docrev 0.11.3 → 0.12.0

package/lib/response.ts

@@ -1,200 +0,0 @@
-/**
- * Response letter generator
- * Extract comments and replies from markdown files for journal resubmission
- */
-
-import * as fs from 'fs';
-import * as path from 'path';
-
-interface Reply {
-  author: string;
-  text: string;
-}
-
-interface CommentWithReplies {
-  author: string;
-  text: string;
-  replies: Reply[];
-  context: string;
-  file: string;
-  line: number;
-}
-
-interface ResponseOptions {
-  title?: string;
-  authorName?: string;
-  includeContext?: boolean;
-  includeLocation?: boolean;
-}
-
-/**
- * Parse a comment with potential replies
- * Format: {>>Author: comment<<} {>>Reply Author: reply<<}
- */
-export function parseCommentsWithReplies(text: string, file: string = ''): CommentWithReplies[] {
-  const comments: CommentWithReplies[] = [];
-  const lines = text.split('\n');
-
-  // Pattern for comments: {>>Author: text<<}
-  const commentPattern = /\{>>([^:]+):\s*([^<]+)<<\}/g;
-
-  for (let lineNum = 0; lineNum < lines.length; lineNum++) {
-    const line = lines[lineNum];
-    if (!line) continue;
-    const matches = [...line.matchAll(commentPattern)];
-
-    if (matches.length === 0) continue;
-
-    // Get context (surrounding text without comments)
-    const contextLine = line.replace(/\{>>[\s\S]+?<<\}/g, '').trim();
-    const context = contextLine.slice(0, 100) + (contextLine.length > 100 ? '...' : '');
-
-    // First match is the original comment, rest are replies
-    const [first, ...rest] = matches;
-    if (!first || !first[1] || !first[2]) continue;
-
-    comments.push({
-      author: first[1].trim(),
-      text: first[2].trim(),
-      replies: rest.map(m => ({
-        author: m[1]?.trim() || '',
-        text: m[2]?.trim() || '',
-      })),
-      context,
-      file,
-      line: lineNum + 1,
-    });
-  }
-
-  return comments;
-}
-
-/**
- * Group comments by reviewer
- */
-export function groupByReviewer(comments: CommentWithReplies[]): Map<string, CommentWithReplies[]> {
-  const grouped = new Map<string, CommentWithReplies[]>();
-
-  for (const comment of comments) {
-    const reviewer = comment.author;
-    if (!grouped.has(reviewer)) {
-      grouped.set(reviewer, []);
-    }
-    grouped.get(reviewer)!.push(comment);
-  }
-
-  return grouped;
-}
-
-/**
- * Generate response letter in Markdown format
- */
-export function generateResponseLetter(comments: CommentWithReplies[], options: ResponseOptions = {}): string {
-  const {
-    title = 'Response to Reviewers',
-    authorName = 'Author',
-    includeContext = true,
-    includeLocation = true,
-  } = options;
-
-  const lines: string[] = [];
-  lines.push(`# ${title}`);
-  lines.push('');
-  lines.push(`We thank the reviewers for their constructive feedback. Below we address each comment.`);
-  lines.push('');
-
-  // Group by reviewer
-  const grouped = groupByReviewer(comments);
-
-  // Sort reviewers (put known reviewer names first, then others)
-  const reviewers = [...grouped.keys()].sort((a, b) => {
-    // "Reviewer" names first, then alphabetical
-    const aIsReviewer = a.toLowerCase().includes('reviewer');
-    const bIsReviewer = b.toLowerCase().includes('reviewer');
-    if (aIsReviewer && !bIsReviewer) return -1;
-    if (!aIsReviewer && bIsReviewer) return 1;
-    return a.localeCompare(b);
-  });
-
-  for (const reviewer of reviewers) {
-    // Skip if this is the author's own comments (replies)
-    if (reviewer.toLowerCase() === authorName.toLowerCase()) continue;
