Update utils.py

utils.py

@@ -71,25 +71,20 @@ SYSTEM PROMPT: Given the Overall Summary and All Chunks of the Text:
    - For example, if the theme is a dark jungle, use deep greens, blacks, misty blues, and dim moonlight.  
    - Ensure that all visual elements (fog, shadows, expressions) support the horror/suspense atmosphere.
 8. NEVER generate prompts that could lead to NSFW images or any explicit content. Use safe and appropriate descriptions.
-
 ### Example:
 **Summary:**  
 This text is a story of a man who ventured into a dark jungle and encountered a mysterious lion.
-
 **Chunks:**  
 1. A man enters the dark jungle, mist swirling around him.  
 2. He comes face-to-face with a majestic yet eerie lion.
-
 **Combined Context:**  
 "A man ventures into a dense, eerie jungle and unexpectedly meets a mysterious lion."
-
 **Generated Prompts:**  
 - **Chunk 1:**  
   "[style: 3D | theme: dark jungle] A lone man steps into a dense, eerie jungle at twilight. Thick mist swirls around his feet as towering, twisted trees loom overhead. Dim, bluish moonlight filters through the foliage, casting long, haunting shadows."
   
 - **Chunk 2:**  
   "[style: 3D | theme: dark jungle] In a clearing within the jungle, a majestic lion appears with an unsettling aura. Its eyes glow faintly in the dim light, and the surrounding trees seem to lean in, enhancing the mysterious tension."
-
 TASK: Here is the summary: {summary}\n\n and \n\n Total of {chunks_count} chunks, generate an Image Prompt for each chunk\n\n {chunks}
 """
     result = extractor.extract(prompt)
@@ -160,56 +155,92 @@ from moviepy.editor import AudioFileClip, ImageClip, concatenate_videoclips
 
 def generate_video(audio_file, images, segments):
     try:
-        video_dir = "./generated_videos"
-        os.makedirs(video_dir, exist_ok=True)
-
+        # Save the uploaded audio file to a temporary location
         file_extension = os.path.splitext(audio_file.name)[1]
         temp_audio_path = tempfile.NamedTemporaryFile(delete=False, suffix=f"{file_extension}")
         temp_audio_path.write(audio_file.read())
         temp_audio_path.close()
 
+        # Load the audio file using MoviePy
         audio = AudioFileClip(temp_audio_path.name)
 
+        # Define YouTube-like dimensions (16:9 aspect ratio)
         frame_width = 1280
         frame_height = 720
 
         video_clips = []
         total_segments = len(segments)
+
         for i, current_segment in enumerate(segments):
             start_time = current_segment["start"]
             end_time = current_segment["end"]
-
+            
+            # Calculate the actual duration including any gap until the next segment
             if i < total_segments - 1:
+                # If there's a next segment, extend until it starts
                 next_segment = segments[i + 1]
                 actual_end_time = next_segment["start"]
             else:
+                # For the last segment, use its end time
                 actual_end_time = end_time
-
+            
+            # Calculate total duration including any gap
             segment_duration = actual_end_time - start_time
-
+            
+            print(f"\nProcessing segment {i + 1}/{total_segments}:")
+            print(f"  Start time: {start_time}s")
+            print(f"  Base end time: {end_time}s")
+            print(f"  Actual end time: {actual_end_time}s")
+            print(f"  Total duration: {segment_duration}s")
+            print(f"  Text: '{current_segment['text']}'")
+            
+            # Ensure the image index is within bounds
             image_path = images[min(i, len(images) - 1)]
-
+            
+            # Create an ImageClip for the current segment
             image_clip = ImageClip(image_path)
+            
+            # Resize and pad the image to fit a 16:9 aspect ratio
             image_clip = image_clip.resize(height=frame_height).on_color(
                 size=(frame_width, frame_height),
-                color=(0, 0, 0),
-                pos="center"
+                color=(0, 0, 0),  # Black background
+                pos="center"      # Center the image
             )
-
-            image_clip = image_clip.set_duration(segment_duration).set_start(start_time)
+            
+            # Set the duration and start time for the clip
+            image_clip = image_clip.set_duration(segment_duration)
+            image_clip = image_clip.set_start(start_time)  # Set the start time explicitly
+            
             video_clips.append(image_clip)
 
+        # Concatenate all the image clips to form the video
+        print("Concatenating video clips...")
         video = concatenate_videoclips(video_clips, method="compose")
-        video = video.set_audio(audio)
 
-        video_filename = f"generated_video_{uuid.uuid4().hex}.mp4"
-        video_path = os.path.join(video_dir, video_filename)
+        # Add the audio to the video
+        video = video.set_audio(audio)
 
+        # Save the video to a temporary file
+        temp_dir = tempfile.gettempdir()
+        video_path = os.path.join(temp_dir, "generated_video.mp4")
+        print(f"Writing video file to {video_path}...")
         video.write_videofile(video_path, fps=30, codec="libx264", audio_codec="aac")
+
+        # Clean up the temporary audio file
         os.remove(temp_audio_path.name)
+        print("Temporary audio file removed.")
 
         return video_path
 
     except Exception as e:
         print(f"Error generating video: {e}")
-        return None
+        return None
+
+
+
+
+
+
+# Example usage:
+if __name__ == "__main__":
+    result = generate_images(["a guy in jungle", "a waterfall","greenery"])