File size: 3,132 Bytes
bc736f8
97fd8dc
bc736f8
 
b640e62
49d7053
bc736f8
49d7053
a63bdf7
bc736f8
 
49d7053
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
bc736f8
 
3a36dd7
896fad4
 
 
 
9817c7c
 
5188c19
bc736f8
 
5188c19
fe55d4b
 
5188c19
 
 
 
 
 
 
fe55d4b
 
5188c19
 
fe55d4b
5188c19
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
import whisper
from pytube import YouTube
import gradio as gr
import os
import re
import logging

logging.basicConfig(level=logging.INFO)
model = whisper.load_model("base")

def get_text(url):
    try:
        if url != '':
            output_text_transcribe = ''
    
        yt = YouTube(url)
        #video_length = yt.length
        #if video_length < 5400:
        video = yt.streams.filter(only_audio=True).first()
        out_file=video.download(output_path=".")
        base, ext = os.path.splitext(out_file)
        new_file = base+'.mp3'
        os.rename(out_file, new_file)
        a = new_file
    
        logging.error("Size of audio file: %s", str(a.length))
        result = model.transcribe(a)
        return result['text'].strip()
        #else:
        #    return "Videos for transcription on this space are limited to 1.5 hours. Sorry about this limit but some joker thought they could stop this tool from working by transcribing many extremely long videos. Please visit https://steve.digital to contact me about this space."
    finally:
        raise gr.Error("Exception: There was a problem transcribing the audio.")

def get_summary(article):
    #try:
    first_sentences = ' '.join(re.split(r'(?<=[.:;])\s', article)[:5])
    b = summarizer(first_sentences, min_length = 20, max_length = 120, do_sample = False)
    b = b[0]['summary_text'].replace(' .', '.').strip()
    return b
    #finally:
    #raise gr.Error("Exception: There was a problem summarizing the transcript.")

  
with gr.Blocks() as demo:
    gr.Markdown("<h1><center>Free Fast YouTube URL Video-to-Text using <a href=https://openai.com/blog/whisper/ target=_blank>OpenAI's Whisper</a> Model</center></h1>")
    #gr.Markdown("<center>Enter the link of any YouTube video to generate a text transcript of the video and then create a summary of the video transcript.</center>")
    gr.Markdown("<center>Enter the link of any YouTube video to generate a text transcript of the video.</center>")
    gr.Markdown("<center><b>'Whisper is a neural net that approaches human level robustness and accuracy on English speech recognition.'</b></center>")
    gr.Markdown("<center>Transcription takes 5-10 seconds per minute of the video (bad audio/hard accents slow it down a bit). #patience<br />If you have time while waiting, check out my <a href=https://www.artificial-intelligence.blog target=_blank>AI blog</a> (opens in new tab).</center>")
    
    input_text_url = gr.Textbox(placeholder='Youtube video URL', label='URL')
    result_button_transcribe = gr.Button('1. Transcribe')
    output_text_transcribe = gr.Textbox(placeholder='Transcript of the YouTube video.', label='Transcript')
    
    #result_button_summary = gr.Button('2. Create Summary')
    #output_text_summary = gr.Textbox(placeholder='Summary of the YouTube video transcript.', label='Summary')
    
    result_button_transcribe.click(get_text, inputs = input_text_url, outputs = output_text_transcribe)
    #result_button_summary.click(get_summary, inputs = output_text_transcribe, outputs = output_text_summary)

demo.queue(default_enabled = True).launch(debug = True)