app.py

import gradio as gr
import azure.cognitiveservices.speech as speechsdk

def assess_pronunciation(audio_file):
    # Configure Azure Speech Service
    speech_key = "12afe22c558a4f8d8bd28d6a67cdb9b0"
    service_region = "westus"
    speech_config = speechsdk.SpeechConfig(subscription=speech_key, region=service_region)
    
    # Set up the audio configuration
    audio_config = speechsdk.audio.AudioConfig(filename=audio_file)
    
    # Create pronunciation assessment config
    pronunciation_config = speechsdk.PronunciationAssessmentConfig(
        reference_text="你好",
        grading_system=speechsdk.PronunciationAssessmentGradingSystem.HundredMark,
        granularity=speechsdk.PronunciationAssessmentGranularity.Phoneme
    )
    pronunciation_config.enable_prosody_assessment()

    # Create the recognizer
    recognizer = speechsdk.SpeechRecognizer(speech_config=speech_config, audio_config=audio_config)
    pronunciation_config.apply_to(recognizer)

    # Recognize speech and assess pronunciation
    result = recognizer.recognize_once()
    
    # Debug information
    print(f"Recognition result reason: {result.reason}")
    if result.reason == speechsdk.ResultReason.RecognizedSpeech:
        pronunciation_result = speechsdk.PronunciationAssessmentResult(result)
        
        # Extract and format the results
        accuracy_score = pronunciation_result.accuracy_score
        fluency_score = pronunciation_result.fluency_score
        completeness_score = pronunciation_result.completeness_score
        prosody_score = pronunciation_result.prosody_score

        return {
            "Accuracy": accuracy_score,
            "Fluency": fluency_score,
            "Completeness": completeness_score,
            "Prosody": prosody_score
        }
    else:
        print(f"Recognition result text: {result.text}")
        print(f"Recognition error details: {result.error_details}")
        return {"Error": "There was an error. Speech could not be recognized."}

# Create Gradio interface
interface = gr.Interface(
    fn=assess_pronunciation,
    inputs=gr.Audio(type="filepath"),  # Corrected input
    outputs="json",
    title="Chinese Pronunciation Checker"
)

if __name__ == "__main__":
    interface.launch()