docrev 0.8.0 → 0.8.5

package/dist/lib/word.js

@@ -0,0 +1,360 @@
+/**
+ * Word document extraction utilities
+ * Handle reading text, comments, and anchors from .docx files
+ */
+import * as fs from 'fs';
+import AdmZip from 'adm-zip';
+import { parseString } from 'xml2js';
+import { promisify } from 'util';
+const parseXml = promisify(parseString);
+// =============================================================================
+// Constants
+// =============================================================================
+/** Characters of context to extract around comment anchors */
+const ANCHOR_CONTEXT_SIZE = 100;
+/** Characters of context before comment range start */
+const CONTEXT_BEFORE_SIZE = 500;
+// =============================================================================
+// Public API
+// =============================================================================
+/**
+ * Extract comments from Word document's comments.xml
+ * @param docxPath - Path to .docx file
+ * @returns Array of extracted comments
+ * @throws {TypeError} If docxPath is not a string
+ * @throws {Error} If file not found or invalid docx
+ */
+export async function extractWordComments(docxPath) {
+    if (typeof docxPath !== 'string') {
+        throw new TypeError(`docxPath must be a string, got ${typeof docxPath}`);
+    }
+    if (!fs.existsSync(docxPath)) {
+        throw new Error(`File not found: ${docxPath}`);
+    }
+    const zip = new AdmZip(docxPath);
+    const commentsEntry = zip.getEntry('word/comments.xml');
+    if (!commentsEntry) {
+        return []; // No comments in document
+    }
+    const commentsXml = zip.readAsText(commentsEntry);
+    const parsed = await parseXml(commentsXml);
+    if (!parsed?.['w:comments'] || !parsed['w:comments']['w:comment']) {
+        return [];
+    }
+    const comments = [];
+    const rawComments = parsed['w:comments']['w:comment'];
+    for (const comment of rawComments) {
+        const id = comment.$?.['w:id'];
+        const author = comment.$?.['w:author'] || 'Unknown';
+        const date = comment.$?.['w:date'];
+        // Extract text from all paragraphs in comment
+        let text = '';
+        const paragraphs = comment['w:p'] || [];
+        for (const para of paragraphs) {
+            const runs = para['w:r'] || [];
+            for (const run of runs) {
+                const texts = run['w:t'] || [];
+                for (const t of texts) {
+                    text += typeof t === 'string' ? t : (t._ || '');
+                }
+            }
+        }
+        if (id && text.trim()) {
+            comments.push({
+                id,
+                author,
+                date,
+                text: text.trim(),
+            });
+        }
+    }
+    return comments;
+}
+/**
+ * Extract comment anchors (where comments are attached) from document.xml
+ * Returns mapping of comment ID to the text they're anchored to
+ * @param docxPath - Path to .docx file
+ * @returns Map of comment ID to anchor info
+ * @throws {TypeError} If docxPath is not a string
+ * @throws {Error} If invalid docx structure
+ */
+export async function extractCommentAnchors(docxPath) {
+    if (typeof docxPath !== 'string') {
+        throw new TypeError(`docxPath must be a string, got ${typeof docxPath}`);
