import spaces
import torch

import gradio as gr
from transformers import pipeline
from transformers.pipelines.audio_utils import ffmpeg_read

import tempfile
import os

MODEL_NAME = "openai/whisper-large-v3"
BATCH_SIZE = 8

device = 0 if torch.cuda.is_available() else "cpu"

pipe = pipeline(
    task="automatic-speech-recognition",
    model=MODEL_NAME,
    chunk_length_s=30,
    device=device,
)

@spaces.GPU
def transcribe(inputs, task):
    if inputs is None:
        raise gr.Error("No audio file submitted! Please upload or record an audio file before submitting your request.")
    text = pipe(inputs, batch_size=BATCH_SIZE, generate_kwargs={"task": task}, return_timestamps=True)["text"]
    return text

demo = gr.Blocks()

file_transcribe = gr.Interface(
    fn=transcribe,
    inputs=[
        gr.Audio(sources="upload", type="filepath", label="Audio file"),
        gr.Radio(["transcribe"], label="Task", value="transcribe"),
    ],
    outputs="text",
    title="Whisper Large V3: Transcribe Audio",
    allow_flagging="never",
)

with demo:
    gr.TabbedInterface([file_transcribe], ["Audio file"])

demo.queue().launch(share=True)