Fix visual clutter and audio sync: enhanced layout zones, audio concatenation, proper timing

backend/compiler.py

@@ -79,15 +79,21 @@ VISUAL RULES (CRITICAL)
    - Do not let squares, shapes, or arrows clip off-screen.
    - Standard frame is [-7, 7] horizontally and [-4, 4] vertically. Keep well within this.
 
-2. No overlapping elements.
-   - All text must be positioned with next_to(), move_to(), or buff>=0.4.
-   - All shapes must have spacing.
-   - If a square represents a², show the label inside or beside — never over other shapes.
+2. MANDATORY SPACING - No overlapping elements.
+   - Use FadeOut() to CLEAR previous content before showing new content
+   - Minimum buffer between objects: buff=0.5 (NEVER less than 0.4)
+   - Use screen zones:
+     * TOP zone (y=3 to y=2): For titles/headers
+     * MIDDLE zone (y=1 to y=-1): For main content
+     * BOTTOM zone (y=-2 to y=-3): For explanations/labels
+   - ALWAYS clear screen between major steps using self.play(FadeOut(VGroup(*self.mobjects)))
+   - Position text with .to_edge(UP/DOWN) or .shift(UP*2 / DOWN*2) to avoid center crowding
 
 3. Visual accuracy FIRST.
    - Show geometry clearly.
    - Avoid rotating or stretching objects unnecessarily.
    - Avoid random effects.
+   - Use scale(0.7) for text if needed to prevent overflow
 
 ────────────────────────────
 
@@ -95,10 +101,12 @@ ANIMATION RULES
 
 ────────────────────────────
 
-1. Slow down animations & make them educational.
-   - Use 0.5–1 second durations for Create(), Write(), FadeIn().
-   - Avoid sudden transitions.
-   - Avoid instant scaling or teleporting.
+1. Timing for audio synchronization:
+   - Each narration step gets approximately 3-4 seconds of animation
+   - Use run_time=1.0 for Write() and Create()
+   - Use run_time=0.8 for FadeIn/FadeOut
+   - Add self.wait(1.5) between major steps for narration
+   - TOTAL scene duration should be 12-20 seconds
 
 2. Only use these animations:
    Create
@@ -112,6 +120,13 @@ ANIMATION RULES
 
 3. No 3D, no camera zoom, no cinematic effects, no physics.
 
+4. STRUCTURE each step clearly:
+   - Clear previous content with FadeOut
+   - Show new title/concept
+   - Display visual elements one by one
+   - Add wait time for narration
+   - Transition to next step
+
 ────────────────────────────
 
 STRUCTURE & PACING

backend/main.py

@@ -74,15 +74,30 @@ async def process_video_generation(prompt: str):
         # 2. Narrator: Generate per-step narration audio files
         steps = outline.get("steps", [])
         step_audio_paths = []
-        from narrator import generate_narration_audio
+        from narrator import generate_narration_audio, concatenate_audio_files, get_audio_duration
+        
+        total_audio_duration = 0.0
         for idx, step in enumerate(steps):
             narration = step.get("narration", "")
             if narration:
                 audio_filename = f"step_{idx+1}_narration.mp3"
                 audio_path = generate_narration_audio(narration, filename=audio_filename)
-                step_audio_paths.append(audio_path)
+                if audio_path:
+                    step_audio_paths.append(audio_path)
+                    duration = get_audio_duration(audio_path)
+                    total_audio_duration += duration
+                    print(f"✓ Step {idx+1} audio: {duration:.2f}s")
+                else:
+                    step_audio_paths.append(None)
             else:
                 step_audio_paths.append(None)
+        
+        # Combine all step audios into one file
+        combined_audio_path = None
+        if step_audio_paths and any(step_audio_paths):
+            combined_audio_path = concatenate_audio_files(step_audio_paths, output_filename="combined_narration.mp3")
+            if combined_audio_path:
+                print(f"✓ Combined audio duration: {total_audio_duration:.2f}s")
 
         # 3. Coding
         job_status["stage"] = "coding"
@@ -94,10 +109,12 @@ async def process_video_generation(prompt: str):
         job_status["message"] = "Rendering Animation Frames..."
         video_path = await render_scene(code)
 
