import streamlit as st
from tempfile import NamedTemporaryFile

st.write("## Whisper Speech to Text 🎧")

import torch
from transformers import AutoModelForSpeechSeq2Seq, AutoProcessor, pipeline


@st.cache_resource
def load_model():
    device = "cuda:0" if torch.cuda.is_available() else "cpu"
    torch_dtype = torch.float16 if torch.cuda.is_available() else torch.float32

    model_id = "openai/whisper-large-v3-turbo"
    model = AutoModelForSpeechSeq2Seq.from_pretrained(
        model_id, torch_dtype=torch_dtype, low_cpu_mem_usage=True, use_safetensors=True
    )
    model.to(device)

    processor = AutoProcessor.from_pretrained(model_id)

    pipe = pipeline(
        "automatic-speech-recognition",
        model=model,
        tokenizer=processor.tokenizer,
        feature_extractor=processor.feature_extractor,
        torch_dtype=torch_dtype,
        device=device,
    )

    return pipe

@st.cache_data
def transcribe(_pipe, sample):
    result = _pipe(sample, return_timestamps=True)
    return result["text"]


audio = st.file_uploader("Upload your Audio Here")

model = load_model()
if st.button("Transcribe"):
    if audio is not None:
        st.write("Transcribing ... ")

        with NamedTemporaryFile(suffix="mp3") as temp:
            temp.write(audio.getvalue())
            temp.seek(0)
            transcript = transcribe(model, temp.name)
            transcript                              # MAGIC!