add moviepy and video 2 audio conversion

Dockerfile

@@ -9,7 +9,7 @@ RUN apt-get update && apt-get install -y \
     && rm -rf /var/lib/apt/lists/*
 
 # Install pip packages
-RUN pip install gradio==3.50.2
+RUN pip install gradio==3.50.2 moviepy==1.0.3
 
 # Set environment variables for matplotlib and fontconfig
 ENV MPLCONFIGDIR=/tmp/matplotlib

app.py

@@ -3,6 +3,17 @@ import gradio as gr
 import tempfile
 from pathlib import Path
 import os
+from moviepy.editor import VideoFileClip
+
+def convert_video_to_audio(video_input):
+    video_clip = VideoFileClip(video_input)
+    audio_clip = video_clip.audio
+    audio_clip_filepath = os.path.normpath(f"{video_input.split('.')[0]}.m4a")
+    audio_clip.write_audiofile(audio_clip_filepath, codec='aac')
+    audio_clip.close()
+    video_clip.close()
+    return audio_clip_filepath
+
 
 # Check permissions and environment setup
 print("Checking directory and file permissions...")
@@ -62,8 +73,12 @@ class Predictor:
         return final_video_path
 
 # Gradio user interface
-def gradio_predict(audio, bg_color, fg_alpha, bars_color, bar_count, bar_width, caption_text):
+def gradio_predict(input_obj, input_obj_type, bg_color, fg_alpha, bars_color, bar_count, bar_width, caption_text):
     predictor = Predictor()
+    if input_obj_type =="video":
+        audio = convert_video_to_audio(video_input=input_obj)
+    else:
+        audio = input_obj
     result = predictor.predict(
         audio=audio,
         bg_color=bg_color,
@@ -75,23 +90,44 @@ def gradio_predict(audio, bg_color, fg_alpha, bars_color, bar_count, bar_width,
     )
     return result
 
-# Launch Gradio interface
-interface = gr.Interface(
-    fn=gradio_predict,
-    inputs=[
-        gr.Audio(source="upload", type="filepath", label="Audio File"),
-        gr.Textbox(value="#000000", label="Background Color"),
-        gr.Slider(0, 1, value=0.75, label="Foreground Opacity"),
-        gr.ColorPicker(value="#ffffff", label="Bars Color"),
-        gr.Slider(10, 500, value=100, step=1, label="Number of Bars"),
-        gr.Slider(0, 1, value=0.4, step=0.1, label="Bar Width"),
-        gr.Textbox(value="", label="Caption Text")
-    ],
-    outputs=gr.Video(label="Waveform Video"),
-    live=False
-    allow_flagging="never"
-)
+def main():
+    with gr.Blocks(title="Speech Enhancement", delete_cache=(86400, 86400), theme=gr.themes.Ocean()) as demo:
+        with gr.Tabs(selected="video") as tabs:
+            with gr.Tab("Video", id="video"):
+                gr.Interface(
+                    fn=gradio_predict,
+                    inputs=[
+                        gr.Video(source="upload", label="Video File"),
+                        gr.Radio(choices=["video"], value="video", label="File Type"),
+                        gr.Textbox(value="#000000", label="Background Color"),
+                        gr.Slider(0, 1, value=0.75, label="Foreground Opacity"),
+                        gr.ColorPicker(value="#ffffff", label="Bars Color"),
+                        gr.Slider(10, 500, value=100, step=1, label="Number of Bars"),
+                        gr.Slider(0, 1, value=0.4, step=0.1, label="Bar Width"),
+                        gr.Textbox(value="", label="Caption Text")
+                    ],
+                    outputs=gr.Video(label="Waveform Video"),
+                    live=False,
+                    allow_flagging="never"
+                )
+            with gr.Tab("Audio", id="audio"):
+                gr.Interface(
+                    fn=gradio_predict,
+                    inputs=[
+                        gr.Audio(source="upload", type="filepath", label="Audio File"),
+                        gr.Radio(choices=["audio"], value="audio", label="File Type"),
+                        gr.Textbox(value="#000000", label="Background Color"),
+                        gr.Slider(0, 1, value=0.75, label="Foreground Opacity"),
+                        gr.ColorPicker(value="#ffffff", label="Bars Color"),
+                        gr.Slider(10, 500, value=100, step=1, label="Number of Bars"),
+                        gr.Slider(0, 1, value=0.4, step=0.1, label="Bar Width"),
+                        gr.Textbox(value="", label="Caption Text")
+                    ],
+                    outputs=gr.Video(label="Waveform Video"),
+                    live=False,
+                    allow_flagging="never"
+                )
+        demo.launch()
 
 if __name__ == "__main__":
-    print("Starting Gradio interface...")
-    interface.launch(server_name="0.0.0.0", server_port=7860)
+    main()