get_sample.py
4.28 KB
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
import argparse
import os
import random
import shutil
def main():
args = parse_arguments()
if not args.input or not args.output:
print('Error: Input and output must be selected!')
get_sample(args.input, args.output)
def parse_arguments():
parser = argparse.ArgumentParser(description='Get conllup corpora sample.')
required_arguments = parser.add_argument_group('required arguments')
required_arguments.add_argument('-o', '--output', help='output directory', required=True)
required_arguments.add_argument('-i', '--input', help='corpora root directory', required=True)
return parser.parse_args()
def get_sample(root_directory, out_corpora_directory):
docs_by_domains = sort_docs_by_domains(root_directory)
for domain in docs_by_domains:
draw_sample_and_copy(docs_by_domains[domain]['abstract'], docs_by_domains[domain]['full'],
out_corpora_directory)
def sort_docs_by_domains(root_directory):
docs_by_domains = {}
for root, dirs, files in os.walk(root_directory):
for filename in files:
print(filename)
if filename.endswith('.conllup') or filename.endswith('.conllu'):
src = os.path.join(root, filename)
domain = get_domain(src)
sentences_count = get_sentences_count(src)
if sentences_count > 3:
if domain not in docs_by_domains:
docs_by_domains[domain] = {'full': [], 'abstract': []}
if is_full_text(src):
docs_by_domains[domain]['full'].append({'path': src, 'sentences': sentences_count})
else:
docs_by_domains[domain]['abstract'].append({'path': src, 'sentences': sentences_count})
return docs_by_domains
def get_domain(filepath):
with open(filepath, 'r') as conllup_file:
for line in conllup_file:
line = line.strip()
if is_segment(line):
continue
elif is_metadata(line):
name, value = get_metadata(line)
if name == 'domain':
return value.strip()
return ''
def is_segment(line):
if line and line[0].isdigit():
return True
return False
def is_metadata(line):
if line.startswith('#'):
return True
return False
def get_metadata(line):
name_value_pair = line.split('=', 1)
name = name_value_pair[0].lstrip('#').strip()
value = name_value_pair[1].strip()
return name, value
def is_full_text(filepath):
with open(filepath, 'r') as conllup_file:
for line in conllup_file:
line = line.strip()
if is_segment(line):
continue
elif is_metadata(line):
name, value = get_metadata(line)
if name == 'content_type':
if value.strip() == 'full_text':
return True
else:
return False
return False
def get_sentences_count(filepath):
sentences = 0
with open(filepath, 'r') as conllup_file:
for line in conllup_file:
line = line.strip()
if is_segment(line):
pass
elif is_metadata(line):
name, value = get_metadata(line)
if name == 'sent_id':
sentences += 1
return sentences
def draw_sample_and_copy(abstracts, full_texts, out_corpora_directory):
number_of_draws = 0
sentences_count = 0
sentences_cap = 6500
random.shuffle(abstracts)
random.shuffle(full_texts)
for abstract in abstracts:
if sentences_count >= sentences_cap:
break
filename = os.path.basename(abstract['path'])
dst = os.path.join(out_corpora_directory, filename)
shutil.copyfile(abstract['path'], dst)
number_of_draws += 1
sentences_count += abstract['sentences']
if number_of_draws % 8 == 0:
full_text = full_texts[number_of_draws // 8]
filename = os.path.basename(full_text['path'])
dst = os.path.join(out_corpora_directory, filename)
shutil.copyfile(full_text['path'], dst)
sentences_count += full_text['sentences']
if __name__ == '__main__':
main()