add audio

baseline_utils.py

@@ -7,7 +7,7 @@ import torch
 from diffusers.utils import export_to_video
 import os
 import spaces
-from moviepy.editor import VideoFileClip, concatenate_videoclips
+from moviepy.editor import VideoFileClip, AudioFileClip, concatenate_videoclips
 
 # Utilize the Google Cloud Vision API to recognize text in the
 # input input_images (diary input_images), https://cloud.google.com/vision.
@@ -177,18 +177,25 @@ def generate_video(scene_list, writer_description, fps=24):  # Lower fps
 
     # Concatenate the generated videos into a single video
     concatenated_video_path = "videos/combined_video.mp4"
-    concatenate_videos(video_paths, concatenated_video_path)
-
+    concatenate_videos(video_paths, concatenated_video_path, audio_path="meow-meow-meow-tiktok.mp3")
     return concatenated_video_path
 
 
-def concatenate_videos(video_paths, output_path):
+def concatenate_videos(video_paths, output_path, audio_path="meow-meow-meow-tiktok.mp3"):
     # Load each video file as a VideoFileClip
     clips = [VideoFileClip(video) for video in video_paths]
 
     # Concatenate the clips
     final_clip = concatenate_videoclips(clips, method="compose")
 
-    # Write the concatenated video to a file
-    final_clip.write_videofile(output_path, codec="libx264")
+    # If an audio file is provided, load it and trim if necessary
+    if audio_path:
+        audio = AudioFileClip(audio_path)
+        # Trim the audio to match the final video's duration if audio is longer
+        if audio.duration > final_clip.duration:
+            audio = audio.subclip(0, final_clip.duration)
+        # Set the trimmed audio to the final video clip
+        final_clip = final_clip.set_audio(audio)
 
+    # Write the concatenated video to a file
+    final_clip.write_videofile(output_path, codec="libx264", audio_codec="aac")