks.push({ original: match, placeholder, type: 'display', simplified });
+    return placeholder;
+  });
+
+  // Then protect inline math ($...$)
+  text = text.replace(/\$([^$\n]+)\$/g, (match, content) => {
+    const idx = mathBlocks.length;
+    const placeholder = `MATHBLOCK${idx}ENDMATH`;
+    const simplified = simplifyMathForMatching(content);
+    mathBlocks.push({ original: match, placeholder, type: 'inline', simplified });
+    return placeholder;
+  });
+
+  return { text, mathBlocks };
+}
+
+/**
+ * Restore math from placeholders
+ */
+export function restoreMath(text: string, mathBlocks: ProtectedMath[]): string {
+  for (const block of mathBlocks) {
+    text = text.split(block.placeholder).join(block.original);
+  }
+  return text;
+}
+
+/**
+ * Replace rendered math in Word text with matching placeholders
+ * This is heuristic-based since Word can render math in various ways
+ */
+export function replaceRenderedMath(wordText: string, mathBlocks: ProtectedMath[]): string {
+  let result = wordText;
+
+  for (const block of mathBlocks) {
+    // For inline math, try to find the simplified form in Word text
+    if (block.simplified.length >= 2) {
+      // Try exact match first
+      if (result.includes(block.simplified)) {
+        result = result.replace(block.simplified, block.placeholder);
+      }
+    }
+  }
+
+  return result;
+}
+
+/**
+ * Protect citations before diffing by replacing with placeholders
+ */
+export function protectCitations(md: string): ProtectCitationsResult {
+  const citations: string[] = [];
+  const text = md.replace(/\[@[^\]]+\]/g, (match) => {
+    const idx = citations.length;
+    citations.push(match);
+    return `CITEREF${idx}ENDCITE`;
+  });
+  return { text, citations };
+}
+
+/**
+ * Restore citations from placeholders
+ */
+export function restoreCitations(text: string, citations: string[]): string {
+  for (let i = 0; i < citations.length; i++) {
+    // Handle cases where placeholder might be inside annotations
+    const placeholder = `CITEREF${i}ENDCITE`;
+    text = text.split(placeholder).join(citations[i]);
+  }
+  return text;
+}
+
+/**
+ * Remove rendered citations from Word text (replace with matching placeholders)
+ */
+export function replaceRenderedCitations(wordText: string, count: number): string {
+  // Match rendered citation patterns: (Author 2021), (Author et al. 2021), etc.
+  const pattern = /\((?:[A-Z][a-zé]+(?:\s+et\s+al\.?)?(?:\s*[&,;]\s*[A-Z][a-zé]+(?:\s+et\s+al\.?)?)*\s+\d{4}(?:[a-z])?(?:\s*[,;]\s*(?:[A-Z][a-zé]+(?:\s+et\s+al\.?)?\s+)?\d{4}(?:[a-z])?)*)\)/g;
+
+  let idx = 0;
+  return wordText.replace(pattern, (match) => {
+    if (idx < count) {
+      const placeholder = `CITEREF${idx}ENDCITE`;
+      idx++;
+      return placeholder;
+    }
+    return match;
+  });
+}
+
+/**
+ * Protect markdown images before diffing by replacing with placeholders
+ * Images are treated as atomic blocks to prevent corruption during diff
+ *
+ * Matches: ![caption](path){#fig:label} or ![caption](path)
+ * Also matches Word-style: ![Figure N: caption](media/path)
+ */
+export function protectImages(md: string, registry: ImageRegistry | null = null): ProtectImagesResult {
+  const images: ProtectedImage[] = [];
+
+  // Match markdown images: ![caption](path){#anchor} or ![caption](path)
+  // The anchor is optional and can have additional attributes
+  const imagePattern = /!\[([^\]]*)\]\(([^)]+)\)(?:\{([^}]+)\})?/g;
+
+  const text = md.replace(imagePattern, (match, caption, path, anchor) => {
+    const idx = images.length;
+    const placeholder = `IMAGEBLOCK${idx}ENDIMAGE`;
+
+    // Extract label from anchor if present (e.g., "#fig:map" -> "map")
+    let label: string | null = null;
+    if (anchor) {
+      const labelMatch = anchor.match(/#(fig|tbl):([a-zA-Z0-9_-]+)/);
+      if (labelMatch) {
+        label = labelMatch[2];
+      }
+    }
+
+    // Try to extract figure number from Word-style caption "Figure N: ..."