-        # 5. Merge audio if available
-        if audio_path:
+        # 5. Merge audio with video if available
+        if combined_audio_path:
             from narrator import merge_audio_video
-            video_path = merge_audio_video(video_path, audio_path)
+            job_status["message"] = "Merging Audio with Video..."
+            video_path = merge_audio_video(video_path, combined_audio_path)
+            print(f"✓ Final video with audio: {video_path}")
 
         # Success
         relative_path = os.path.relpath(video_path, start=MEDIA_DIR).replace("\\", "/")

backend/narrator.py

@@ -1,6 +1,6 @@
 import os
 from gtts import gTTS
-from moviepy import VideoFileClip, AudioFileClip, CompositeAudioClip
+from moviepy import VideoFileClip, AudioFileClip, CompositeAudioClip, concatenate_audioclips
 import uuid
 
 # Get the directory of the current script (backend/)
@@ -11,6 +11,61 @@ AUDIO_DIR = os.path.join(MEDIA_DIR, "audio")
 # Ensure audio directory exists
 os.makedirs(AUDIO_DIR, exist_ok=True)
 
+def get_audio_duration(audio_path: str) -> float:
+    """Get the duration of an audio file in seconds"""
+    try:
+        if not audio_path or not os.path.exists(audio_path):
+            return 0.0
+        audio = AudioFileClip(audio_path)
+        duration = audio.duration
+        audio.close()
+        return duration
+    except Exception as e:
+        print(f"Error getting audio duration: {e}")
+        return 0.0
+
+def concatenate_audio_files(audio_paths: list, output_filename: str = None) -> str:
+    """
+    Concatenates multiple audio files into one.
+    Returns the path to the combined audio file.
+    """
+    try:
+        # Filter out None values and non-existent files
+        valid_paths = [p for p in audio_paths if p and os.path.exists(p)]
+        
+        if not valid_paths:
+            print("No valid audio files to concatenate")
+            return None
+            
+        if len(valid_paths) == 1:
+            # Only one file, just return it
+            return valid_paths[0]
+        
+        # Load all audio clips
+        clips = [AudioFileClip(path) for path in valid_paths]
+        
+        # Concatenate
+        final_audio = concatenate_audioclips(clips)
+        
+        # Generate output path
+        if not output_filename:
+            run_id = str(uuid.uuid4())
+            output_filename = f"combined_narration_{run_id}.mp3"
+        
+        output_path = os.path.join(BASE_DIR, output_filename)
+        final_audio.write_audiofile(output_path, logger=None)
+        
+        # Cleanup
+        for clip in clips:
+            clip.close()
+        final_audio.close()
+        
+        return output_path
+        
+    except Exception as e:
+        print(f"Error concatenating audio files: {e}")
+        return None
+
 def generate_narration_audio(text: str, filename: str = None) -> str:
     """
     Generates an MP3 audio file from the given text using gTTS.

backend/runner.py

@@ -35,7 +35,7 @@ def cleanup_old_files():
                 except Exception as e:
                     print(f"⚠️ Could not remove {old_folder}: {e}")
         
-        # Remove old audio files (keep only last 5)
+        # Remove old step audio files (keep only last 5)
         audio_files = sorted(glob.glob(os.path.join(BASE_DIR, "step_*_narration.mp3")), key=os.path.getmtime)
         for old_audio in audio_files[:-5]:
             try:
@@ -43,6 +43,15 @@ def cleanup_old_files():
                 print(f"🗑️ Cleaned up old audio file: {old_audio}")
             except Exception as e:
                 print(f"⚠️ Could not remove {old_audio}: {e}")
+        
+        # Remove old combined audio files (keep only last 2)
+        combined_files = sorted(glob.glob(os.path.join(BASE_DIR, "combined_narration*.mp3")), key=os.path.getmtime)
+        for old_combined in combined_files[:-2]:
+            try:
+                os.remove(old_combined)
+                print(f"🗑️ Cleaned up old combined audio: {old_combined}")
+            except Exception as e:
+                print(f"⚠️ Could not remove {old_combined}: {e}")
                 
     except Exception as e:
         print(f"⚠️ Cleanup error: {e}")