+    }
+    const zip = new AdmZip(docxPath);
+    const documentEntry = zip.getEntry('word/document.xml');
+    if (!documentEntry) {
+        throw new Error('Invalid docx: no document.xml');
+    }
+    const documentXml = zip.readAsText(documentEntry);
+    const anchors = new Map();
+    // Find commentRangeStart and commentRangeEnd pairs
+    // The text between them is what the comment is anchored to
+    const startPattern = /<w:commentRangeStart w:id="(\d+)"\/>/g;
+    const endPattern = /<w:commentRangeEnd w:id="(\d+)"\/>/g;
+    let match;
+    const starts = new Map();
+    const ends = new Map();
+    while ((match = startPattern.exec(documentXml)) !== null) {
+        if (match[1]) {
+            starts.set(match[1], match.index);
+        }
+    }
+    while ((match = endPattern.exec(documentXml)) !== null) {
+        if (match[1]) {
+            ends.set(match[1], match.index);
+        }
+    }
+    // For each comment, extract the text between start and end
+    for (const [id, startPos] of starts) {
+        const endPos = ends.get(id);
+        if (!endPos)
+            continue;
+        const segment = documentXml.slice(startPos, endPos);
+        // Extract all text content from the segment
+        const textPattern = /<w:t[^>]*>([^<]*)<\/w:t>/g;
+        let text = '';
+        let textMatch;
+        while ((textMatch = textPattern.exec(segment)) !== null) {
+            text += textMatch[1] ?? '';
+        }
+        // Get surrounding context (text before the anchor)
+        const contextStart = Math.max(0, startPos - CONTEXT_BEFORE_SIZE);
+        const contextSegment = documentXml.slice(contextStart, startPos);
+        let context = '';
+        while ((textMatch = textPattern.exec(contextSegment)) !== null) {
+            context += textMatch[1] ?? '';
+        }
+        anchors.set(id, {
+            text: text.trim(),
+            context: context.slice(-ANCHOR_CONTEXT_SIZE),
+        });
+    }
+    return anchors;
+}
+/**
+ * Extract plain text from Word document using mammoth
+ * @param docxPath - Path to .docx file
+ * @returns Extracted plain text
+ * @throws {TypeError} If docxPath is not a string
+ * @throws {Error} If file not found
+ */
+export async function extractTextFromWord(docxPath) {
+    if (typeof docxPath !== 'string') {
+        throw new TypeError(`docxPath must be a string, got ${typeof docxPath}`);
+    }
+    if (!fs.existsSync(docxPath)) {
+        throw new Error(`File not found: ${docxPath}`);
+    }
+    const mammoth = await import('mammoth');
+    const result = await mammoth.extractRawText({ path: docxPath });
+    return result.value;
+}
+/**
+ * Extract rich content from Word with basic formatting
+ * @param docxPath - Path to .docx file
+ * @returns Text and HTML content
+ * @throws {TypeError} If docxPath is not a string
+ * @throws {Error} If file not found
+ */
+export async function extractFromWord(docxPath) {
+    if (typeof docxPath !== 'string') {
+        throw new TypeError(`docxPath must be a string, got ${typeof docxPath}`);
+    }
+    if (!fs.existsSync(docxPath)) {
+        throw new Error(`File not found: ${docxPath}`);
+    }
+    const mammoth = await import('mammoth');
