Update app.py

app.py

@@ -1,67 +1,66 @@
 import gradio as gr
 import cv2
 import numpy as np
-from pydub import AudioSegment
-from moviepy.editor import VideoFileClip, AudioFileClip
 import tempfile
-import os
+import soundfile as sf
+from moviepy.editor import VideoFileClip, concatenate_videoclips,AudioFileClip
 
-def image_to_video(image, audio):
+def image_to_video(image, audio, fps=30):
     """
-    Converts an image and an audio file into a video.
+    Converts an image (NumPy array) and audio (tuple of sample rate and audio data) 
+    into a video file.
 
     Parameters:
-    - image: Uploaded image file.
-    - audio: Uploaded audio file.
+    - image: Input image as a NumPy array.
+    - audio: Input audio as a tuple of (sample rate, audio data as a NumPy array).
+    - fps: Frames per second of the output video.
+    
+    Returns:
+    - Path to the generated video file.
     """
-    # Create temporary paths for the files
-    image_path = tempfile.mktemp(suffix=".png")
-    audio_path = tempfile.mktemp(suffix=".mp3")
-    video_path = tempfile.mktemp(suffix=".mp4")
+    # Create a temporary video file
+    video_path = tempfile.mktemp('.mp4')
     
-    # Save uploaded files to temporary paths
-    image.save(image_path)
-    audio.save(audio_path)
+    # Save the image to a temporary file
+    image_path = tempfile.mktemp('.png')
+    cv2.imwrite(image_path, image)
     
-    # Load the image
-    img = cv2.imread(image_path)
-
-    # Get image dimensions
-    height, width, layers = img.shape
-
-    # Define the codec and create VideoWriter object
+    # Create VideoWriter to write frames
+    height, width, _ = image.shape
     fourcc = cv2.VideoWriter_fourcc(*'mp4v')
-    video = cv2.VideoWriter(video_path, fourcc, 30, (width, height))
-
-    audio_segment = AudioSegment.from_file(audio_path)
-    duration_sec = len(audio_segment) / 1000.0  # Convert duration from ms to seconds
-
-    # Calculate the number of frames needed
-    num_frames = int(duration_sec * 30)  # Assuming 30 fps
-
+    video = cv2.VideoWriter(video_path, fourcc, fps, (width, height))
+    
+    # Determine the duration from the audio sample count and sample rate
+    sample_rate, audio_data = audio
+    duration_sec = len(audio_data) / sample_rate
+    
     # Write the image to the video file for the required number of frames
+    num_frames = int(duration_sec * fps)
     for _ in range(num_frames):
-        video.write(img)
-
-    # Release the video writer
+        video.write(cv2.imread(image_path))
     video.release()
-
-    # Add audio to the video
+    
+    # Save the audio to a temporary file
+    audio_path = tempfile.mktemp('.wav')
+    sf.write(audio_path, audio_data.T, sample_rate)  # Transpose if necessary for multi-channel audio
+    
+    # Combine the video and audio
     video_clip = VideoFileClip(video_path)
     audio_clip = AudioFileClip(audio_path)
     final_clip = video_clip.set_audio(audio_clip)
-    final_clip.write_videofile(video_path, codec="libx264", audio_codec="aac")
-
+    final_clip.write_videofile(video_path, codec="libx264", audio_codec="aac")    
     return video_path
 
-# Define the Gradio interface
+# Gradio interface
 iface = gr.Interface(
     fn=image_to_video,
-    inputs=[gr.Image(label="Upload Image"), gr.Audio(label="Upload Audio")],
-    outputs=gr.Video(label="Output Video"),
-    title="Image to Video Converter",
-    description="Converts an image and an audio file into a video."
+    inputs=[
+        gr.Image(type="numpy"),
+        gr.Audio(type="numpy", label="Audio (Sample Rate, Audio Data)")
+    ],
+    outputs=gr.Video("video"),
+    title="Image and Audio to Video Converter",
+    description="Upload an image and audio to convert them into a video."
 )
 
-# Launch the app
-iface.launch()
+iface.launch(debug=True,share=True)