-
Notifications
You must be signed in to change notification settings - Fork 0
/
helper.py
executable file
·188 lines (148 loc) · 5.92 KB
/
helper.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
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
from pymongo import MongoClient
import numpy as np, sys, unicodedata, requests, os, random, pdb, requests, json
from random import randint
from pprint import pprint
import logging, logging.config, itertools, pathlib
from sklearn.metrics import precision_recall_fscore_support
np.set_printoptions(precision=4)
# c_dosa = MongoClient('mongodb://10.24.28.104:27017/')
# db_word2vec = c_dosa['word2vec']['google_news_300']
def checkFile(filename):
return pathlib.Path(filename).is_file()
def getWord2vec(wrd_list):
dim = 300
embeds = np.zeros((len(wrd_list), dim), np.float32)
embed_map = {}
res = db_word2vec.find({"_id": {"$in": wrd_list}})
for ele in res:
embed_map[ele['_id']] = ele['vec']
count = 0
for wrd in wrd_list:
if wrd in embed_map: embeds[count, :] = np.float32(embed_map[wrd])
else: embeds[count, :] = np.random.randn(dim)
count += 1
return embeds
# def getPhr2vec(phr_list, embed_type):
# dim = int(embed_type.split('_')[1])
# db_glove = c_dosa['glove'][embed_type]
# wrd_list = []
# embeds = np.zeros((len(phr_list), dim), np.float32)
# embed_map = {}
# for phr in phr_list:
# wrd_list += phr.split('_')
# wrd_list = list(set(wrd_list))
# res = db_glove.find({"_id": {"$in": wrd_list}})
# for ele in res:
# embed_map[ele['_id']] = ele['vec']
# count = 0
# for phr in phr_list:
# wrds = phr.split('_')
# vec = np.zeros((dim,), np.float32)
# for wrd in wrds:
# if wrd in embed_map: vec += np.float32(embed_map[wrd])
# else: vec += np.float32(np.random.randn(dim))
# vec = vec / len(wrds)
# embeds[count, :] = vec
# return embeds
# def signal(message):
# requests.post( 'http://10.24.28.210:9999/jobComplete', data=message)
# def len_key(tp):
# return len(tp[1])
def set_gpu(gpus):
os.environ["CUDA_DEVICE_ORDER"] = "PCI_BUS_ID"
os.environ["CUDA_VISIBLE_DEVICES"] = gpus
def shape(tensor):
s = tensor.get_shape()
return tuple([s[i].value for i in range(0, len(s))])
# coreNLP_url = [ 'http://10.24.28.106:9006/', 'http://10.24.28.106:9007/', 'http://10.24.28.106:9008/', 'http://10.24.28.106:9009/', 'http://10.24.28.106:9010/', 'http://10.24.28.106:9011/',
# 'http://10.24.28.106:9012/', 'http://10.24.28.106:9013/', 'http://10.24.28.106:9014/', 'http://10.24.28.106:9015/', 'http://10.24.28.106:9016/']
# def callnlpServer(text):
# params = {
# 'properties': '{"annotators":"tokenize"}',
# 'outputFormat': 'json'
# }
# res = requests.post( coreNLP_url[randint(0, len(coreNLP_url)-1)],
# params=params, data=text,
# headers={'Content-type': 'text/plain'})
# if res.status_code == 200: return res.json()
# else: print("CoreNLP Error, status code:{}".format(res.status_codet))
def debug_nn(res_list,feed_dict):
import tensorflow as tf
# ph = np.zeros(self.p.batch_size, dtype = np.int32)
# pt = np.zeros(self.p.batch_size, dtype = np.int32)
# r = np.zeros(self.p.batch_size, dtype = np.int32)
# nh = np.zeros(self.p.batch_size, dtype = np.int32)
# nt = np.zeros(self.p.batch_size, dtype = np.int32)
# ph_addr = ph.__array_interface__['data'][0]
# pt_addr = pt.__array_interface__['data'][0]
# r_addr = r.__array_interface__['data'][0]
# nh_addr = nh.__array_interface__['data'][0]
# nt_addr = nt.__array_interface__['data'][0]
# lib.init(self.max_ent, self.max_rel, 483142, self.p.batch_size)
# lib.getBatch(ph_addr, pt_addr, r_addr, nh_addr, nt_addr, batch_size,1)
# feed_dict = {pos_head : ph,
# pos_tail : pt,
# rel : r,
# neg_head : nh,
# neg_tail : nt}
# facts = open('../data/train.txt','wb')
# kg_adj_in, kg_adj_out = self.get_adj(facts, self.max_ent, self.max_rel) # max_et + 1(DCT)
# for lbl in range(self.max_rel):
# feed_dict[self.kg_adj_mat_in[i][lbl]] = tf.SparseTensorValue( indices = np.array([kg_adj_in[i][lbl].row, kg_adj_in[i][lbl].col]).T,
# values = kg_adj_in[i][lbl].data,
# dense_shape = kg_adj_in[i][lbl].shape)
# feed_dict[self.kg_adj_mat_out[i][lbl]] = tf.SparseTensorValue( indices = np.array([kg_adj_out[i][lbl].row, kg_adj_out[i][lbl].col]).T,
# values = kg_adj_out[i][lbl].data,
# dense_shape = kg_adj_out[i][lbl].shape)
# if dtype != 'train':
# feed_dict[self.dropout] = 1.0
# feed_dict[self.rec_dropout] = 1.0
config = tf.ConfigProto()
config.gpu_options.allow_growth=True
sess = tf.Session(config=config)
sess.run(tf.global_variables_initializer())
summ_writer = tf.summary.FileWriter("tf_board/debug_nn", sess.graph)
res = sess.run(res_list, feed_dict = feed_dict)
pdb.set_trace()
def stanford_tokenize(text):
res = callnlpServer(text)
toks = [ele['word'] for ele in res['tokens']]
return toks
def is_number(s):
try:
float(s)
return True
except ValueError:
pass
try:
import unicodedata
unicodedata.numeric(s)
return True
except (TypeError, ValueError):
pass
return False
def is_int(s):
try:
int(s)
return True
except ValueError:
return False
def get_logger(name):
config_dict = json.load(open('/scratchd/home/shikhar/gcn/config/log_config.json'))
config_dict['handlers']['file_handler']['filename'] = '/scratchd/home/shikhar/gcn/main/log/' + name.replace('/', '-')
logging.config.dictConfig(config_dict)
logger = logging.getLogger(name)
std_out_format = '%(asctime)s - [%(levelname)s] - %(message)s'
consoleHandler = logging.StreamHandler(sys.stdout)
consoleHandler.setFormatter(logging.Formatter(std_out_format))
logger.addHandler(consoleHandler)
return logger
def partition(lst, n):
division = len(lst) / float(n)
return [ lst[int(round(division * i)): int(round(division * (i + 1)))] for i in range(n) ]
def getChunks(inp_list, chunk_size):
return [inp_list[x:x+chunk_size] for x in range(0, len(inp_list), chunk_size)]
def mergeList(list_of_list):
return list(itertools.chain.from_iterable(list_of_list))
# doc = 'Delhi is the capital of India. Mumbai is not the capital of India.'
# pprint(callnlpServer(doc))