+    const [textResult, htmlResult] = await Promise.all([
+        mammoth.extractRawText({ path: docxPath }),
+        mammoth.convertToHtml({ path: docxPath }),
+    ]);
+    return {
+        text: textResult.value,
+        html: htmlResult.value,
+    };
+}
+/**
+ * Get document metadata from Word file
+ * @param docxPath - Path to .docx file
+ * @returns Document metadata
+ * @throws {TypeError} If docxPath is not a string
+ */
+export async function getWordMetadata(docxPath) {
+    if (typeof docxPath !== 'string') {
+        throw new TypeError(`docxPath must be a string, got ${typeof docxPath}`);
+    }
+    const zip = new AdmZip(docxPath);
+    const coreEntry = zip.getEntry('docProps/core.xml');
+    if (!coreEntry) {
+        return {};
+    }
+    const coreXml = zip.readAsText(coreEntry);
+    const metadata = {};
+    // Extract common metadata fields
+    const patterns = {
+        title: /<dc:title>([^<]*)<\/dc:title>/,
+        author: /<dc:creator>([^<]*)<\/dc:creator>/,
+        created: /<dcterms:created[^>]*>([^<]*)<\/dcterms:created>/,
+        modified: /<dcterms:modified[^>]*>([^<]*)<\/dcterms:modified>/,
+    };
+    for (const [key, pattern] of Object.entries(patterns)) {
+        const match = coreXml.match(pattern);
+        if (match) {
+            metadata[key] = match[1];
+        }
+    }
+    return metadata;
+}
+/**
+ * Check if file is a valid Word document
+ * @param filePath - Path to file to check
+ * @returns True if valid .docx file
+ */
+export function isWordDocument(filePath) {
+    if (typeof filePath !== 'string')
+        return false;
+    if (!fs.existsSync(filePath))
+        return false;
+    if (!filePath.toLowerCase().endsWith('.docx'))
+        return false;
+    try {
+        const zip = new AdmZip(filePath);
+        return zip.getEntry('word/document.xml') !== null;
+    }
+    catch {
+        return false;
+    }
+}
+/**
+ * Extract text content from XML element, handling nested elements
+ * @param xml - XML string
+ * @returns Plain text content
+ */
+function extractTextFromXml(xml) {
+    let text = '';
+    // Match w:t elements (regular text)
+    const textPattern = /<w:t[^>]*>([^<]*)<\/w:t>/g;
+    let match;
+    while ((match = textPattern.exec(xml)) !== null) {
+        text += match[1];
+    }
+    // Also match w:delText (deleted text)
+    const delTextPattern = /<w:delText[^>]*>([^<]*)<\/w:delText>/g;
+    while ((match = delTextPattern.exec(xml)) !== null) {
+        text += match[1];
+    }
+    return text;
+}
+/**
+ * Extract track changes (insertions and deletions) from Word document
+ * Converts Word's w:ins and w:del elements to CriticMarkup format
+ *
+ * @param docxPath - Path to Word document
+ * @returns Track changes result with content and stats
+ */
+export async function extractTrackChanges(docxPath) {
+    if (!fs.existsSync(docxPath)) {
+        throw new Error(`File not found: ${docxPath}`);
+    }
+    const zip = new AdmZip(docxPath);
+    const documentEntry = zip.getEntry('word/document.xml');
+    if (!documentEntry) {
+        throw new Error('Invalid docx: no document.xml');
+    }
+    let xml = zip.readAsText(documentEntry);
+    let insertions = 0;
+    let deletions = 0;
+    // Check if there are any track changes
