-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathfeature_data_helper.py
171 lines (154 loc) · 6.07 KB
/
feature_data_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
"""
feature_utils.py
- process metaphor features
"""
import argparse
import logging
import os
import sys
import numpy as np
import pickle
import jsonlines
logger = logging.getLogger(__name__)
feature2file = {"biasdown": "C-BiasDown.jsonlines", \
"biasup": "C-BiasUp.jsonlines", \
"biasupdown": "CCDB-BiasUpDown.jsonlines", \
"corp": "Corpus.jsonlines", \
"topic": "T.jsonlines", \
"verbnet": "VN-Raw.jsonlines", \
"wordnet": "WordNet.jsonlines"}
def readFeatureVocab(feature_fn_list):
feature_set = set([])
for feature_fn in feature_fn_list:
with jsonlines.open(feature_fn) as reader:
for obj in reader:
feature_dict = obj["x"]
for feature in feature_dict:
feature_set.add(feature)
feature_vocab = list(feature_set)
feature_vocab.sort()
print("feature vocab size: {}, examples: {}".format(len(feature_vocab),
feature_vocab[:10]))
feature2idx = {}
for (idx, feature) in enumerate(feature_vocab):
feature2idx[feature] = idx
return feature_vocab, feature2idx
def mapTokenIdToFeature(feature_fn_list, train_type, feature_vocab_fn):
if train_type == "train":
# save vocabulary of features
feature_vocab, feature2idx = readFeatureVocab(feature_fn_list)
with open(feature_vocab_fn, "wb") as handle:
pickle.dump((feature_vocab, feature2idx), handle)
elif train_type == "test":
# load vocabulary of features
with open(feature_vocab_fn, "rb") as handle:
feature_vocab, feature2idx = pickle.load(handle)
feature_dim = len(feature_vocab)
tok_id_to_feature = {}
for feature_fn in feature_fn_list:
with jsonlines.open(feature_fn) as reader:
for obj in reader:
tok_id = obj["id"]
feature_dict = obj["x"]
feature_vector = [0] * len(feature_vocab)
for feature in feature_dict:
feature_idx = feature2idx[feature]
feature_vector[feature_idx] = feature_dict[feature]
tok_id_to_feature[tok_id] = feature_vector[:]
return tok_id_to_feature, feature_dim
def mapSentToTokenId(tok_id_fn):
sent_tok_ids = []
with open(tok_id_fn, "r") as f:
for line in f:
tok_ids = line.strip().split()
sent_tok_ids.append(tok_ids[:])
return sent_tok_ids
def genFeatureFile(feature_type, train_type, tok_id_fn,
feature_folder=None, output_folder=None):
# read files for feature_type
feature_affix = feature2file[feature_type]
feature_fn_list = []
queue = [os.path.join(feature_folder, "all_pos")]
while queue != []:
fn = queue.pop(0)
if os.path.isdir(fn):
for subfolder in os.listdir(fn):
queue.append(os.path.join(fn, subfolder))
elif os.path.isfile(fn) and fn.split("/")[-1] == feature_affix:
feature_fn_list.append(fn)
else:
continue
if feature_fn_list == []:
print("No feature file for feature type: {}!".format(feature_type))
return
print("# of feature files for feature type: {}".format(feature_type))
# map tok_id to feature vector
feature_vocab_fn = os.path.join(output_folder, feature_type+".vocab.pkl")
tok_id_to_feature, feature_dim = mapTokenIdToFeature(feature_fn_list,
train_type, feature_vocab_fn)
# a list of tok_ids in a sentence
sent_tok_ids = mapSentToTokenId(tok_id_fn)
non_feature_cnt = 0
# save mask in mask.txt
mask_fout = open(os.path.join(output_folder, train_type+"_"+"mask"+".txt"), "w")
# save feature vector in [feature_type].txt
feature_fout = open(os.path.join(output_folder, train_type+"_"+feature_type+".txt"), "w")
for tok_ids in sent_tok_ids:
tok_masks = []
tok_features = []
for tok_id in tok_ids:
if tok_id in tok_id_to_feature:
tok_masks.append(1)
feature = tok_id_to_feature[tok_id]
if type(feature) == type([]):
tok_features.append(",".join([str(val) for val in feature]))
else:
#tok_features.append(str(feature))
print("Not a feature vector: {}".format(feature))
sys.exit(0)
else:
tok_masks.append(0)
tok_features.append(",".join([str(0) for i in range(feature_dim)]))
non_feature_cnt += 1
mask_fout.write(" ".join([str(val) for val in tok_masks])+"\n")
feature_fout.write(" ".join(tok_features)+"\n")
mask_fout.close()
feature_fout.close()
print("tok # without features: {}".format(non_feature_cnt))
print("Saving {} feature to {}".format(feature_type,
os.path.join(output_folder, feature_type+".txt")))
if __name__=="__main__":
parser = argparse.ArgumentParser()
parser.add_argument(
"--feature_dir",
default=None,
type=str,
required=True,
help="The input data dir. Should contain the training files for the CoNLL-2003 NER task.",
)
parser.add_argument(
"--train_type",
default="train",
type=str,
required=True,
help="Train or test"
)
parser.add_argument(
"--tok_id_fn",
default=None,
type=str,
required=True,
help="The token identifier file name"
)
parser.add_argument(
"--output_dir",
default=None,
type=str,
required=True,
help="The output directory where the model predictions and checkpoints will be written.",
)
args = parser.parse_args()
for feature_type in feature2file:
genFeatureFile(feature_type, args.train_type, args.tok_id_fn,
args.feature_dir, args.output_dir)
print("\n")