app.py

import torchaudio
from transformers import pipeline

# Load the Hugging Face ASR model
asr_pipeline = pipeline("automatic-speech-recognition", model="facebook/wav2vec2-base-960h")

def inference(audio, sentiment_option):
    # Load and preprocess audio
    audio_tensor, _ = torchaudio.load(audio)
    
    # Perform ASR
    transcription = asr_pipeline(audio_tensor.numpy()[0], sampling_rate=audio_tensor.sampling_rate)

    # Perform sentiment analysis
    sentiment_results = analyze_sentiment(transcription[0]['sentence'])
    sentiment_output = display_sentiment_results(sentiment_results, sentiment_option)

    return "N/A", transcription[0]['sentence'], sentiment_output