+    const hasInsertions = xml.includes('<w:ins ');
+    const hasDeletions = xml.includes('<w:del ');
+    if (!hasInsertions && !hasDeletions) {
+        return { hasTrackChanges: false, content: null, stats: { insertions: 0, deletions: 0 } };
+    }
+    // Process insertions: <w:ins ...>...</w:ins> -> {++...++}
+    // Match the full w:ins element including nested content
+    xml = xml.replace(/<w:ins\b[^>]*>([\s\S]*?)<\/w:ins>/g, (match, content) => {
+        const text = extractTextFromXml(content);
+        if (text.trim()) {
+            insertions++;
+            return `{++${text}++}`;
+        }
+        return text;
+    });
+    // Process deletions: <w:del ...>...</w:del> -> {--...--}
+    xml = xml.replace(/<w:del\b[^>]*>([\s\S]*?)<\/w:del>/g, (match, content) => {
+        const text = extractTextFromXml(content);
+        if (text.trim()) {
+            deletions++;
+            return `{--${text}--}`;
+        }
+        return '';
+    });
+    return {
+        hasTrackChanges: true,
+        content: xml,
+        stats: { insertions, deletions },
+    };
+}
+/**
+ * Extract Word document content with track changes preserved as CriticMarkup
+ * Uses pandoc with track-changes=all option to preserve insertions/deletions
+ *
+ * @param docxPath - Path to Word document
+ * @param options - Options
+ * @returns Track changes result with text and stats
+ */
+export async function extractWithTrackChanges(docxPath, options = {}) {
+    const { mediaDir } = options;
+    if (!fs.existsSync(docxPath)) {
+        throw new Error(`File not found: ${docxPath}`);
+    }
+    const { execSync } = await import('child_process');
+    // Use pandoc with --track-changes=all to preserve track changes
+    // This outputs insertions as [insertion]{.insertion} and deletions as [deletion]{.deletion}
+    let pandocArgs = `"${docxPath}" -t markdown --wrap=none --track-changes=all`;
+    if (mediaDir) {
+        pandocArgs += ` --extract-media="${mediaDir}"`;
+    }
+    let text;
+    try {
+        text = execSync(`pandoc ${pandocArgs}`, {
+            encoding: 'utf-8',
+            maxBuffer: 50 * 1024 * 1024,
+        });
+    }
+    catch (err) {
+        throw new Error(`Pandoc extraction failed: ${err.message}`);
+    }
+    // Count track changes from pandoc output
+    let insertions = 0;
+    let deletions = 0;
+    // Pandoc outputs track changes as:
+    // [inserted text]{.insertion author="..."}
+    // [deleted text]{.deletion author="..."}
+    // Convert pandoc's track change format to CriticMarkup
+    // Insertions: [text]{.insertion ...} -> {++text++}
+    text = text.replace(/\[([^\]]*)\]\{\.insertion[^}]*\}/g, (match, content) => {
+        if (content.trim()) {
+            insertions++;
+            return `{++${content}++}`;
+        }
+        return '';
+    });
+    // Deletions: [text]{.deletion ...} -> {--text--}
+    text = text.replace(/\[([^\]]*)\]\{\.deletion[^}]*\}/g, (match, content) => {
+        if (content.trim()) {
+            deletions++;
+            return `{--${content}--}`;
+        }
+        return '';
+    });
+    const hasTrackChanges = insertions > 0 || deletions > 0;
+    return {
+        text,
+        hasTrackChanges,
+        stats: { insertions, deletions },
+    };
+}
