# Module Imports
from pytube import YouTube
import whisper
import gradio as gr

model = whisper.load_model("base.en")

# Functions
def transcribe(file):
    options = dict(task="transcribe", best_of=5)
    text = model.transcribe(file, **options)["text"]
    return text.strip()

def translate(file):
    options = dict(task="translate", best_of=5)
    text = model.transcribe(file, **options)["text"]
    return text.strip()

def get_filename(file_obj):
    return file_obj.orig_name

def inference(link):
  yt = YouTube(link)
  path = yt.streams.filter(only_audio=True)[0].download(filename="audio.mp4")
  options = whisper.DecodingOptions(without_timestamps=True)
  results = model.transcribe(path)
  return results['text']

def populate_metadata(link):
  yt = YouTube(link)
  return yt.thumbnail_url, yt.title

def transcribe_file(file):
    options = dict(task="transcribe", best_of=5)
    file = get_filename(file)
    text = model.transcribe(file, **options)["text"]
    return text.strip()

# Gradio Blocks
demo = gr.Blocks()
with demo:
    gr.Markdown(""" <p style="text-align: center;"> Speech To Text Grammar Checker </p>""")
    with gr.Tabs():
        with gr.TabItem("Voice Record"):
            with gr.Row():
                audio = gr.Audio(show_label=False,source="microphone",type="filepath")
                text_output1 = gr.Textbox(label="Transcription", placeholder="Text Output")
            transcribe_button1 = gr.Button("Transcribe")
            CG_button1 = gr.Button("Check Grammar (Work In Progress)")
        with gr.TabItem("Upload File"):
            with gr.Row():
                file_upload = gr.File()
                text_output2 = gr.Textbox(label="Transcription", placeholder="Text Output")
            transcribe_button2 = gr.Button("Transcribe")
            CG_button3 = gr.Button("Check Grammar (Work In Progress)")
        with gr.TabItem("Youtube Link"):
            with gr.Box():
                link = gr.Textbox(label="YouTube Link")
                with gr.Row().style(mobile_collapse=False, equal_height=True):
                  title = gr.Label(label="Video Title", placeholder="Title")
                  img = gr.Image(label="Thumbnail")
                text_link_output = gr.Textbox(label="Transcription", placeholder="Text Output",lines=5)
                with gr.Row().style(mobile_collapse=False, equal_height=True): 
                  transcribe_button3 = gr.Button("Transcribe")
                  CG_button3 = gr.Button("Check Grammar (Work In Progress)")
        with gr.TabItem("About"):
          with gr.Row():
            gr.Markdown("""
            <p style="text-align: center;"> Thesis System presented by  <br/> <br/>
            • <b>Jomari A. Buenaobra</b> <br/>
            • <b>Christian G. Eslit</b> <br/>
            • <b>Daniel L. Espinola</b> <br/>
            • <b>Jhon Vincent A. Gupo</b> <br/>
            • <b>Ryan M. Ibay</b> <br/> <br/>
            In partial fulfillment of the requirements for the degree <br/>   
            Bachelor of Science in Computer Science Specialized in Intelligent Systems <br/>
            Laguna State Polytechnic University - Los Baños Campus . <br/>  <br/> 
            We would also like to thank our fellow adviser and subject specialist for their guidance in making this idea a reality. <br/>
            • <b>Crisanto F. Gulay</b> - Adviser <br/>
            • <b>Jonardo R. Asor</b> - Subject Specialist <br/>
            </p>
            """)

            link.change(populate_metadata, inputs=[link], outputs=[img, title])
            transcribe_button1.click(transcribe, inputs=audio, outputs=text_output1)
            transcribe_button2.click(transcribe_file, inputs=file_upload, outputs=text_output2)
            transcribe_button3.click(inference, inputs=link, outputs=text_link_output)
            
demo.launch()