import gradio as gr
from dataset import TranscriptDataset
from downloader import WhisperPP, YoutubeDownloader
from interpreter import WhisperInterpreter

model_size = "base"
mode = "transcribe"
write = False
download_path = "tmp/"

def dataset(url, name):    
    ds = TranscriptDataset(name)
    data = []
    #whisper_options = dict(
    #    model_size=model_size, mode=mode, write=write, number_videos=500)
    #whisperPP = WhisperPP(data,name, **whisper_options)
    #downloader = YoutubeDownloader(download_path)
    #downloader.download(url, whisperPP)
    params = dict(model_size=model_size,write=write, number_videos=500)
    overwrite = True
    ds.generate_dataset(url, download_path, overwrite, params)
    ds.upload()
  
    return "Playlist Name: " + name + "!!"

gr.Markdown("""## Create Transcription Dataset for Youtube using OpenAI Whisper """)
            
yt_input = gr.Textbox(label = 'Youtube Link')
name_input = gr.Textbox(label = 'Dataset Name')

repo_output = gr.Textbox(label = "Outcome")

iface = gr.Interface(fn=dataset, inputs=[yt_input, name_input], outputs=repo_output)
iface.launch()