-    if (reviewer.toLowerCase() === 'claude') continue;
-
-    const reviewerComments = grouped.get(reviewer)!;
-    lines.push(`## ${reviewer}`);
-    lines.push('');
-
-    for (let i = 0; i < reviewerComments.length; i++) {
-      const c = reviewerComments[i];
-      if (!c) continue;
-
-      lines.push(`### Comment ${i + 1}`);
-      if (includeLocation) {
-        lines.push(`*${c.file}:${c.line}*`);
-      }
-      lines.push('');
-
-      // Original comment
-      lines.push(`> **${c.author}:** ${c.text}`);
-      lines.push('');
-
-      // Context if available
-      if (includeContext && c.context) {
-        lines.push(`*Context:* "${c.context}"`);
-        lines.push('');
-      }
-
-      // Replies
-      if (c.replies.length > 0) {
-        lines.push('**Response:**');
-        lines.push('');
-        for (const reply of c.replies) {
-          lines.push(`${reply.text}`);
-        }
-      } else {
-        lines.push('**Response:**');
-        lines.push('');
-        lines.push('*[TODO: Add response]*');
-      }
-      lines.push('');
-      lines.push('---');
-      lines.push('');
-    }
-  }
-
-  // Summary stats
-  const totalComments = comments.filter(c =>
-    !c.author.toLowerCase().includes('claude') &&
-    c.author.toLowerCase() !== authorName.toLowerCase()
-  ).length;
-  const answered = comments.filter(c => c.replies.length > 0).length;
-
-  lines.push('## Summary');
-  lines.push('');
-  lines.push(`- Total comments: ${totalComments}`);
-  lines.push(`- Addressed: ${answered}`);
-  lines.push(`- Pending: ${totalComments - answered}`);
-
-  return lines.join('\n');
-}
-
-/**
- * Collect comments from multiple files
- */
-export async function collectComments(files: string[]): Promise<CommentWithReplies[]> {
-  const { readAnnotatedInput } = await import('./input.js');
-  const allComments: CommentWithReplies[] = [];
-
-  for (const file of files) {
-    if (!fs.existsSync(file)) continue;
-
-    // Route each file through the shared reader so a returned `.docx` is
-    // converted to CriticMarkup first, rather than regexing the binary ZIP.
-    const text = await readAnnotatedInput(file);
-    const comments = parseCommentsWithReplies(text, path.basename(file));
-    allComments.push(...comments);
-  }
-
-  return allComments;
-}

package/lib/restore-references.ts

@@ -1,240 +0,0 @@
-/**
- * Post-extraction reference restoration and comment parsing
- */
-
-import { readImageRegistry } from './image-registry.js';
-
-// ============================================
-// Type Definitions
-// ============================================
-
-export interface RestoreCrossrefResult {
-  text: string;
-  restored: number;
-  messages: string[];
-  restoredLabels: Set<string>;
-}
-
-export interface RestoreImagesResult {
-  text: string;
-  restored: number;
-  messages: string[];
-}
-
-// ============================================
-// Functions
-// ============================================
-
-/**
- * Parse visible comment markers from Word text
- */
-export function parseVisibleComments(text: string): Array<{ author: string; text: string; position: number }> {
-  const comments: Array<{ author: string; text: string; position: number }> = [];
-  const pattern = /\[([^\]:]+):\s*([^\]]+)\]/g;
-
-  let match;
-  while ((match = pattern.exec(text)) !== null) {
-    comments.push({
-      author: match[1].trim(),
-      text: match[2].trim(),
-      position: match.index,
-    });
-  }
-
-  return comments;
