app.py

import gradio as gr
from transformers import pipeline
import torch

def load_models():
    """Load and verify models with error checking"""
    try:
        # Check CUDA availability
        device = "cuda" if torch.cuda.is_available() else "cpu"
        print(f"Using device: {device}")
        
        # Load a smaller Whisper model for better compatibility
        transcriber = pipeline(
            "automatic-speech-recognition",
            model="openai/whisper-tiny",
            device=device
        )
        
        # Load a simpler sentiment model
        sentiment = pipeline(
            "sentiment-analysis",
            model="distilbert-base-uncased-finetuned-sst-2-english",
            device=device
        )
        
        return transcriber, sentiment
    except Exception as e:
        print(f"Error loading models: {str(e)}")
        return None, None

def analyze_audio(audio_path):
    """
    Analyze audio tone with robust error handling
    """
    if audio_path is None:
        return "Please provide an audio input", "No audio detected"
        
    try:
        # Load models
        transcriber, sentiment = load_models()
        if transcriber is None or sentiment is None:
            return "Error loading models", "Model initialization failed"
            
        # Transcribe with error checking
        try:
            result = transcriber(audio_path)
            text = result["text"]
            if not text.strip():
                return "No speech detected", "Empty transcription"
        except Exception as e:
            return f"Transcription error: {str(e)}", "Failed to process audio"
            
        # Analyze sentiment
        try:
            sentiment_result = sentiment(text)[0]
            tone = sentiment_result["label"]
            confidence = f"{sentiment_result['score']:.2%}"
            return tone, confidence
        except Exception as e:
            return f"Sentiment analysis error: {str(e)}", "Analysis failed"
            
    except Exception as e:
        return f"Unexpected error: {str(e)}", "Process failed"

# Create interface with simplified components
interface = gr.Interface(
    fn=analyze_audio,
    inputs=gr.Audio(
        sources=["microphone", "upload"],
        type="filepath",
    ),
    outputs=[
        gr.Textbox(label="Tone"),
        gr.Textbox(label="Confidence Level")
    ],
    title="Simple Speech Tone Analyzer",
    description="Record or upload audio to analyze its tone. The analysis may take a few moments.",
    examples=None,
    cache_examples=False,
    theme=gr.themes.Base(),
)

# Launch with specific parameters for better stability
if __name__ == "__main__":
    interface.launch(
        debug=True,
        server_name="0.0.0.0",
        server_port=7860,
        share=True
    )