Update app.py

app.py

@@ -103,24 +103,19 @@ def text_to_audio(text, api_key, voice):
     return audio_filename, audio_bytes_io
     
 def merge_audio_video(video_filename, audio_filename, output_filename):
-    print("Merging audio and video ...")
-    # Load the video file
     video_clip = VideoFileClip(video_filename)
-    # Load the audio file
     audio_clip = AudioFileClip(audio_filename)
+    
+    if audio_clip.duration > video_clip.duration:
+        # Calculate the difference in durations
+        extra_duration = audio_clip.duration - video_clip.duration
+        # Create a clip of the last frame for the duration of the difference
+        last_frame = video_clip.subclip(video_clip.duration - 1).to_ImageClip(duration=extra_duration)
+        # Concatenate the last frame clip to the end of the original video clip
+        video_clip = concatenate_videoclips([video_clip, last_frame])
 
-    # Determine the shortest duration between audio and video
-    min_duration = min(video_clip.duration, audio_clip.duration)
-
-    # Set the audio of the video clip as the audio file, trimming to the shortest duration
-    video_clip = video_clip.subclip(0, min_duration)
-    audio_clip = audio_clip.subclip(0, min_duration)
     final_clip = video_clip.set_audio(audio_clip)
-
-    # Write the result to a file
     final_clip.write_videofile(output_filename, codec='libx264', audio_codec="aac")
-
-    # Close the clips
     video_clip.close()
     audio_clip.close()
 
@@ -178,7 +173,7 @@ def main():
 
     # Define unique prompt templates for each script type, including the dynamic content for "Product Tutorial"
     script_templates = {
-        'Product Tutorial': f"Generate a short voiceover  that is approximately {selected_duration} seconds long. Lets roleplay you are a script generator for tutorials. Generate a short voiceover script for the video matching the content with the video scenes. Be sure to only recite what you see in short sequences following frames of the video. You are allowed to comment on UI and UX even faces. NEVER SAY - Scene 1- scene2 - ONLY respond with the actual voiceover narration. Never add Timestamps to your response! You look at the website and create tutorial style content!! The app is called Pixio. The style should be fun and engaging. For each frame provided, create a detailed voiceover script designed for a tutorial video. Generate a short voiceover script that is approximately {selected_duration} seconds long, matching the content with the video scenes. " + 
+        'Product Tutorial': f"Generate a short voiceover  that is approximately {selected_duration} seconds long.Your script should be limited to {selected_duration} seconds only! DO NOT exceed {selected_duration} seconds. Lets roleplay you are a script generator for tutorials. Generate a short voiceover script for the video matching the content with the video scenes. Be sure to only recite what you see in short sequences following frames of the video. You are allowed to comment on UI and UX even faces. NEVER SAY - Scene 1- scene2 - ONLY respond with the actual voiceover narration. Never add Timestamps to your response! You look at the website and create tutorial style content!! The app is called Pixio. The style should be fun and engaging. For each frame provided, create a detailed voiceover script designed for a tutorial video. Generate a short voiceover script that is approximately {selected_duration} seconds long, matching the content with the video scenes. " + 
         "The app is called Pixio. The style should be fun and engaging. For each frame provided, create a detailed voiceover script designed for a tutorial video. " +
         "Never say 'Female 2' or 'VoiceOver' in responses. You output a script to be spoken! - Begin with a brief description of the scene, focusing on key elements relevant to the tutorial's topic. " +
         "- Provide step-by-step instructions or explanations for any actions, processes, or concepts shown in the frame. Use clear and concise language suitable for educational content. " +