app.py

import os
import openai
import time
import numpy
import gradio as gr
import soundfile as sf
from pydub import AudioSegment

from openai import OpenAI

# Load API key from an environment variable
OPENAI_SECRET_KEY = os.environ.get("OPENAI_SECRET_KEY")
client = OpenAI(api_key = OPENAI_SECRET_KEY)


####################### Choose Question Type
language_type_map = {
  "French": "FrenchTutor.txt",
  "Spanish": "SpanishTutor.txt",
}

def generate_question(language_type):

    ######################## Create Sample Question ##################
    if language_type == "French":
        with open(f"Format_Library/FrenchQuestion.txt", "r") as f:
            FrenchQGen = f.read()
        print(FrenchQGen+"/n/n")
        messages = [
            {"role": "system", "content": FrenchQGen},
            {"role": "user", "content": "Donnez moi le context et la question."}
        ]
    else:
        messages = [
            {"role": "system", "content": "Eres una máquina de preguntas que crea preguntas para estimular una respuesta de un estudiante intermedio de Espangol. La pregunta debe tener un contexto, pero el contexto puede ser cualquier cosa. Favorece los contextos cotidianos y evita los temas turísticos. Proporciona el contexto seguido de la pregunta."},
            {"role": "user", "content": "Proporcióneme el contexto y la pregunta."}
        ]
        
    question = client.chat.completions.create(model="gpt-3.5-turbo-1106", temperature=1, messages=messages)
    question_text=question.choices[0].message.content
    return question_text


def generate_feedback(recorder,language_type):
    audio_data, sr = sf.read(recorder)
    with sf.SoundFile("Audio_Files/test.mp3", mode='w', samplerate=sr, channels=len(audio_data.shape)) as file:
      file.write(audio_data)

    file_size = os.path.getsize("Audio_Files/test.mp3")
    print(f"File size after writing {file_size} bytes")
  
  ##################################Send file to Whisper for Transcription
    with open("Audio_Files/test.mp3", "rb") as audio_file:
      max_attempts = 3
      attempt = 0
      while attempt < max_attempts:
          try:
              audio_transcript = client.audio.transcriptions.create(model="whisper-1", file=audio_file)
              break
          except openai.error.APIConnectionError as e:
              print(f"Attempt {attempt + 1} failed with error: {e}")
              attempt += 1
              time.sleep(3) # wait for 3 seconds before retrying
      else:
          print("Failed to transcribe audio after multiple attempts")  
    print(audio_transcript.text+"\n\n")    
  
  
  ####################################Ask OpenAI to provide feedback on audio response

    file_name = language_type_map.get(language_type, "FrenchTutor.txt")
    with open(f"Format_Library/{file_name}", "r") as f:
      role = f.read()
  
    print(role)
        
    messages = [{"role": "system", "content": role},
               {"role": "user", "content": audio_transcript.text }
               ]
    response = client.chat.completions.create(model="gpt-3.5-turbo-1106", temperature=0, messages=messages)
    note_transcript = response.choices[0].message.content
    print(note_transcript) 
    

#######################################Create Text to Speech of Response

    speech_file_path = "Audio_Files/speech.mp3"
    response = client.audio.speech.create(
      model="tts-1",
      voice="fable",
      input=note_transcript
    )
    
    response.stream_to_file(speech_file_path)

    return note_transcript, speech_file_path
    
    
########################## Define Gradio Interface Using Blocks
with gr.Blocks() as demo:
    output_file = "Audio_Files/speech.mp3"
    
    language_type = gr.Radio(["French","Spanish"], show_label=False)
    
    txt = gr.Textbox(label="Question")
    btn = gr.Button(value="Generate Question")
    btn.click(generate_question, inputs=[language_type], outputs=[txt])
    
    recorder = gr.Audio(sources=["microphone"],type="filepath")
    txt_in = gr.Textbox(value="", label ="Voice Transcript")
    btn2 = gr.Button(value="Generate Feedback")
     
    
    output = gr.Audio(value=output_file)
    txt_2 = gr.Textbox(value="", label="Suggestions")
    btn2.click(generate_feedback, inputs=[recorder, language_type], outputs=[txt_2,output])
    
if __name__ == "__main__":
    demo.launch(share=False, debug=True)