-
Notifications
You must be signed in to change notification settings - Fork 0
/
app.py
112 lines (79 loc) · 3.17 KB
/
app.py
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
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
from flask import Flask, jsonify
from flask_cors import CORS
from flask import request
from datetime import datetime
from youtube_transcript_api import YouTubeTranscriptApi
import json
from urllib.parse import parse_qs, urlparse
from transformers import pipeline
from transformers import T5ForConditionalGeneration, T5Tokenizer
from transformers import AutoModelForSeq2SeqLM, AutoTokenizer
app = Flask(__name__)
CORS(app)
@app.route('/')
def index_page():
return "Hello world"
@app.route('/time', methods=['GET'])
def get_time():
return str(datetime.now())
@app.route('/hello')
def welcome_message():
name = request.args.get('name', '')
return f'Welcome {name}'
###################################################################################
@app.route('/api/summarize', methods=['GET'])
def get_transcripts():
#Example:https://www.youtube.com/watch?v=Mus_vwhTCq0
yt_url = request.args.get('youtube_url', '')
yt_id = get_youtube_id(yt_url)[2:]
transcripts = handle_transcript(yt_id) #change
#Summarize
summary = do_NLP(transcripts) #change
#print(summary)
#Return with HTTP Status OK and handle HTTP exceptions
#return ("ugh")
return str(summary)
def handle_transcript(youtubeid):
list_of_dictionaries = YouTubeTranscriptApi.get_transcript(youtubeid)
sentence = []
for transcript in list_of_dictionaries:
#print(transcript['text'])
sentence.append(transcript['text'])
return ' '.join(map(str,sentence))
def get_youtube_id(url):
q = urlparse(url).query
return q
#Function for NLP
#accept YouTube transcript as an input parameter and return summarized transcript as output
def do_NLP(transcripts):
# using pipeline API for summarization task
#summarization = pipeline("summarization")
#summary_text = summarization(transcripts)[0]['summary_text']
# # initialize the model architecture and weights
# model = T5ForConditionalGeneration.from_pretrained("t5-base")
# # initialize the model tokenizer
# tokenizer = T5Tokenizer.from_pretrained("t5-base")
# inputs = tokenizer.encode("summarize: " + transcripts, return_tensors="pt", max_length=512, truncation=True)
# # generate the summarization output
# outputs = model.generate(
# inputs,
# max_length=150,
# min_length=40,
# length_penalty=2.0,
# num_beams=4,
# early_stopping=True)
# # just for debugging
# print(outputs)
# summary_text = tokenizer.decode(outputs[0])
# print(summary_text)
model = AutoModelForSeq2SeqLM.from_pretrained("t5-base")
tokenizer = AutoTokenizer.from_pretrained("t5-base")
inputs = tokenizer("summarize: " + transcripts, return_tensors="pt", max_length=512, truncation=True)
outputs = model.generate(
inputs["input_ids"], max_length=150, min_length=40, length_penalty=2.0, num_beams=4, early_stopping=True
)
summary_text = tokenizer.decode(outputs[0], skip_special_tokens=True)
print(summary_text)
return summary_text
if __name__ == '__main__':
app.run(debug=True)