Update app.py

app.py

@@ -3,7 +3,7 @@ from transformers import WhisperForConditionalGeneration, WhisperProcessor
 from transformers import pipeline
 import librosa
 import torch
-from spleeter.separator import Separator
+# from spleeter.separator import Separator
 from pydub import AudioSegment
 from IPython.display import Audio
 import os
@@ -11,6 +11,9 @@ import accelerate
 
 
 
+
+
+
 # preprocess and crop audio file
 def audio_preprocess(file_name = '/test1/vocals.wav'):
    # separate music and vocal
@@ -23,13 +26,17 @@ def audio_preprocess(file_name = '/test1/vocals.wav'):
    end_time = 110000  # e.g. 40 seconds, 40000
 
 
-   audio = AudioSegment.from_file(output_file) #(file_name)
+
+
+   audio = AudioSegment.from_file(file_name)
    cropped_audio = audio[start_time:end_time]
    processed_audio = cropped_audio
    # .export('cropped_vocals.wav', format='wav') # save vocal audio file
    return processed_audio
 
 
+
+
 # ASR transcription
 def asr_model(processed_audio):
    # load audio file
@@ -81,7 +88,7 @@ def senti_model(transcription):
 
 
 # main
-def main():
+def main(input_file):
 
 
    # processed_audio = audio_preprocess(input_file)
@@ -108,7 +115,7 @@ if __name__ == '__main__':
    # steamlit setup
    st.set_page_config(page_title="Sentiment Analysis on Your Cantonese Song",)
    st.header("Cantonese Song Sentiment Analyzer")
-   input_file = st.file_uploader("upload a song in mp3 format") #, type="mp3") # upload song
+   input_file = st.file_uploader("upload a song in mp3 format", type="mp3") # upload song
    if input_file is not None:
        st.write("File uploaded successfully!")
        st.write(input_file)
@@ -119,8 +126,8 @@ if __name__ == '__main__':
 
    # load song
    #input_file = os.path.isfile("test1.mp3")
-   output_file = os.path.isdir("")
+   # output_file = os.path.isdir("")
 
 
    if button_click:
-       main()
+       main(input_file=input_file)