app.py

import gradio as gr
from scipy.io import wavfile
from wav2vec_aligen import speaker_pronunciation_assesment



def analyze_audio(audio):
# Write the processed audio to a temporary WAV file
    if audio is None:
      return 'the audio is missing'
    temp_filename = 'temp_audio.wav'
    wavfile.write(temp_filename, audio[0], audio[1])


    result = speaker_pronunciation_assesment(temp_filename)
    accuracy_score = result['pronunciation_accuracy']
    fluency_score   = result['fluency_score']
    total_score       = result['total_score']
    content_scores = result['content_scores']
  
    result_markdown = f"""|Language Aspect| Score|
    |---|---| 
    |Pronunciation Accuracy| {accuracy_score}|
    |Fluency| {fluency_score}|
    |Total Score| {total_score}|
    |Content Score| {content_scores}|
    """
    return result_markdown
  
import gradio as gr

CHOICES     = ['2861', '2862'] # '2860', 

AUDIOS = [r'assets/2861.wav', r'assets/2862.wav']

PAIRED_AUDIOS = {k: v for k, v in zip(CHOICES, AUDIOS)}

def get_paired_text(value):
    a_path = PAIRED_AUDIOS.get(value, '')
    return a_path

with gr.Blocks() as demo:
    with gr.Row():
        with gr.Column():
            with gr.Row():
                drp_down = gr.Dropdown(choices=CHOICES, scale=2)
                show_text_btn = gr.Button("Select", scale=1)
            listen_audio = gr.Audio(label='Listen to speech')
            show_text_btn.click(get_paired_text, inputs=drp_down, outputs=listen_audio)
            audio_area = gr.Audio(label='Reapet the sentence')
            analyize_audio_btn = gr.Button("Submit", scale=1)
        with gr.Column():
            capt_area = gr.Markdown(label='CAPT Scores')
            analyize_audio_btn.click(analyze_audio, inputs=audio_area, outputs=capt_area)
demo.launch()