app.py

import gradio as gr
import openai
import shutil
import tempfile
import os

vocabolario_LOUD = os.environ.get("VOCABOLARIO_LOUD")
open_ai_key = os.environ.get("OPEN_AI_KEY")
openai.api_key = open_ai_key

def transcribe(audio):

    # the audio file has no extension. We mae a .wav copy:
    with tempfile.NamedTemporaryFile(suffix=".wav", delete=True) as temp_audio_file:
        
        shutil.copy(audio, temp_audio_file.name)

        audio_filepath = temp_audio_file.name
        audio_file= open(audio_filepath, "rb")
    
    transcript = openai.Audio.transcribe("whisper-1", audio_file, prompt=vocabolario_LOUD, language="it")

    return transcript["text"]

app = gr.Interface(fn=transcribe, inputs=gr.Microphone(type="filepath"), outputs="text", title="Demo di trascrizione con vocabolario LOUD",description="Demo implementante l'API di Whisper personalizzato con vocabolario LOUD", theme="dark-peach")

app.launch()