Spaces:
Sleeping
Sleeping
Create app.py
Browse files
app.py
ADDED
|
@@ -0,0 +1,63 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
from youtube_transcript_api import YouTubeTranscriptApi
|
| 2 |
+
from urllib.parse import urlparse, parse_qs
|
| 3 |
+
import torch
|
| 4 |
+
import gradio as gr
|
| 5 |
+
# Use a pipeline as a high-level helper
|
| 6 |
+
from transformers import pipeline
|
| 7 |
+
|
| 8 |
+
# model_path = r"../Models/models--sshleifer--distilbart-cnn-12-6/snapshots/a4f8f3ea906ed274767e9906dbaede7531d660ff"
|
| 9 |
+
# text_summary = pipeline ( task= "summarization", model= model_path, torch_dtype=torch.bfloat16)
|
| 10 |
+
|
| 11 |
+
text_summary = pipeline ( task= "summarization", model="sshleifer/distilbart-cnn-12-6", torch_dtype=torch.bfloat16)
|
| 12 |
+
|
| 13 |
+
def summary(input):
|
| 14 |
+
output = text_summary(input)
|
| 15 |
+
return output[0]['summary_text']
|
| 16 |
+
|
| 17 |
+
|
| 18 |
+
def get_video_id(youtube_url):
|
| 19 |
+
# Extract the video ID from the YouTube URL
|
| 20 |
+
query = urlparse(youtube_url)
|
| 21 |
+
if query.hostname == 'youtu.be':
|
| 22 |
+
return query.path[1:]
|
| 23 |
+
if query.hostname in ('www.youtube.com', 'youtube.com'):
|
| 24 |
+
if query.path == '/watch':
|
| 25 |
+
return parse_qs(query.query)['v'][0]
|
| 26 |
+
if query.path[:7] == '/embed/':
|
| 27 |
+
return query.path.split('/')[2]
|
| 28 |
+
if query.path[:3] == '/v/':
|
| 29 |
+
return query.path.split('/')[2]
|
| 30 |
+
return None
|
| 31 |
+
|
| 32 |
+
def get_transcript(youtube_url):
|
| 33 |
+
video_id = get_video_id(youtube_url)
|
| 34 |
+
if not video_id:
|
| 35 |
+
return "Invalid YouTube URL."
|
| 36 |
+
|
| 37 |
+
try:
|
| 38 |
+
transcript = YouTubeTranscriptApi.get_transcript(video_id)
|
| 39 |
+
transcript_text = "\n".join([entry['text'] for entry in transcript])
|
| 40 |
+
summary_text = summary(transcript_text)
|
| 41 |
+
return summary_text
|
| 42 |
+
except Exception as e:
|
| 43 |
+
return f"Error fetching transcript: {e}"
|
| 44 |
+
|
| 45 |
+
# Example usage
|
| 46 |
+
# url = input("Enter YouTube URL: ")
|
| 47 |
+
# transcript = get_transcript(url)
|
| 48 |
+
# print("\n--- Transcript ---\n")
|
| 49 |
+
# print(transcript)
|
| 50 |
+
|
| 51 |
+
|
| 52 |
+
gr.close_all()
|
| 53 |
+
|
| 54 |
+
demo = gr.Interface(fn=get_transcript,
|
| 55 |
+
inputs=[gr.Textbox(label="Input Youtube URL to summarize",lines=1)],
|
| 56 |
+
outputs=[gr.Textbox(label="summarized text",lines=4)],
|
| 57 |
+
title="@GenAILearniverse Project 2: Youtube Script Summarizer",
|
| 58 |
+
description="THIS APPLICATION WILL BE USED TO SUMMARIZE Youtube Video Script")
|
| 59 |
+
|
| 60 |
+
|
| 61 |
+
print("🚀 App is launching...")
|
| 62 |
+
demo.launch(share=True)
|
| 63 |
+
|