Add debug endpoint to find control chars in project data

routes/project_routes.py

@@ -488,3 +488,76 @@ def delete_thumbnail(project_id):
     cursor.execute('UPDATE projects SET thumbnail_data = NULL WHERE id = ?', (project_id,))
     db.commit()
     return jsonify({'success': True})
+
+# ============================================
+# DEBUG: Identify corrupt data
+# ============================================
+
+@project_bp.route('/api/projects/<int:project_id>/debug', methods=['GET'])
+@login_required
+def debug_project(project_id):
+    """
+    Scan a project for control characters and report which fields are dirty.
+    Visit: /api/projects/<id>/debug after logging in.
+    """
+    db = get_db()
+    cursor = db.cursor()
+    
+    cursor.execute('SELECT id, name FROM projects WHERE id = ?', (project_id,))
+    project = cursor.fetchone()
+    if not project:
+        return jsonify({'error': 'Project not found'}), 404
+    
+    def find_bad_chars(s):
+        """Return list of (position, char_code) for any control char found."""
+        if not s or not isinstance(s, str):
+            return []
+        bad = []
+        for i, ch in enumerate(s):
+            code = ord(ch)
+            # Allow \t (9), \n (10), \r (13). Anything else <32 or 127 is bad.
+            if (code < 32 and code not in (9, 10, 13)) or code == 127:
+                bad.append({'pos': i, 'code': code, 'hex': f'0x{code:02x}'})
+                if len(bad) >= 5:  # cap at 5 per field
+                    break
+        return bad
+    
+    report = {
+        'project_id': project['id'],
+        'project_name': project['name'],
+        'issues': []
+    }
+    
+    cursor.execute('SELECT * FROM chapters WHERE project_id = ? ORDER BY chapter_number', (project_id,))
+    chapters = cursor.fetchall()
+    
+    for chapter in chapters:
+        ch_num = chapter['chapter_number']
+        
+        for field in ('title', 'voice'):
+            bad = find_bad_chars(chapter[field])
+            if bad:
+                report['issues'].append({
+                    'where': f'chapter {ch_num} -> {field}',
+                    'bad_chars': bad,
+                    'sample': repr((chapter[field] or '')[:80])
+                })
+        
+        cursor.execute('SELECT * FROM markdown_blocks WHERE chapter_id = ? ORDER BY block_order', (chapter['id'],))
+        blocks = cursor.fetchall()
+        
+        for block in blocks:
+            b_order = block['block_order']
+            for field in ('block_type', 'content', 'tts_text', 'audio_data', 'audio_format', 'transcription'):
+                bad = find_bad_chars(block[field])
+                if bad:
+                    val = block[field] or ''
+                    report['issues'].append({
+                        'where': f'chapter {ch_num}, block {b_order} -> {field}',
+                        'field_length': len(val),
+                        'bad_chars': bad,
+                        'sample_around_first_bad': repr(val[max(0, bad[0]['pos']-20):bad[0]['pos']+20])
+                    })
+    
+    report['total_issues'] = len(report['issues'])
+    return jsonify(report)