Update app.py

app.py

@@ -657,9 +657,24 @@ Return only valid JSON. Do not include explanation, markdown, or comments.
             for i, item in enumerate(podcast_json['podcast']):
                 try:
                     add_log(f"🎵 Processing line {i+1}/{total_lines}: Speaker {item['speaker']}")
-                    
+                    clean_line = item['line']
+
+                    # 🔧 Sanitize malformed lines
+                    if not isinstance(clean_line, str) or len(clean_line.strip()) == 0 or clean_line.strip().startswith('"') or "{" in clean_line:
+                    add_log(f"⚠️ Malformed line detected for speaker {item['speaker']}: {repr(clean_line[:80])}")
+                    # Try to recover from JSON-like noise
+                    candidates = re.findall(r'\"line\"\s*:\s*\"([^\"]+)\"', clean_line)
+                    if candidates:
+                        clean_line = candidates[0]
+                        add_log(f"✅ Recovered line: {clean_line}")
+                    else:
+                        # Fallback: strip bad characters
+                        clean_line = re.sub(r'[^A-Za-z0-9\s.,!?;:\-\'"]+', '', clean_line)
+                        add_log(f"🛠️ Cleaned fallback line: {clean_line}")
+
                     audio_file = await self.tts_generate(
-                        item['line'], 
+                        clean_line,
+                        #item['line'], 
                         item['speaker'], 
                         speaker1, 
                         speaker2