-
Notifications
You must be signed in to change notification settings - Fork 0
/
transform_data.py
206 lines (183 loc) · 9.44 KB
/
transform_data.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
203
204
205
206
import sys
import os
import nltk
from nltk.tokenize import word_tokenize
import json
if __name__ == '__main__':
dirname = os.path.dirname(__file__)
def get_path(dirname, type, file_type =0):
url = "data\\" + type
# file_type = 0 -> .txt
# file_type = 1 -> .ann
if file_type == 0:
url = url + "\\text\\"
else:
url = url + "\\attributions\\"
return os.path.join(dirname, url)
type_of_dataset = sys.argv[1]
print("Transform ",type_of_dataset," data")
articles_names = os.listdir(get_path(dirname, type_of_dataset))
attributions_names = os.listdir(get_path(dirname, type_of_dataset, 1))
def get_attribution_filename_by_article_name(article_name, attributions_names):
for attribution in attributions_names:
if article_name[:-4] in attribution:
return attribution
def get_article(dirname, type, article_name, skip_line=0):
path = get_path(dirname, type)
f = open(path+article_name, "r", encoding="utf8")
article = []
for x in f:
#Remove last characther "\n" if exists
if(x[len(x)-1]=="\n"):
x = x[:len(x)-1]
article.append(x)
for i in range(skip_line):
f.readline()
f.close()
return article
articles = {}
print("Read articles")
for articles_name in articles_names:
articles[articles_name] = get_article(dirname, type_of_dataset, articles_name, 1)
#use brat parser to decode .ann files
import brat_parser
from brat_parser import get_entities_relations_attributes_groups, Entity
# it is used in quicksort
def partition(array, low, high):
pivot = array[high]
i = low - 1
for j in range(low, high):
start_item = array[j].span[0][0] if type(array[j].span[0]) is tuple else array[j].span[0]
start_pivot = pivot.span[0][0] if type(pivot.span[0]) is tuple else pivot.span[0]
if start_item <= start_pivot:
i = i + 1
(array[i], array[j]) = (array[j], array[i])
(array[i + 1], array[high]) = (array[high], array[i + 1])
return i + 1
# https://www.geeksforgeeks.org/python-program-for-quicksort/
def quickSort(array, low, high):
if low < high:
pi = partition(array, low, high)
quickSort(array, low, pi - 1)
quickSort(array, pi + 1, high)
def get_entities_from_ann(dirname, type, filename):
file = get_path(dirname, type, 1) + filename
entities, relations, attributes, groups = get_entities_relations_attributes_groups(file)
desired_entitites = []
for id in entities:
if(entities[id].type == "Cue" or entities[id].type == "Source" or entities[id].type == "Content"):
start_index = 0
end_index = 0
i = 0
# split into multiple entities if entity has discontinuous span
if len(entities[id].span) > 1:
for span in entities[id].span:
end_index = end_index + span[1] - span[0] + 1
entity = Entity(id = str(id)+"_"+str(i), type = entities[id].type, span = (span[0], span[1]), text = entities[id].text[start_index: end_index])
desired_entitites.append(entity)
start_index = end_index
i = i + 1
else:
desired_entitites.append(entities[id])
size = len(desired_entitites)
# sort entities in order to have ascending spans
quickSort(desired_entitites, 0, size - 1)
return desired_entitites
attributions = {}
print("Read attributions")
for attributions_name in attributions_names:
attributions[attributions_name] = get_entities_from_ann(dirname, type_of_dataset, attributions_name)
# def get_percentage_of_quotations_in_article(article, attributions):
# start = 0
# counter = 0
# for instance in article:
# for index in range(len(attributions)):
# print(attributions[index]," ", attributions[index].span)
# if attributions[index].start >= start and attributions[index].end <= (start+len(instance)-1):
# counter = counter + 1
# break
# start = start + len(instance)
# return counter / len(article)
# per = get_percentage_of_quotations_in_article(articles[articles_names[0]],attributions[attribution_filename])
# print("Number of instances: ", len(articles[articles_names[0]]), " / Percentage of instances that contain source, cue or content: ", round(per*100,2), "%")
class Article:
