Spaces:
Paused
Paused
| import gradio as gr | |
| from tqdm import tqdm | |
| from pipeline import ( | |
| extract_audio_from_video, | |
| transcribe_and_preprocess_audio, | |
| translation_main, | |
| tts_main, | |
| combine_audio_streams, | |
| create_combined_output, | |
| ) | |
| from pipeline import translation_hdr, translation_url, LANG | |
| async def process_video_translation( | |
| input_video, speaker, progress=gr.Progress(track_tqdm=True) | |
| ): | |
| total_stages = 6 | |
| output_video = f"{input_video.split('.')[0]}_translated.mp4" | |
| with tqdm(total=total_stages, desc="Processing video translation") as pbar: | |
| progress(0.1, desc="Extracting audio from video") | |
| pbar.update(1) | |
| output_audio_path = extract_audio_from_video(input_video) | |
| # transcribe audio | |
| pbar.set_description("Transcribing audio") | |
| pbar.update(1) | |
| sentences = transcribe_and_preprocess_audio(output_audio_path) | |
| # translate to twi | |
| pbar.set_description("Translating to Twi") | |
| khaya_translations = await translation_main( | |
| sentences, translation_url, translation_hdr, LANG | |
| ) | |
| # create output files | |
| print("Creating output files") | |
| list_of_output_chunks = [ | |
| f"translated_{i}.wav" for i in range(len(khaya_translations)) | |
| ] | |
| pbar.update(1) | |
| # convert to speech | |
| pbar.set_description("Converting to speech") | |
| await tts_main(khaya_translations, speaker, list_of_output_chunks) | |
| pbar.update(1) | |
| # combine audio streams | |
| print("Combining audio streams") | |
| pbar.set_description("Combining audio streams") | |
| output_audio = combine_audio_streams( | |
| list_of_output_chunks, "combined_audio.wav" | |
| ) | |
| pbar.update(1) | |
| pbar.set_description("Combining audio and video") | |
| create_combined_output(input_video, output_audio, output_video) | |
| pbar.update(1) | |
| print("Video translation completed") | |
| gr.Info(f"Video translation completed", duration=2) | |
| return output_video | |
| with gr.Blocks( | |
| theme=gr.themes.Soft(), | |
| title="Video Dubbing Interface", | |
| ) as demo: | |
| with gr.Row(variant="default"): | |
| with gr.Column( | |
| scale=1, | |
| min_width=0, | |
| ): | |
| gr.Image( | |
| "logo_2.jpeg", | |
| show_label=False, | |
| width=150, | |
| height=150, | |
| show_download_button=False, | |
| show_fullscreen_button=False, | |
| container=False, | |
| ) | |
| with gr.Column( | |
| scale=2, | |
| ): | |
| gr.Markdown("# Video Dubbing Interface", height=100) | |
| with gr.Column( | |
| scale=1, | |
| min_width=0, | |
| ): | |
| gr.Image( | |
| "NLPGhana_logo_2.png", | |
| show_label=False, | |
| width=50, | |
| height=150, | |
| show_download_button=False, | |
| show_fullscreen_button=False, | |
| container=False, | |
| ) | |
| # main interface components | |
| with gr.Row(): | |
| input_video = gr.Video(label="Input Video", sources=["upload"]) | |
| input_speaker = gr.Radio( | |
| label="Select Speaker", | |
| choices=["male", "female"], | |
| value="female", | |
| min_width=50, | |
| container=True, | |
| ) | |
| output_video = gr.Video(label="Processed Video") | |
| with gr.Row(): | |
| # process video translation | |
| submit = gr.Button("Process Video", scale=1) | |
| submit.click( | |
| process_video_translation, | |
| inputs=[input_video, input_speaker], | |
| outputs=output_video, | |
| ) | |
| # # Define the Gradio interface | |
| # interface = gr.Interface( | |
| # fn=process_video_translation, # Function to process the video | |
| # inputs=gr.Video(label="Input Video"), # Video file input | |
| # outputs=gr.Video(label="Processed Video"), # Video file output | |
| # title="Video Processing Interface", | |
| # description="Upload a video, and the processed video will be returned.", | |
| # theme="light", | |
| # ) | |
| # Launch the interface | |
| demo.launch(debug=True) | |