+    let figureNumber: string | null = null;
+    const figNumMatch = caption.match(/^(?:Figure|Fig\.?|Table|Tbl\.?)\s+(\d+|S\d+)[:\.]?\s*/i);
+    if (figNumMatch) {
+      figureNumber = figNumMatch[1];
+    }
+
+    images.push({
+      original: match,
+      placeholder,
+      label,
+      caption: caption.trim(),
+      path,
+      figureNumber,
+    });
+
+    return placeholder;
+  });
+
+  return { text, images };
+}
+
+/**
+ * Restore images from placeholders
+ */
+export function restoreImages(text: string, images: ProtectedImage[]): string {
+  return restoreProtectedItems(text, images);
+}
+
+/**
+ * Match Word-extracted images to original images using registry
+ * Returns a mapping of Word image placeholders to original image placeholders
+ */
+export function matchWordImagesToOriginal(
+  originalImages: ProtectedImage[],
+  wordImages: ProtectedImage[],
+  registry: ImageRegistry | null = null
+): Map<string, string> {
+  const mapping = new Map<string, string>();
+  const usedOriginals = new Set<string>();
+
+  for (const wordImg of wordImages) {
+    let bestMatch: ProtectedImage | null = null;
+    let bestScore = 0;
+
+    for (const origImg of originalImages) {
+      if (usedOriginals.has(origImg.placeholder)) continue;
+
+      let score = 0;
+
+      // Match by label (most reliable)
+      if (wordImg.label && origImg.label && wordImg.label === origImg.label) {
+        score += 100;
+      }
+
+      // Match by figure number via registry
+      if (wordImg.figureNumber && registry) {
+        const entry = registry.byNumber?.get(`fig:${wordImg.figureNumber}`);
+        if (entry && entry.label === origImg.label) {
+          score += 90;
+        }
+      }
+
+      // Match by caption similarity (first 50 chars, normalized)
+      const wordCaption = wordImg.caption.replace(/^(?:Figure|Fig\.?|Table|Tbl\.?)\s+\d+[:\.]?\s*/i, '').toLowerCase().slice(0, 50);
+      const origCaption = origImg.caption.toLowerCase().slice(0, 50);
+      if (wordCaption && origCaption && wordCaption === origCaption) {
+        score += 80;
+      } else if (wordCaption && origCaption && (wordCaption.includes(origCaption.slice(0, 30)) || origCaption.includes(wordCaption.slice(0, 30)))) {
+        score += 40;
+      }
+
+      // Match by path similarity (filename)
+      const wordFile = wordImg.path.split('/').pop()?.toLowerCase() || '';
+      const origFile = origImg.path.split('/').pop()?.toLowerCase() || '';
+      if (wordFile === origFile) {
+        score += 30;
+      }
+
+      if (score > bestScore) {
+        bestScore = score;
+        bestMatch = origImg;
+      }
+    }
+
+    if (bestMatch && bestScore >= 40) {
+      mapping.set(wordImg.placeholder, bestMatch.placeholder);
+      usedOriginals.add(bestMatch.placeholder);
+    }
+  }
+
+  return mapping;
+}
+
+/**
+ * Protect markdown tables before diffing by replacing with placeholders
+ * Tables are treated as atomic blocks to prevent corruption during diff
+ */
+export function protectTables(md: string): ProtectTablesResult {
+  const tables: ProtectedTable[] = [];
+
+  // Match markdown tables: lines starting with | and containing |
+  // A table is: optional caption, header row, separator row (|---|), data rows
+  const tablePattern = /(?:^(?:\*\*)?Table[^\n]*\n\n?)?(?:^\|[^\n]+\|\n)+/gm;
+
+  const text = md.replace(tablePattern, (match) => {
+    // Verify it's actually a table (has separator row with dashes)
+    if (!match.includes('|---') && !match.includes('| ---') && !match.includes('|:--')) {
+      return match; // Not a real table, just lines with pipes
+    }
+
+    const idx = tables.length;
+    const placeholder = `\n\nTABLEBLOCK${idx}ENDTABLE\n\n`;
+
+    // Count cells for matching in Word (approximate)
+    const cellCount = (match.match(/\|/g) || []).length;
+
+    tables.push({ original: match.trim(), placeholder: placeholder.trim(), cellCount });
+    return placeholder;
+  });
+
+  return { text, tables };
+}
+
+/**
+ * Restore tables from placeholders
+ */
+export function restoreTables(text: string, tables: ProtectedTable[]): string {
+  return restoreProtectedItems(text, tables);
+}