import locale
from pathlib import Path
# from whisper import Whisper
# from pyannote.audio import Pipeline
# import torch
import static_ffmpeg
import file_operations

def prep() -> None:
    locale.getpreferredencoding = lambda: "UTF-8"
    # download and add ffmpeg to env
    static_ffmpeg.add_paths()


def audiofile(url: str, input_path: Path) -> Path:
    file = file_operations.download_from_url(url, input_path)
    file_wav = file_operations.convert_to_wav(file, input_path)
    file.unlink()
    return file_wav

#
# def diarization(access_token: str | None) -> Pipeline:
#     pipeline = Pipeline.from_pretrained(
#         "pyannote/speaker-diarization", use_auth_token=access_token
#     )
#     device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
#     return pipeline.to(device)
#
#
# def whisper() -> Whisper:
#     # LOAD MODEL INTO VRAM
#     device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
#     return whisper.load_model("large", device=device)