app.py

import gradio as gr
import openai
import time
import os
from pydub import AudioSegment
from pydub.silence import split_on_silence

openai_client = openai.OpenAI()
OPENAI_TRANSCRIPTION_MODEL = 'whisper-1'
whisper = openai_client.audio.transcriptions.create
WHISPER_TEMPERATURE = 0.0

def transcribe_audio(audio_file):
    audio = AudioSegment.from_file(audio_file)
    chunks = split_on_silence(audio, min_silence_len=500, silence_thresh=-40)
    
    transcription = ""
    
    for i, chunk in enumerate(chunks):
        chunk_file = f"CHUNKS/chunk_{i}.mp3"
        chunk.export(chunk_file, format="mp3")
        response = whisper(model = OPENAI_TRANSCRIPTION_MODEL,
                         file = open(chunk_file, 'rb'), 
                         language = 'es',
                         temperature = WHISPER_TEMPERATURE)        # to do, explore temperature
        print(dir(response))
        transcription += response.text + " "
        # os.remove(chunk_file)   # concatenate at the end
        
        yield transcription

iface = gr.Interface(
    fn=transcribe_audio,
    inputs = gr.Audio(sources=["microphone"], type="filepath"),
    outputs="text",
    live=True,  # Enable live streaming
    title="💊💊💊 Lucero-CETRAM-CORFO Transcripcion con Whisper 💊💊💊",

    description="record using your microphone to get a transcription using OpenAI's Whisper API."
)

iface.launch()