import gradio as gr
import speech_recognition as sr
from Levenshtein import ratio
import tempfile
import numpy as np
import soundfile as sf

def transcribe_audio(file_info):
    r = sr.Recognizer()
    with tempfile.NamedTemporaryFile(delete=True, suffix=".wav") as tmpfile:
        sf.write(tmpfile.name, data=file_info, samplerate=44100, format='WAV')
        tmpfile.seek(0)
        with sr.AudioFile(tmpfile.name) as source:
            audio_data = r.record(source)
    try:
        text = r.recognize_google(audio_data)
        return text
    except sr.UnknownValueError:
        return "Could not understand audio"
    except sr.RequestError as e:
        return f"Could not request results; {e}"

def pronunciation_correction(expected_text, file_info):
    user_spoken_text = transcribe_audio(file_info)
    similarity = ratio(expected_text.lower(), user_spoken_text.lower())
    description = f"{similarity:.2f}"

    if similarity >= 0.9:
        feedback = "Excellent pronunciation!"
    elif similarity >= 0.7:
        feedback = "Good pronunciation!"
    elif similarity >= 0.5:
        feedback = "Needs improvement."
    else:
        feedback = "Poor pronunciation, try to focus more on clarity."

    return feedback, description

def get_sentence():
    return gr.inputs.Textbox(label="Enter a Sentence")

sentence_input = get_sentence()
audio_input = gr.inputs.Audio(label="Upload Audio File")

gr.Interface(pronunciation_correction,
             inputs=[sentence_input, audio_input],
             outputs=["text", "text"]).launch()