app.py

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

def assess_pronunciation(audio_file, reference_text):
    try:
        # 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=reference_text,
            grading_system=speechsdk.PronunciationAssessmentGradingSystem.HundredMark,
            granularity=speechsdk.PronunciationAssessmentGranularity.Phoneme
        )

        # 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
            }
        elif result.reason == speechsdk.ResultReason.NoMatch:
            print("NOMATCH: Speech could not be recognized.")
            return {"Error": "Speech could not be recognized. Please try again with a clearer audio."}
        elif result.reason == speechsdk.ResultReason.Canceled:
            cancellation_details = speechsdk.CancellationDetails(result)
            print(f"CANCELED: Reason={cancellation_details.reason}")
            print(f"CANCELED: ErrorDetails={cancellation_details.error_details}")
            return {"Error": f"Speech recognition canceled: {cancellation_details.error_details}"}
    except Exception as e:
        print(f"An error occurred: {str(e)}")
        return {"Error": f"An unexpected error occurred: {str(e)}"}

# Create Gradio interface
interface = gr.Interface(
    fn=assess_pronunciation,
    inputs=[
        gr.Audio(type="filepath"),  # Audio input
        gr.Textbox(label="Reference Text", placeholder="Enter the reference text you are pronouncing")  # Reference text input
    ],
    outputs="json",
    title="Chinese Pronunciation Checker"
)

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