def __init__(self):
self.article_name = ""
self.attribution_name = ""
self.article = []
self.attributions = []
articles_items = []
for article_name in articles_names:
article = Article()
article.article_name = article_name
article.attribution_name = get_attribution_filename_by_article_name(article_name, attributions_names)
article.article = articles[article_name]
article.attributions = attributions[article.attribution_name]
articles_items.append(article)
final_data = []
print("Construct final data to store in file")
for article_item in articles_items:
instances = {}
start = 0
for instance in article_item.article:
instances[instance] = (start, start + len(instance))
start = start + len(instance) + 2
instances_keys = list(instances.keys())
def reformat_data(splitted_instance, splitted_instance_tags):
final_splitted_instance = []
final_splitted_instance_tags = []
for j in range(len(splitted_instance)):
splitted_instance_item = splitted_instance[j]
splitted_instance_tags_item = splitted_instance_tags[j]
excepted_symbols = []#[']', '_', '{', ':', '(', '}', '$', ';', ')', '[', '%', '#', '@', '-', '^', '`', '\\', '?', '|', ',', '/', '~', '>', '<', '!', '=', '.', '+', '*', '&']
temp_tokens = word_tokenize(splitted_instance_item)
tokens=[]
for token in temp_tokens:
if token not in excepted_symbols:
tokens.append(token)
final_splitted_instance.extend(tokens)
for i in range(len(tokens)):
if splitted_instance_tags_item == 'Source' and i == 0:
final_splitted_instance_tags.append('B-Source')
elif splitted_instance_tags_item == 'Source':
final_splitted_instance_tags.append('I-Source')
elif splitted_instance_tags_item == 'Cue' and i == 0:
final_splitted_instance_tags.append('B-Cue')
elif splitted_instance_tags_item == 'Cue':
final_splitted_instance_tags.append('I-Cue')
elif splitted_instance_tags_item == 'Content' and i == 0:
final_splitted_instance_tags.append('B-Content')
elif splitted_instance_tags_item == 'Content':
final_splitted_instance_tags.append('I-Content')
else:
final_splitted_instance_tags.append('O')
return final_splitted_instance, final_splitted_instance_tags
tagged_instances = []
for key in instances_keys:
splitted_instance = [key]
splitted_instance_tags = ['O']
temp_index = instances[key][0]
for attribution in articles_items[0].attributions:
attribution_start = attribution.span[0][0] if type(attribution.span[0]) is tuple else attribution.span[0]
attribution_end = attribution.span[0][1] if type(attribution.span[0]) is tuple else attribution.span[1]
if instances[key][0] <= attribution_start and attribution_end <= instances[key][1]:
last_item = splitted_instance.pop()
splitted_instance_tags.pop()
if last_item[:attribution_start - temp_index] != "" and last_item[:attribution_start - temp_index] != " " :
splitted_instance.append(last_item[:attribution_start-temp_index])
splitted_instance_tags.append('O')
splitted_instance.append(last_item[attribution_start - temp_index:attribution_end - temp_index])
splitted_instance_tags.append(attribution.type)
if last_item[attribution_end- temp_index:] != "" and last_item[attribution_end - temp_index:] != " ":
splitted_instance.append(last_item[attribution_end -temp_index:])
splitted_instance_tags.append('O')
temp_index = attribution_end
if instances[key][1] < attribution_start:
break
splitted_instance, splitted_instance_tags = reformat_data(splitted_instance, splitted_instance_tags)
tagged_instances.append([splitted_instance, splitted_instance_tags])
final_data.append(tagged_instances)
json_string = json.dumps(final_data)
f = open(os.path.join(dirname, type_of_dataset+"_data.json"), "a")
f.write(json_string)
f.close()