Update app.py

app.py

@@ -42,54 +42,7 @@ def clean_file_name(file_path):
 
     return clean_file_path
 
-def get_audio_file(uploaded_file):
-    global base_path
-    # ,device
-    device = "cuda" if torch.cuda.is_available() else "cpu"
-    # Detect the file type (audio/video)
-    mime_type, _ = mimetypes.guess_type(uploaded_file)
-    # Create the folder path to store audio files
-    audio_folder = f"{base_path}/subtitle_audio"
-    os.makedirs(audio_folder, exist_ok=True)
-    # Initialize variable for the audio file path
-    audio_file_path = ""
-    if mime_type and mime_type.startswith('audio'):
-        # If it's an audio file, save it as is
-        audio_file_path = os.path.join(audio_folder, os.path.basename(uploaded_file))
-        audio_file_path=clean_file_name(audio_file_path)
-        shutil.copy(uploaded_file, audio_file_path)  # Move file to audio folder
-
-    elif mime_type and mime_type.startswith('video'):
-        # If it's a video file, extract the audio
-        audio_file_name = os.path.splitext(os.path.basename(uploaded_file))[0] + ".mp3"
-        audio_file_path = os.path.join(audio_folder, audio_file_name)
-        audio_file_path=clean_file_name(audio_file_path)
-
-        # Extract the file extension from the uploaded file
-        file_extension = os.path.splitext(uploaded_file)[1]  # Includes the dot, e.g., '.mp4'
-
-        # Generate a random UUID and create a new file name with the same extension
-        random_uuid = uuid.uuid4().hex[:6]
-        new_file_name = random_uuid + file_extension
-
-        # Set the new file path in the subtitle_audio folder
-        new_file_path = os.path.join(audio_folder, new_file_name)
-
-        # Copy the original video file to the new location with the new name
-        shutil.copy(uploaded_file, new_file_path)
-        if device=="cuda":
-          command = f"ffmpeg -hwaccel cuda -i {new_file_path} {audio_file_path} -y"
-        else:
-          command = f"ffmpeg -i {new_file_path} {audio_file_path} -y"
-
-        subprocess.run(command, shell=True)
-        if os.path.exists(new_file_path):
-          os.remove(new_file_path)
-    # Return the saved audio file path
-    audio = AudioSegment.from_file(audio_file_path)
-    # Get the duration in seconds
-    duration_seconds = len(audio) / 1000.0  # pydub measures duration in milliseconds
-    return audio_file_path,duration_seconds
+
 
 def format_segments(segments):
     saved_segments = list(segments)
@@ -242,8 +195,7 @@ def whisper_subtitle(uploaded_file,Source_Language,max_words_per_subtitle=8):
       device = "cpu"
       compute_type = "int8"
   faster_whisper_model = WhisperModel("deepdml/faster-whisper-large-v3-turbo-ct2",device=device, compute_type=compute_type)
-  audio_path,audio_duration=get_audio_file(uploaded_file)
-
+  audio_path=uploaded_file
   if Source_Language=="Automatic":
       segments,d = faster_whisper_model.transcribe(audio_path, word_timestamps=True)
       lang_code=d.language
@@ -267,11 +219,11 @@ def whisper_subtitle(uploaded_file,Source_Language,max_words_per_subtitle=8):
   original_srt_name=clean_file_name(save_name)
   original_txt_name=original_srt_name.replace(".srt",".txt")
   word_level_srt_name=original_srt_name.replace(".srt","_word_level.srt")
-  default_srt_name=original_srt_name.replace(".srt","_default.srt")
+  customize_srt_name=original_srt_name.replace(".srt","_customize.srt")
     
-  generate_srt_from_sentences(sentence_timestamp, srt_path=default_srt_name)
+  generate_srt_from_sentences(sentence_timestamp, srt_path=original_srt_name)
   word_level_srt(words_timestamp, srt_path=word_level_srt_name)
-  write_subtitles_to_file(word_segments, filename=original_srt_name)
+  write_subtitles_to_file(word_segments, filename=customize_srt_name)
   with open(original_txt_name, 'w', encoding='utf-8') as f1:
     f1.write(text)
   return default_srt_name,original_srt_name,word_level_srt_name,original_txt_name
@@ -304,9 +256,9 @@ source_lang_list.extend(available_language)
 @click.option("--debug", is_flag=True, default=False, help="Enable debug mode.")
 @click.option("--share", is_flag=True, default=False, help="Enable sharing of the interface.")
 def main(debug, share):
-    description = """
-    **Note**: For large video files, upload audio instead. FFmpeg video-to-audio conversion may take a long time.
-    """
+    # description = """
+    # **Note**: For large video files, upload audio instead. FFmpeg video-to-audio conversion may take a long time.
+    # """
     # Define Gradio inputs and outputs
     gradio_inputs = [
         gr.File(label="Upload Audio or Video File"),
@@ -322,9 +274,9 @@ def main(debug, share):
     ]
 
     # Create Gradio interface
-    demo = gr.Interface(fn=subtitle_maker, inputs=gradio_inputs, outputs=gradio_outputs, title="Auto Subtitle Generator Using Whisper-Large-V3-Turbo-Ct2",description=description)
+    demo = gr.Interface(fn=subtitle_maker, inputs=gradio_inputs, outputs=gradio_outputs, title="Auto Subtitle Generator Using Whisper-Large-V3-Turbo-Ct2")#,description=description)
 
     # Launch Gradio with command-line options
     demo.queue().launch(debug=debug, share=share)
 if __name__ == "__main__":
-    main()
+    main()