-}
-
-/**
- * Convert visible comments to CriticMarkup format
- */
-export function convertVisibleComments(text: string): string {
-  return text.replace(/\[([^\]:]+):\s*([^\]]+)\]/g, '{>>$1: $2<<}');
-}
-
-/**
- * Restore pandoc-crossref figure/table references from Word-rendered format
- */
-export function restoreCrossrefFromWord(
-  text: string,
-  projectDir: string,
-  restoredLabels: Set<string> | null = null
-): RestoreCrossrefResult {
-  const messages: string[] = [];
-  let restored = 0;
-  let result = text;
-
-  const registry = readImageRegistry(projectDir);
-
-  if (!restoredLabels) {
-    restoredLabels = new Set<string>();
-  }
-
-  // Pattern 1: [Figure]{.mark} [N]{.mark}
-  result = result.replace(/\[(Figure|Table|Fig\.?)\]\{\.mark\}\s*\[(\d+|S\d+)\]\{\.mark\}/gi, (match, type, num) => {
-    const prefix = type.toLowerCase().startsWith('tab') ? 'tbl' : 'fig';
-    if (registry) {
-      const entry = registry.byNumber?.get(`${prefix}:${num}`);
-      if (entry && entry.label) {
-        restored++;
-        return `@${prefix}:${entry.label}`;
-      }
-    }
-    restored++;
-    messages.push(`Restored ${type} ${num} (no label found, using placeholder)`);
-    return `@${prefix}:fig${num}`;
-  });
-
-  // Pattern 2: Plain "Figure N" or "Fig. N"
-  result = result.replace(/(?<!!)\b(Figure|Fig\.?|Table|Tbl\.?)\s+(\d+|S\d+)\b(?!\s*:)/gi, (match, type, num) => {
-    const prefix = type.toLowerCase().startsWith('tab') ? 'tbl' : 'fig';
-    if (registry) {
-      const entry = registry.byNumber?.get(`${prefix}:${num}`);
-      if (entry && entry.label) {
-        restored++;
-        return `@${prefix}:${entry.label}`;
-      }
-    }
-    return match;
-  });
-
-  // Pattern 3: Remove duplicate plain-text captions
-  result = result.replace(/(\!\[[^\]]+\]\([^)]+\)(?:\{[^}]*\})?)\s*\n+\s*(?:Figure|Fig\.?|Table|Tbl\.?)\s+\d+[:\.]?\s*[^\n]+/gi, '$1');
-
-  // Pattern 4: Clean up image captions that start with "Figure N: "
-  result = result.replace(/!\[(Figure|Fig\.?|Table|Tbl\.?)\s+(\d+|S\d+)[:\.]?\s*([^\]]*)\]\(([^)]+)\)(?:\{[^}]*\})?/gi,
-    (match, type, num, caption, imgPath) => {
-      const prefix = type.toLowerCase().startsWith('tab') ? 'tbl' : 'fig';
-      const labelKey = `${prefix}:${num}`;
-
-      if (registry) {
-        const entry = registry.byNumber?.get(labelKey);
-        if (entry) {
-          if (restoredLabels!.has(labelKey)) {
-            messages.push(`Skipped duplicate ${prefix}:${entry.label} (already restored)`);
-            return `![${entry.caption}](${entry.path})`;
-          }
-          restoredLabels!.add(labelKey);
-          restored++;
-          messages.push(`Restored image ${prefix}:${entry.label} from Figure ${num}`);
-          return `![${entry.caption}](${entry.path}){#${prefix}:${entry.label}}`;
-        }
-      }
-      const cleanCaption = caption.trim();
-      return `![${cleanCaption}](${imgPath})`;
-    });
-
-  return { text: result, restored, messages, restoredLabels };
-}
-
-/**
- * Restore proper markdown image syntax from Word-extracted text using image registry
- */
-export function restoreImagesFromRegistry(
-  text: string,
-  projectDir: string,
-  restoredLabels: Set<string> | null = null
-): RestoreImagesResult {
-  const messages: string[] = [];
-  let restored = 0;
-
-  const registry = readImageRegistry(projectDir);
-  if (!registry || !registry.figures || registry.figures.length === 0) {
-    return { text, restored: 0, messages: ['No image registry found'] };
