-
Notifications
You must be signed in to change notification settings - Fork 2
/
5_create_mturk_files.py
136 lines (110 loc) · 6.23 KB
/
5_create_mturk_files.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
import csv
from itertools import chain
import random
import json
from utilwebisadb import set_csv_field_size, get_ids_in_range, search_start_end_index_in_sentence
def create_relation_template():
for i in range(0, 20):
print(
"""<tr><td style="text-align: left;">${r_?}</td><td><input type="radio" name="${id_?}" value="yes" required="required"/></td><td><input type="radio" name="${id_?}" value="uncertain" required="required"/></td><td><input type="radio" name="${id_?}" value="no" required="required"/></td></tr>"""
.replace("?", str(i)))
def create_sentence_template():
for i in range(0, 10):
print(
"""<li>
<p>${sentence_?}</p>
<div class="form-horizontal">
<div class="form-group">
<div class="col-sm-2"><label class="control-label">${label_?_instance}</label></div>
<div class="col-sm-5"><input type="url" class="form-control" name="${id_?}_instance_page" placeholder="English Wikipedia page URL or 'not:possible'" required></div>
<div class="col-sm-5"><input type="url" class="form-control" name="${id_?}_instance_category" placeholder="English Wikipedia category URL or 'not:possible'" required></div>
</div>
<div class="form-group">
<div class="col-sm-2"><label class="control-label">${label_?_class}</label></div>
<div class="col-sm-5"><input type="url" class="form-control" name="${id_?}_class_page" placeholder="English Wikipedia page URL or 'not:possible'" required></div>
<div class="col-sm-5"><input type="url" class="form-control" name="${id_?}_class_category" placeholder="English Wikipedia category URL or 'not:possible'" required></div>
</div>
</div>
</li>""".replace("?", str(i)))
def replace_csv_in_template(template_path, data, out):
template = ""
with open(template_path, 'r') as template_file:
template = template_file.read()
with open(data) as csvfile:
reader = csv.DictReader(csvfile)
for row in reader:
for name in reader.fieldnames:
template = template.replace("${"+ name + "}", row[name])
print("replace {} with {}".format("${"+ name + "}", row[name]))
break #only first line to test
with open(out, "w") as out:
out.write(template)
def write_mturk_relation_csv(csv_file, mturk_csv_file, k, has_headers=False):
with open(mturk_csv_file, "w", newline='') as out :
writer = csv.writer(out)
writer.writerow(chain.from_iterable(("r_" + str(i), "id_" + str(i)) for i in range(0, k)))
relation_id_list = []
with open(csv_file) as f:
reader = csv.reader(f)
if has_headers:
next(reader, None) # skip the headers
for row in reader:
if len(relation_id_list) == (k*2):
writer.writerow(relation_id_list)
relation_id_list.clear()
relation_id_list.append(row[1] + " is a " + row[2]) # relation string
relation_id_list.append(row[0]) # id
if len(relation_id_list) == (k * 2): # write last line if it has enough content
writer.writerow(relation_id_list)
def write_mturk_sentence_csv(csv_file, mturk_sentence_csv_file, k, has_headers=False):
#load sentences info
with open('sentences_info.json', 'r') as fp:
data = json.load(fp)
max_length_sentence = data['max_length_sentence']
max_length_pld = data['max_length_pld']
length_one_entry = max_length_sentence + max_length_pld
min_id = data['min_id']
max_id = data['max_id']
with open(mturk_sentence_csv_file, "w", newline='') as out, open('sentences_sorted_equidistant.bin', 'rb') as sentences_sorted_equidistant_file:
writer = csv.writer(out)
writer.writerow(chain.from_iterable(("sentence_{}".format(i), "label_{}_instance".format(i), "id_{}".format(i), "label_{}_class".format(i)) for i in range(0, k)))
entity_list = []
with open(csv_file) as f:
reader = csv.reader(f)
if has_headers:
next(reader, None) # skip the headers
for row in reader:
if len(entity_list) == (k*4):
writer.writerow(entity_list)
entity_list.clear()
prov_id_list = get_ids_in_range(row[14], min_id, max_id)
random.shuffle(prov_id_list)
sentence = ""
for id in prov_id_list:
skip = (id - min_id) * length_one_entry
sentences_sorted_equidistant_file.seek(skip)
sentence = sentences_sorted_equidistant_file.read(max_length_sentence).decode("utf-8").rstrip()
if sentence:
break
if sentence:
instance_start, instance_end = search_start_end_index_in_sentence(sentence.lower(), row[1].lower())
if instance_start != -1 and instance_end != -1:
sentence = sentence[:instance_start] + '<strong>' + sentence[instance_start:instance_end] + '</strong>' + sentence[instance_end:]
class_start, class_end = search_start_end_index_in_sentence(sentence.lower(), row[2].lower())
if class_start != -1 and class_end != -1:
sentence = sentence[:class_start] + '<strong>' + sentence[class_start:class_end] + '</strong>' + sentence[class_end:]
else:
print("No sentence for {}".format(row))
entity_list.append(sentence)
entity_list.append(row[1])#instance label
entity_list.append(row[0])#id
entity_list.append(row[2])#class label
if len(entity_list) == (k * 4): # write last line if it has enough content
writer.writerow(entity_list)
if __name__ == "__main__":
set_csv_field_size()
#create_sentence_template()
#for i in [1]: #[20, 10, 5, 3, 2, 1, 0]:
# write_mturk_relation_csv('webisa_{}_sample_500.csv'.format(i), 'webisa_{}_sample_500_mturk.csv'.format(i), 20)
#write_mturk_sentence_csv('webisa_1_sentence_sample_1000.csv', 'webisa_1_sentence_sample_1000_mturk.csv', 10)
replace_csv_in_template('mTurk_Sentence.html', 'webisa_1_sentence_sample_1000_mturk.csv','mTurk_Sentence_test_replaced.html')