-
Notifications
You must be signed in to change notification settings - Fork 0
/
loadmodel.py
62 lines (57 loc) · 1.82 KB
/
loadmodel.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
import gensim.downloader as api
import spacy
# export LC_ALL=en_US.UTF-8
# export LANG=en_US.UTF-8
import numpy as np
import veriservice as vs
import time
# http://opendatacommons.org/licenses/pddl/
model = api.load("glove-wiki-gigaword-100") # download the model and return as object ready for use
# model.most_similar("car")
# https://dumps.wikimedia.org/legal.html
corpus = api.load('wiki-english-20171001')
nlp = spacy.load('en_core_web_sm')
vc = vs.VeriClient("localhost:10000")
it = 0
error_count = 0
code = 0
for l in corpus:
# print(l['title'])
# print(l['section_texts'])
text = ''.join(l['section_texts'])
# python -m spacy link en_core_web_md en_default
doc = nlp(text)
for sentence in doc.sents:
words = nlp(sentence.text)
# print(len(words))
features = np.array([])
for token in words:
try:
# print( model.wv[token.text.lower()] )
features = np.append(features, model.wv[token.text.lower().strip()])
# break
except:
error_count+=1
# word = token.text.lower().strip()
#if (len(word) > 0):
# print("Unknown word:_"+token.text.lower().strip()+"_")
response = vc.insert(features[:150000], l['title'], l['title'], 0)
code = response.code
while code == 1 :
time.sleep(5)
response = vc.insert(features[:150000], l['title'], l['title'], 0)
code = response.code
it+=1
print(it, code, l['title'])
if it == 5000 or code == 1:
break
count = 0
for l in corpus:
# print(l['title'])
# print(l['section_texts'])
text = ''.join(l['section_texts'])
# python -m spacy link en_core_web_md en_default
doc = nlp(text)
for sentence in doc.sents:
count+=1
print(count)