-  }
-
-  if (!restoredLabels) {
-    restoredLabels = new Set<string>();
-  }
-
-  let result = text;
-
-  // Pattern 1: Caption-like text
-  const captionPatterns = [
-    /@(fig|tbl):([a-zA-Z0-9_-]+):\s*([^\n]+)/gi,
-    /^(Figure|Fig\.?)\s+(\d+|S\d+)[.:]\s*([^\n]+)/gim,
-    /\|\s*@(fig|tbl):([a-zA-Z0-9_-]+):\s*([^|]+)\s*\|/gi,
-  ];
-
-  // Fix @fig:label: caption patterns
-  result = result.replace(captionPatterns[0], (match, type, label, caption) => {
-    const key = `${type}:${label}`;
-    const entry = registry.byLabel.get(key);
-    if (entry) {
-      if (restoredLabels!.has(key)) {
-        messages.push(`Skipped duplicate ${key} (already restored)`);
-        return `![${entry.caption}](${entry.path})`;
-      }
-      restoredLabels!.add(key);
-      restored++;
-      messages.push(`Restored ${type}:${label} from registry`);
-      return `![${entry.caption}](${entry.path}){#${type}:${label}}`;
-    }
-    return match;
-  });
-
-  // Fix table-wrapped captions
-  result = result.replace(captionPatterns[2], (match, type, label, caption) => {
-    const key = `${type}:${label}`;
-    const entry = registry.byLabel.get(key);
-    if (entry) {
-      if (restoredLabels!.has(key)) {
-        messages.push(`Skipped duplicate ${key} from table wrapper`);
-        return `![${entry.caption}](${entry.path})`;
-      }
-      restoredLabels!.add(key);
-      restored++;
-      messages.push(`Restored ${type}:${label} from table wrapper`);
-      return `![${entry.caption}](${entry.path}){#${type}:${label}}`;
-    }
-    return match;
-  });
-
-  // Clean up empty table structures
-  result = result.replace(/\|\s*\|\s*\n\|:--:\|\s*\n/g, '');
-
-  // Fix "Figure N:" standalone lines
-  result = result.replace(captionPatterns[1], (match, prefix, num, caption) => {
-    const numKey = `fig:${num}`;
-    const entry = registry.byNumber.get(numKey);
-    if (entry) {
-      const labelKey = `fig:${entry.label}`;
-      if (restoredLabels!.has(labelKey)) {
-        messages.push(`Skipped duplicate Figure ${num} (already restored)`);
-        return `![${entry.caption}](${entry.path})`;
-      }
-      restoredLabels!.add(labelKey);
-      restored++;
-      messages.push(`Restored Figure ${num} by number lookup`);
-      return `![${entry.caption}](${entry.path}){#fig:${entry.label}}`;
-    }
-    return match;
-  });
-
-  // Fix generic media paths by matching caption text
-  const genericImagePattern = /!\[([^\]]*)\]\(media\/[^)]+\)/g;
-  result = result.replace(genericImagePattern, (match, caption) => {
-    if (!caption || caption.trim() === '') {
-      return match;
-    }
-
-    const captionKey = caption.slice(0, 50).toLowerCase().trim();
-    const entry = registry.byCaption.get(captionKey);
-    if (entry) {
-      const labelKey = entry.label ? `${entry.type}:${entry.label}` : null;
-      if (labelKey && restoredLabels!.has(labelKey)) {
-        messages.push(`Skipped duplicate by caption match: ${captionKey.slice(0, 30)}...`);
-        return `![${entry.caption}](${entry.path})`;
-      }
-      if (labelKey) {
-        restoredLabels!.add(labelKey);
-      }
-      restored++;
-      messages.push(`Restored image by caption match: ${captionKey.slice(0, 30)}...`);
-      const anchor = (entry.label && !restoredLabels!.has(labelKey!)) ? `{#${entry.type}:${entry.label}}` : '';
-      return `![${entry.caption}](${entry.path})${anchor}`;
-    }
-    return match;
-  });
-
-  return { text: result, restored, messages };
-}