Update app.py

app.py

@@ -485,14 +485,14 @@ def overlay_text_on_image(image, text):
         return None
         
 def generate_combined_audio_from_story(story_text, voice='af_heart', speed=1):
-    """Generate a single audio file for all paragraphs in the story."""
+    print("Starting audio generation...")
     # Split story into paragraphs (reuse logic from generate_image_prompts)
     paragraphs = []
     current_paragraph = []
     
     for line in story_text.split('\n'):
         line = line.strip()
-        if not line:  # Empty line indicates paragraph break
+        if not line:
             if current_paragraph:
                 paragraphs.append(' '.join(current_paragraph))
                 current_paragraph = []
@@ -502,28 +502,32 @@ def generate_combined_audio_from_story(story_text, voice='af_heart', speed=1):
     if current_paragraph:
         paragraphs.append(' '.join(current_paragraph))
     
-    # Combine audio for all paragraphs
+    print(f"Found {len(paragraphs)} paragraphs")
+    
     combined_audio = []
-    for paragraph in paragraphs:
+    for i, paragraph in enumerate(paragraphs):
         if not paragraph.strip():
-            continue  # Skip empty paragraphs
+            continue
 
+        print(f"Processing paragraph {i+1}: {paragraph[:100]}...")
         generator = pipeline(
             paragraph,
             voice=voice,
             speed=speed,
-            split_pattern=r'\n+'  # Split on newlines
+            split_pattern=r'\n+'
         )
-        for _, _, audio in generator:
-            combined_audio.extend(audio)  # Append audio data
+        print(f"Generator created for paragraph {i+1}")
+        for batch_idx, metadata, audio in generator:
+            print(f"Got audio batch {batch_idx}, length: {len(audio) if audio is not None else 0}")
+            combined_audio.extend(audio)
 
-    # Convert combined audio to NumPy array and save
+    print("Converting to array...")
     combined_audio = np.array(combined_audio)
     filename = "combined_story.wav"
-    sf.write(filename, combined_audio, 24000)  # Save audio as .wav
+    sf.write(filename, combined_audio, 24000)
     clear_memory()
     return filename
-
+    
 # Helper functions
 def clean_story_output(story):
     """Clean up the generated story text."""