-
Notifications
You must be signed in to change notification settings - Fork 6
/
format.py
202 lines (177 loc) 路 7.48 KB
/
format.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
189
190
191
192
193
194
195
196
197
198
199
200
201
202
import os
import json
import argparse
from bs4 import BeautifulSoup
from pathlib import Path
from tqdm import tqdm
from collections import OrderedDict
from udpipe import Model
from ace_parser import Parser
from prettytable import PrettyTable
model_map = {
'en': 'udpipe/english-ewt-ud-2.5-191206.udpipe',
'zh': 'udpipe/chinese-gsd-ud-2.5-191206.udpipe',
'ar': 'udpipe/arabic-padt-ud-2.5-191206.udpipe'
}
lang_name = {
'en': 'English',
'ar': 'Arabic',
'zh': 'Chinese'
}
def check_duplicate_files(train, dev, test):
num_train = len(set(train))
num_dev = len(set(dev))
num_test = len(set(test))
if num_train != len(train):
print('Warning: duplicate file found in train split')
if num_dev != len(dev):
print('Warning: duplicate file found in dev split')
if num_test != len(test):
print('Warning: duplicate file found in test split')
total_files = len(train) + len(dev) + len(test)
if total_files != num_train + num_dev + num_test:
print('Warning: duplicate file found in train/dev/test split')
def get_filenames(args):
"""Check if file exists, return a file list for train/dev/test split."""
def read_files(split):
filepath = os.path.join(args.filelist, 'ace.%s.%s.txt' % (args.lang, split))
filelist = []
not_found = []
with open(filepath) as f:
for line in f:
filename = '.'.join(line.strip().split('.')[:-2])
sgm_file = os.path.join(args.data, '{}.sgm'.format(filename))
xml_file = os.path.join(args.data, '{}.apf.xml'.format(filename))
if os.path.exists(sgm_file) and os.path.exists(xml_file):
filelist.append(filename)
else:
not_found.append(filename)
print('[{:>5}] out of {:>3} files, {:>2} were not found.'.format(
split.upper(), len(filelist) + len(not_found), len(not_found)))
if not_found:
print('Files that were not found - ', not_found)
return filelist
train = read_files('train')
dev = read_files('dev')
test = read_files('test')
check_duplicate_files(train, dev, test)
return train, dev, test
def parse_sgm(model, sgm_path):
sgm_file = os.path.join(args.data, '{}.sgm'.format(sgm_path))
with open(sgm_file, 'r') as f:
soup = BeautifulSoup(f.read(), features='html.parser')
sgm_text = soup.text
sentences = model.tokenize(sgm_text, 'ranges')
total_words = 0
for s in sentences:
total_words += len(s.words)
model.tag(s)
model.parse(s)
conllu = model.write(sentences, "conllu")
return conllu, len(sentences), total_words
def parse_xml(xml_path):
parser = Parser(os.path.join(args.data, xml_path))
return OrderedDict([
('entities', parser.entity_mentions),
('events', parser.event_mentions),
('relations', parser.relation_mentions)
])
def process_data(opt, model, filenames, split):
outdir = os.path.join(opt.output, split)
Path(outdir).mkdir(parents=True, exist_ok=True)
total_entities = 0
total_events = 0
total_relations = 0
total_event_arguments = 0
total_sentences = 0
total_words = 0
for filename in tqdm(filenames, total=len(filenames)):
outfile = os.path.split(filename)[-1]
conllu_file = os.path.join(outdir, '{}.conllu'.format(outfile))
with open(conllu_file, 'w') as fw:
conllu, num_sent, num_words = parse_sgm(model, filename)
total_sentences += num_sent
total_words += num_words
fw.write(conllu)
json_file = os.path.join(outdir, '{}.v1.json'.format(outfile))
with open(json_file, 'w') as fw:
jsonobj = parse_xml(filename)
total_entities += len(jsonobj['entities'])
total_events += len(jsonobj['events'])
total_relations += len(jsonobj['relations'])
total_event_arguments += sum([len(em['arguments']) for em in jsonobj['events']])
json.dump(jsonobj, fw)
return {
'total_files': len(filenames),
'total_sentences': total_sentences,
'total_words': total_words,
'total_entities': total_entities,
'total_events': total_events,
'total_relations': total_relations,
'total_event_arguments': total_event_arguments
}
def main(args):
args.data = os.path.join(args.data, lang_name[args.lang])
args.output = os.path.join(args.output, lang_name[args.lang])
Path(args.output).mkdir(parents=True, exist_ok=True)
train_files, dev_files, test_files = get_filenames(args)
model = Model(model_map[args.lang])
train_stat = process_data(args, model, train_files, 'train')
dev_stat = process_data(args, model, dev_files, 'dev')
test_stat = process_data(args, model, test_files, 'test')
table = PrettyTable()
table.field_names = ["Attribute", "Train", "Dev", "Test", "Total"]
table.align["Attribute"] = "l"
table.align["Train"] = "r"
table.align["Dev"] = "r"
table.align["Test"] = "r"
table.align["Total"] = "r"
table.add_row([
'#Documents', train_stat['total_files'],
dev_stat['total_files'], test_stat['total_files'],
train_stat['total_files'] + dev_stat['total_files'] + test_stat['total_files']
])
table.add_row([
'#Sentences', train_stat['total_sentences'],
dev_stat['total_sentences'], test_stat['total_sentences'],
train_stat['total_sentences'] + dev_stat['total_sentences'] + test_stat['total_sentences']
])
table.add_row([
'#Words', train_stat['total_words'],
dev_stat['total_words'], test_stat['total_words'],
train_stat['total_words'] + dev_stat['total_words'] + test_stat['total_words']
])
table.add_row([
'Entity Mentions', train_stat['total_entities'],
dev_stat['total_entities'], test_stat['total_entities'],
train_stat['total_entities'] + dev_stat['total_entities'] + test_stat['total_entities']
])
table.add_row([
'Relation Mentions', train_stat['total_relations'],
dev_stat['total_relations'], test_stat['total_relations'],
train_stat['total_relations'] + dev_stat['total_relations'] + test_stat['total_relations']
])
table.add_row([
'Event Mentions', train_stat['total_events'],
dev_stat['total_events'], test_stat['total_events'],
train_stat['total_events'] + dev_stat['total_events'] + test_stat['total_events']
])
table.add_row([
'Event Arguments', train_stat['total_event_arguments'],
dev_stat['total_event_arguments'], test_stat['total_event_arguments'],
train_stat['total_event_arguments'] + dev_stat['total_event_arguments'] + test_stat['total_event_arguments']
])
print(table)
if __name__ == '__main__':
parser = argparse.ArgumentParser()
parser.add_argument('--data', type=str, default='./ace_2005/data/',
help="Path of ACE2005 data")
parser.add_argument('--filelist', type=str, default='./filelist/',
help="List of files for train/dev/test split")
parser.add_argument('--lang', type=str, help="Name of the language", default='en',
choices=['en', 'ar', 'zh'])
parser.add_argument('--output', type=str, default='./processed-data/',
help="Path of the output directory")
args = parser.parse_args()
print('\n' + '*' * 20 + lang_name[args.lang] + '*' * 20 + '\n')
main(args)