Create app.py

app.py

@@ -0,0 +1,53 @@
+import gradio as gr
+from moviepy.editor import VideoFileClip
+from transformers import pipeline
+import os
+
+# Initialize the Whisper model
+whisper_model = pipeline("automatic-speech-recognition", model="openai/whisper-large")
+
+def convert_video_to_wav(video_path):
+    # Extract audio from video using moviepy and save as WAV
+    video_clip = VideoFileClip(video_path)
+    audio = video_clip.audio
+    wav_file = "temp_audio.wav"
+    audio.write_audiofile(wav_file, codec='pcm_s16le')  # Write as WAV format
+    return wav_file
+
+def convert_audio_to_srt(wav_file):
+    # Transcribe the audio using the Whisper model
+    transcription = whisper_model(wav_file)
+    
+    # Save the transcription to an SRT file with simple formatting
+    srt_file = "transcription.srt"
+    with open(srt_file, "w", encoding="utf-8") as f:
+        for i, segment in enumerate(transcription['text'].split('.')):
+            f.write(f"{i+1}\n")  # Subtitle index
+            f.write(f"00:00:{i*2:02d},000 --> 00:00:{i*2+2:02d},000\n")  # Timestamp (basic)
+            f.write(f"{segment.strip()}\n\n")  # Transcription text
+    
+    # Clean up temp audio file
+    os.remove(wav_file)
+    
+    return srt_file
+
+def process_video(video):
+    # Save the uploaded video file to a temporary location
+    video_path = video.name
+    
+    # Process the video to extract audio and convert to srt
+    wav_file = convert_video_to_wav(video_path)  # Convert video to WAV
+    srt_file = convert_audio_to_srt(wav_file)  # Convert WAV to SRT
+    
+    return srt_file  # Return the path of the generated SRT file
+
+# Gradio Interface
+interface = gr.Interface(
+    fn=process_video, 
+    inputs=gr.File(label="Upload video file", file_types=['mp4', 'avi', 'mkv']),  # Video file input
+    outputs=gr.File(label="Download SRT File"),  # Output the SRT file for download
+    title="Video to SRT Subtitle Generator",
+    description="Upload a video file (e.g., .mp4), and the app will generate a subtitle file (SRT format) using Whisper model."
+)
+
+interface.launch()