+//# sourceMappingURL=word.js.map

package/dist/lib/word.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"word.js","sourceRoot":"","sources":["../../lib/word.ts"],"names":[],"mappings":"AAAA;;;GAGG;AAEH,OAAO,KAAK,EAAE,MAAM,IAAI,CAAC;AAEzB,OAAO,MAAM,MAAM,SAAS,CAAC;AAC7B,OAAO,EAAE,WAAW,EAAE,MAAM,QAAQ,CAAC;AACrC,OAAO,EAAE,SAAS,EAAE,MAAM,MAAM,CAAC;AAGjC,MAAM,QAAQ,GAAG,SAAS,CAAC,WAAW,CAAC,CAAC;AAExC,gFAAgF;AAChF,YAAY;AACZ,gFAAgF;AAEhF,8DAA8D;AAC9D,MAAM,mBAAmB,GAAG,GAAG,CAAC;AAEhC,uDAAuD;AACvD,MAAM,mBAAmB,GAAG,GAAG,CAAC;AAEhC,gFAAgF;AAChF,aAAa;AACb,gFAAgF;AAEhF;;;;;;GAMG;AACH,MAAM,CAAC,KAAK,UAAU,mBAAmB,CAAC,QAAgB;IACxD,IAAI,OAAO,QAAQ,KAAK,QAAQ,EAAE,CAAC;QACjC,MAAM,IAAI,SAAS,CAAC,kCAAkC,OAAO,QAAQ,EAAE,CAAC,CAAC;IAC3E,CAAC;IACD,IAAI,CAAC,EAAE,CAAC,UAAU,CAAC,QAAQ,CAAC,EAAE,CAAC;QAC7B,MAAM,IAAI,KAAK,CAAC,mBAAmB,QAAQ,EAAE,CAAC,CAAC;IACjD,CAAC;IAED,MAAM,GAAG,GAAG,IAAI,MAAM,CAAC,QAAQ,CAAC,CAAC;IACjC,MAAM,aAAa,GAAG,GAAG,CAAC,QAAQ,CAAC,mBAAmB,CAAC,CAAC;IAExD,IAAI,CAAC,aAAa,EAAE,CAAC;QACnB,OAAO,EAAE,CAAC,CAAC,0BAA0B;IACvC,CAAC;IAED,MAAM,WAAW,GAAG,GAAG,CAAC,UAAU,CAAC,aAAa,CAAC,CAAC;IAClD,MAAM,MAAM,GAAG,MAAM,QAAQ,CAAC,WAAW,CAAQ,CAAC;IAElD,IAAI,CAAC,MAAM,EAAE,CAAC,YAAY,CAAC,IAAI,CAAC,MAAM,CAAC,YAAY,CAAC,CAAC,WAAW,CAAC,EAAE,CAAC;QAClE,OAAO,EAAE,CAAC;IACZ,CAAC;IAED,MAAM,QAAQ,GAAkB,EAAE,CAAC;IACnC,MAAM,WAAW,GAAG,MAAM,CAAC,YAAY,CAAC,CAAC,WAAW,CAAC,CAAC;IAEtD,KAAK,MAAM,OAAO,IAAI,WAAW,EAAE,CAAC;QAClC,MAAM,EAAE,GAAG,OAAO,CAAC,CAAC,EAAE,CAAC,MAAM,CAAC,CAAC;QAC/B,MAAM,MAAM,GAAG,OAAO,CAAC,CAAC,EAAE,CAAC,UAAU,CAAC,IAAI,SAAS,CAAC;QACpD,MAAM,IAAI,GAAG,OAAO,CAAC,CAAC,EAAE,CAAC,QAAQ,CAAC,CAAC;QAEnC,8CAA8C;QAC9C,IAAI,IAAI,GAAG,EAAE,CAAC;QACd,MAAM,UAAU,GAAG,OAAO,CAAC,KAAK,CAAC,IAAI,EAAE,CAAC;QACxC,KAAK,MAAM,IAAI,IAAI,UAAU,EAAE,CAAC;YAC9B,MAAM,IAAI,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,EAAE,CAAC;YAC/B,KAAK,MAAM,GAAG,IAAI,IAAI,EAAE,CAAC;gBACvB,MAAM,KAAK,GAAG,GAAG,CAAC,KAAK,CAAC,IAAI,EAAE,CAAC;gBAC/B,KAAK,MAAM,CAAC,IAAI,KAAK,EAAE,CAAC;oBACtB,IAAI,IAAI,OAAO,CAAC,KAAK,QAAQ,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC;gBAClD,CAAC;YACH,CAAC;QACH,CAAC;QAED,IAAI,EAAE,IAAI,IAAI,CAAC,IAAI,EAAE,EAAE,CAAC;YACtB,QAAQ,CAAC,IAAI,CAAC;gBACZ,EAAE;gBACF,MAAM;gBACN,IAAI;gBACJ,IAAI,EAAE,IAAI,CAAC,IAAI,EAAE;aAClB,CAAC,CAAC;QACL,CAAC;IACH,CAAC;IAED,OAAO,QAAQ,CAAC;AAClB,CAAC;AAED;;;;;;;GAOG;AACH,MAAM,CAAC,KAAK,UAAU,qBAAqB,CAAC,QAAgB;IAC1D,IAAI,OAAO,QAAQ,KAAK,QAAQ,EAAE,CAAC;QACjC,MAAM,IAAI,SAAS,CAAC,kCAAkC,OAAO,QAAQ,EAAE,CAAC,CAAC;IAC3E,CAAC;IAED,MAAM,GAAG,GAAG,IAAI,MAAM,CAAC,QAAQ,CAAC,CAAC;IACjC,MAAM,aAAa,GAAG,GAAG,CAAC,QAAQ,CAAC,mBAAmB,CAAC,CAAC;IAExD,IAAI,CAAC,aAAa,EAAE,CAAC;QACnB,MAAM,IAAI,KAAK,CAAC,+BAA+B,CAAC,CAAC;IACnD,CAAC;IAED,MAAM,WAAW,GAAG,GAAG,CAAC,UAAU,CAAC,aAAa,CAAC,CAAC;IAClD,MAAM,OAAO,GAAG,IAAI,GAAG,EAAyB,CAAC;IAEjD,mDAAmD;IACnD,2DAA2D;IAC3D,MAAM,YAAY,GAAG,uCAAuC,CAAC;IAC7D,MAAM,UAAU,GAAG,qCAAqC,CAAC;IAEzD,IAAI,KAA6B,CAAC;IAClC,MAAM,MAAM,GAAG,IAAI,GAAG,EAAkB,CAAC;IACzC,MAAM,IAAI,GAAG,IAAI,GAAG,EAAkB,CAAC;IAEvC,OAAO,CAAC,KAAK,GAAG,YAAY,CAAC,IAAI,CAAC,WAAW,CAAC,CAAC,KAAK,IAAI,EAAE,CAAC;QACzD,IAAI,KAAK,CAAC,CAAC,CAAC,EAAE,CAAC;YACb,MAAM,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,KAAK,CAAC,KAAK,CAAC,CAAC;QACpC,CAAC;IACH,CAAC;IAED,OAAO,CAAC,KAAK,GAAG,UAAU,CAAC,IAAI,CAAC,WAAW,CAAC,CAAC,KAAK,IAAI,EAAE,CAAC;QACvD,IAAI,KAAK,CAAC,CAAC,CAAC,EAAE,CAAC;YACb,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,KAAK,CAAC,KAAK,CAAC,CAAC;QAClC,CAAC;IACH,CAAC;IAED,2DAA2D;IAC3D,KAAK,MAAM,CAAC,EAAE,EAAE,QAAQ,CAAC,IAAI,MAAM,EAAE,CAAC;QACpC,MAAM,MAAM,GAAG,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;QAC5B,IAAI,CAAC,MAAM;YAAE,SAAS;QAEtB,MAAM,OAAO,GAAG,WAAW,CAAC,KAAK,CAAC,QAAQ,EAAE,MAAM,CAAC,CAAC;QAEpD,4CAA4C;QAC5C,MAAM,WAAW,GAAG,2BAA2B,CAAC;QAChD,IAAI,IAAI,GAAG,EAAE,CAAC;QACd,IAAI,SAAiC,CAAC;QACtC,OAAO,CAAC,SAAS,GAAG,WAAW,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,KAAK,IAAI,EAAE,CAAC;YACxD,IAAI,IAAI,SAAS,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;QAC7B,CAAC;QAED,mDAAmD;QACnD,MAAM,YAAY,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,QAAQ,GAAG,mBAAmB,CAAC,CAAC;QACjE,MAAM,cAAc,GAAG,WAAW,CAAC,KAAK,CAAC,YAAY,EAAE,QAAQ,CAAC,CAAC;QACjE,IAAI,OAAO,GAAG,EAAE,CAAC;QACjB,OAAO,CAAC,SAAS,GAAG,WAAW,CAAC,IAAI,CAAC,cAAc,CAAC,CAAC,KAAK,IAAI,EAAE,CAAC;YAC/D,OAAO,IAAI,SAAS,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;QAChC,CAAC;QAED,OAAO,CAAC,GAAG,CAAC,EAAE,EAAE;YACd,IAAI,EAAE,IAAI,CAAC,IAAI,EAAE;YACjB,OAAO,EAAE,OAAO,CAAC,KAAK,CAAC,CAAC,mBAAmB,CAAC;SAC7C,CAAC,CAAC;IACL,CAAC;IAED,OAAO,OAAO,CAAC;AACjB,CAAC;AAED;;;;;;GAMG;AACH,MAAM,CAAC,KAAK,UAAU,mBAAmB,CAAC,QAAgB;IACxD,IAAI,OAAO,QAAQ,KAAK,QAAQ,EAAE,CAAC;QACjC,MAAM,IAAI,SAAS,CAAC,kCAAkC,OAAO,QAAQ,EAAE,CAAC,CAAC;IAC3E,CAAC;IACD,IAAI,CAAC,EAAE,CAAC,UAAU,CAAC,QAAQ,CAAC,EAAE,CAAC;QAC7B,MAAM,IAAI,KAAK,CAAC,mBAAmB,QAAQ,EAAE,CAAC,CAAC;IACjD,CAAC;IAED,MAAM,OAAO,GAAG,MAAM,MAAM,CAAC,SAAS,CAAC,CAAC;IACxC,MAAM,MAAM,GAAG,MAAM,OAAO,CAAC,cAAc,CAAC,EAAE,IAAI,EAAE,QAAQ,EAAE,CAAC,CAAC;IAChE,OAAO,MAAM,CAAC,KAAK,CAAC;AACtB,CAAC;AAED;;;;;;GAMG;AACH,MAAM,CAAC,KAAK,UAAU,eAAe,CAAC,QAAgB;IACpD,IAAI,OAAO,QAAQ,KAAK,QAAQ,EAAE,CAAC;QACjC,MAAM,IAAI,SAAS,CAAC,kCAAkC,OAAO,QAAQ,EAAE,CAAC,CAAC;IAC3E,CAAC;IACD,IAAI,CAAC,EAAE,CAAC,UAAU,CAAC,QAAQ,CAAC,EAAE,CAAC;QAC7B,MAAM,IAAI,KAAK,CAAC,mBAAmB,QAAQ,EAAE,CAAC,CAAC;IACjD,CAAC;IAED,MAAM,OAAO,GAAG,MAAM,MAAM,CAAC,SAAS,CAAC,CAAC;IAExC,MAAM,CAAC,UAAU,EAAE,UAAU,CAAC,GAAG,MAAM,OAAO,CAAC,GAAG,CAAC;QACjD,OAAO,CAAC,cAAc,CAAC,EAAE,IAAI,EAAE,QAAQ,EAAE,CAAC;QAC1C,OAAO,CAAC,aAAa,CAAC,EAAE,IAAI,EAAE,QAAQ,EAAE,CAAC;KAC1C,CAAC,CAAC;IAEH,OAAO;QACL,IAAI,EAAE,UAAU,CAAC,KAAK;QACtB,IAAI,EAAE,UAAU,CAAC,KAAK;KACvB,CAAC;AACJ,CAAC;AAED;;;;;GAKG;AACH,MAAM,CAAC,KAAK,UAAU,eAAe,CAAC,QAAgB;IACpD,IAAI,OAAO,QAAQ,KAAK,QAAQ,EAAE,CAAC;QACjC,MAAM,IAAI,SAAS,CAAC,kCAAkC,OAAO,QAAQ,EAAE,CAAC,CAAC;IAC3E,CAAC;IAED,MAAM,GAAG,GAAG,IAAI,MAAM,CAAC,QAAQ,CAAC,CAAC;IACjC,MAAM,SAAS,GAAG,GAAG,CAAC,QAAQ,CAAC,mBAAmB,CAAC,CAAC;IAEpD,IAAI,CAAC,SAAS,EAAE,CAAC;QACf,OAAO,EAAE,CAAC;IACZ,CAAC;IAED,MAAM,OAAO,GAAG,GAAG,CAAC,UAAU,CAAC,SAAS,CAAC,CAAC;IAC1C,MAAM,QAAQ,GAAiB,EAAE,CAAC;IAElC,iCAAiC;IACjC,MAAM,QAAQ,GAA2B;QACvC,KAAK,EAAE,+BAA+B;QACtC,MAAM,EAAE,mCAAmC;QAC3C,OAAO,EAAE,kDAAkD;QAC3D,QAAQ,EAAE,oDAAoD;KAC/D,CAAC;IAEF,KAAK,MAAM,CAAC,GAAG,EAAE,OAAO,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,QAAQ,CAAC,EAAE,CAAC;QACtD,MAAM,KAAK,GAAG,OAAO,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC;QACrC,IAAI,KAAK,EAAE,CAAC;YACT,QAAgB,CAAC,GAAG,CAAC,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC;QACpC,CAAC;IACH,CAAC;IAED,OAAO,QAAQ,CAAC;AAClB,CAAC;AAED;;;;GAIG;AACH,MAAM,UAAU,cAAc,CAAC,QAAgB;IAC7C,IAAI,OAAO,QAAQ,KAAK,QAAQ;QAAE,OAAO,KAAK,CAAC;IAC/C,IAAI,CAAC,EAAE,CAAC,UAAU,CAAC,QAAQ,CAAC;QAAE,OAAO,KAAK,CAAC;IAC3C,IAAI,CAAC,QAAQ,CAAC,WAAW,EAAE,CAAC,QAAQ,CAAC,OAAO,CAAC;QAAE,OAAO,KAAK,CAAC;IAE5D,IAAI,CAAC;QACH,MAAM,GAAG,GAAG,IAAI,MAAM,CAAC,QAAQ,CAAC,CAAC;QACjC,OAAO,GAAG,CAAC,QAAQ,CAAC,mBAAmB,CAAC,KAAK,IAAI,CAAC;IACpD,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,KAAK,CAAC;IACf,CAAC;AACH,CAAC;AAED;;;;GAIG;AACH,SAAS,kBAAkB,CAAC,GAAW;IACrC,IAAI,IAAI,GAAG,EAAE,CAAC;IACd,oCAAoC;IACpC,MAAM,WAAW,GAAG,2BAA2B,CAAC;IAChD,IAAI,KAA6B,CAAC;IAClC,OAAO,CAAC,KAAK,GAAG,WAAW,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,KAAK,IAAI,EAAE,CAAC;QAChD,IAAI,IAAI,KAAK,CAAC,CAAC,CAAC,CAAC;IACnB,CAAC;IACD,sCAAsC;IACtC,MAAM,cAAc,GAAG,uCAAuC,CAAC;IAC/D,OAAO,CAAC,KAAK,GAAG,cAAc,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,KAAK,IAAI,EAAE,CAAC;QACnD,IAAI,IAAI,KAAK,CAAC,CAAC,CAAC,CAAC;IACnB,CAAC;IACD,OAAO,IAAI,CAAC;AACd,CAAC;AAED;;;;;;GAMG;AACH,MAAM,CAAC,KAAK,UAAU,mBAAmB,CAAC,QAAgB;IACxD,IAAI,CAAC,EAAE,CAAC,UAAU,CAAC,QAAQ,CAAC,EAAE,CAAC;QAC7B,MAAM,IAAI,KAAK,CAAC,mBAAmB,QAAQ,EAAE,CAAC,CAAC;IACjD,CAAC;IAED,MAAM,GAAG,GAAG,IAAI,MAAM,CAAC,QAAQ,CAAC,CAAC;IACjC,MAAM,aAAa,GAAG,GAAG,CAAC,QAAQ,CAAC,mBAAmB,CAAC,CAAC;IAExD,IAAI,CAAC,aAAa,EAAE,CAAC;QACnB,MAAM,IAAI,KAAK,CAAC,+BAA+B,CAAC,CAAC;IACnD,CAAC;IAED,IAAI,GAAG,GAAG,GAAG,CAAC,UAAU,CAAC,aAAa,CAAC,CAAC;IACxC,IAAI,UAAU,GAAG,CAAC,CAAC;IACnB,IAAI,SAAS,GAAG,CAAC,CAAC;IAElB,uCAAuC;IACvC,MAAM,aAAa,GAAG,GAAG,CAAC,QAAQ,CAAC,SAAS,CAAC,CAAC;IAC9C,MAAM,YAAY,GAAG,GAAG,CAAC,QAAQ,CAAC,SAAS,CAAC,CAAC;IAE7C,IAAI,CAAC,aAAa,IAAI,CAAC,YAAY,EAAE,CAAC;QACpC,OAAO,EAAE,eAAe,EAAE,KAAK,EAAE,OAAO,EAAE,IAAI,EAAE,KAAK,EAAE,EAAE,UAAU,EAAE,CAAC,EAAE,SAAS,EAAE,CAAC,EAAE,EAAE,CAAC;IAC3F,CAAC;IAED,0DAA0D;IAC1D,wDAAwD;IACxD,GAAG,GAAG,GAAG,CAAC,OAAO,CAAC,oCAAoC,EAAE,CAAC,KAAK,EAAE,OAAO,EAAE,EAAE;QACzE,MAAM,IAAI,GAAG,kBAAkB,CAAC,OAAO,CAAC,CAAC;QACzC,IAAI,IAAI,CAAC,IAAI,EAAE,EAAE,CAAC;YAChB,UAAU,EAAE,CAAC;YACb,OAAO,MAAM,IAAI,KAAK,CAAC;QACzB,CAAC;QACD,OAAO,IAAI,CAAC;IACd,CAAC,CAAC,CAAC;IAEH,yDAAyD;IACzD,GAAG,GAAG,GAAG,CAAC,OAAO,CAAC,oCAAoC,EAAE,CAAC,KAAK,EAAE,OAAO,EAAE,EAAE;QACzE,MAAM,IAAI,GAAG,kBAAkB,CAAC,OAAO,CAAC,CAAC;QACzC,IAAI,IAAI,CAAC,IAAI,EAAE,EAAE,CAAC;YAChB,SAAS,EAAE,CAAC;YACZ,OAAO,MAAM,IAAI,KAAK,CAAC;QACzB,CAAC;QACD,OAAO,EAAE,CAAC;IACZ,CAAC,CAAC,CAAC;IAEH,OAAO;QACL,eAAe,EAAE,IAAI;QACrB,OAAO,EAAE,GAAG;QACZ,KAAK,EAAE,EAAE,UAAU,EAAE,SAAS,EAAE;KACjC,CAAC;AACJ,CAAC;AAMD;;;;;;;GAOG;AACH,MAAM,CAAC,KAAK,UAAU,uBAAuB,CAC3C,QAAgB,EAChB,UAA0C,EAAE;IAE5C,MAAM,EAAE,QAAQ,EAAE,GAAG,OAAO,CAAC;IAE7B,IAAI,CAAC,EAAE,CAAC,UAAU,CAAC,QAAQ,CAAC,EAAE,CAAC;QAC7B,MAAM,IAAI,KAAK,CAAC,mBAAmB,QAAQ,EAAE,CAAC,CAAC;IACjD,CAAC;IAED,MAAM,EAAE,QAAQ,EAAE,GAAG,MAAM,MAAM,CAAC,eAAe,CAAC,CAAC;IAEnD,gEAAgE;IAChE,4FAA4F;IAC5F,IAAI,UAAU,GAAG,IAAI,QAAQ,+CAA+C,CAAC;IAC7E,IAAI,QAAQ,EAAE,CAAC;QACb,UAAU,IAAI,qBAAqB,QAAQ,GAAG,CAAC;IACjD,CAAC;IAED,IAAI,IAAY,CAAC;IACjB,IAAI,CAAC;QACH,IAAI,GAAG,QAAQ,CAAC,UAAU,UAAU,EAAE,EAAE;YACtC,QAAQ,EAAE,OAAO;YACjB,SAAS,EAAE,EAAE,GAAG,IAAI,GAAG,IAAI;SAC5B,CAAC,CAAC;IACL,CAAC;IAAC,OAAO,GAAQ,EAAE,CAAC;QAClB,MAAM,IAAI,KAAK,CAAC,6BAA6B,GAAG,CAAC,OAAO,EAAE,CAAC,CAAC;IAC9D,CAAC;IAED,yCAAyC;IACzC,IAAI,UAAU,GAAG,CAAC,CAAC;IACnB,IAAI,SAAS,GAAG,CAAC,CAAC;IAElB,mCAAmC;IACnC,2CAA2C;IAC3C,yCAAyC;IAEzC,uDAAuD;IACvD,mDAAmD;IACnD,IAAI,GAAG,IAAI,CAAC,OAAO,CAAC,mCAAmC,EAAE,CAAC,KAAK,EAAE,OAAO,EAAE,EAAE;QAC1E,IAAI,OAAO,CAAC,IAAI,EAAE,EAAE,CAAC;YACnB,UAAU,EAAE,CAAC;YACb,OAAO,MAAM,OAAO,KAAK,CAAC;QAC5B,CAAC;QACD,OAAO,EAAE,CAAC;IACZ,CAAC,CAAC,CAAC;IAEH,iDAAiD;IACjD,IAAI,GAAG,IAAI,CAAC,OAAO,CAAC,kCAAkC,EAAE,CAAC,KAAK,EAAE,OAAO,EAAE,EAAE;QACzE,IAAI,OAAO,CAAC,IAAI,EAAE,EAAE,CAAC;YACnB,SAAS,EAAE,CAAC;YACZ,OAAO,MAAM,OAAO,KAAK,CAAC;QAC5B,CAAC;QACD,OAAO,EAAE,CAAC;IACZ,CAAC,CAAC,CAAC;IAEH,MAAM,eAAe,GAAG,UAAU,GAAG,CAAC,IAAI,SAAS,GAAG,CAAC,CAAC;IAExD,OAAO;QACL,IAAI;QACJ,eAAe;QACf,KAAK,EAAE,EAAE,UAAU,EAAE,SAAS,EAAE;KACjC,CAAC;AACJ,CAAC"}

package/dist/lib/wordcomments.d.ts

@@ -0,0 +1,51 @@
+/**
+ * Word comment injection with reply threading
+ *
+ * Flow:
+ * 1. prepareMarkdownWithMarkers() - Parse comments, detect reply relationships
+ *    - First comment in a cluster = parent (gets markers: ⟦CMS:n⟧anchor⟦CME:n⟧)
+ *    - Subsequent adjacent comments = replies (no markers, attach to parent)
+ * 2. Pandoc converts to DOCX
+ * 3. injectCommentsAtMarkers() - Insert comment ranges for parents only
+ *    - Replies go in comments.xml with parent reference in commentsExtended.xml
+ */
+interface ParsedComment {
+    author: string;
+    text: string;
+    anchor: string | null;
+    start: number;
+    end: number;
+    fullMatch: string;
+}
+interface PreparedComment extends ParsedComment {
+    isReply: boolean;
+    parentIdx: number | null;
+    commentIdx: number;
+    anchorFromReply?: boolean;
+    placesParentMarkers?: boolean;
+}
+interface PrepareResult {
+    markedMarkdown: string;
+    comments: PreparedComment[];
+}
+interface InjectionResult {
+    success: boolean;
+    commentCount: number;
+    replyCount?: number;
+    skippedComments: number;
+    error?: string;
+}
+/**
+ * Parse comments and create markers
+ *
+ * Returns:
+ * - markedMarkdown: markdown with markers for parent comments only
+ * - comments: array with author, text, isReply, parentIdx
+ */
+export declare function prepareMarkdownWithMarkers(markdown: string): PrepareResult;
+/**
+ * Inject comments at marker positions
+ */
+export declare function injectCommentsAtMarkers(docxPath: string, comments: PreparedComment[], outputPath: string): Promise<InjectionResult>;
+export {};
+//# sourceMappingURL=wordcomments.d.ts.map

package/dist/lib/wordcomments.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"wordcomments.d.ts","sourceRoot":"","sources":["../../lib/wordcomments.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;GAUG;AASH,UAAU,aAAa;IACrB,MAAM,EAAE,MAAM,CAAC;IACf,IAAI,EAAE,MAAM,CAAC;IACb,MAAM,EAAE,MAAM,GAAG,IAAI,CAAC;IACtB,KAAK,EAAE,MAAM,CAAC;IACd,GAAG,EAAE,MAAM,CAAC;IACZ,SAAS,EAAE,MAAM,CAAC;CACnB;AAED,UAAU,eAAgB,SAAQ,aAAa;IAC7C,OAAO,EAAE,OAAO,CAAC;IACjB,SAAS,EAAE,MAAM,GAAG,IAAI,CAAC;IACzB,UAAU,EAAE,MAAM,CAAC;IACnB,eAAe,CAAC,EAAE,OAAO,CAAC;IAC1B,mBAAmB,CAAC,EAAE,OAAO,CAAC;CAC/B;AAED,UAAU,aAAa;IACrB,cAAc,EAAE,MAAM,CAAC;IACvB,QAAQ,EAAE,eAAe,EAAE,CAAC;CAC7B;AAUD,UAAU,eAAe;IACvB,OAAO,EAAE,OAAO,CAAC;IACjB,YAAY,EAAE,MAAM,CAAC;IACrB,UAAU,CAAC,EAAE,MAAM,CAAC;IACpB,eAAe,EAAE,MAAM,CAAC;IACxB,KAAK,CAAC,EAAE,MAAM,CAAC;CAChB;AAmBD;;;;;;GAMG;AACH,wBAAgB,0BAA0B,CAAC,QAAQ,EAAE,MAAM,GAAG,aAAa,CAuJ1E;AA4ID;;GAEG;AACH,wBAAsB,uBAAuB,CAC3C,QAAQ,EAAE,MAAM,EAChB,QAAQ,EAAE,eAAe,EAAE,EAC3B,UAAU,EAAE,MAAM,GACjB,OAAO,CAAC,eAAe,CAAC,CAwU1B"}