forked from oscarqpe/cnn-multilabel-text-classification
/
class_DatasetBoW.py
171 lines (164 loc) · 4.85 KB
/
class_DatasetBoW.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
import xml.etree.ElementTree as et
import numpy as np
import os
import utils
import config
import random
class Dataset:
def __init__(self, path_data = "", batch = 25):
#assert os.path.exists(path_data), 'No existe el archivo con los datos de entrada ' + path_data
self.path_data = path_data
self.names = []
self.names_test = []
self.texts_train = []
self.labels_train = []
self.batch = batch
self.total_texts = 12288#12337
#self.total_texts = 7168#7270 # first 3
self.total_test = 4864#4891
self.start = 0
self.end = 0
self.start_test = 0
self.end_test = 0
self.first0 = 0
self.first1 = 0
self.first2 = 0
self.first3 = 0
self.first4 = 0
self.first5 = 0
self.first6 = 0
self.first7 = 0
self.first8 = 0
self.first9 = 0
self.label_examples = np.zeros(config.label_size)
for i in range(0, 12337):
self.names.append("text_" + str(i) + ".xml")
for i in range(0, self.total_test):
self.names_test.append("text_" + str(i) + ".xml")
def read_data(self, name, type = 1):
#print "extract: " + self.path_data + name
ruta = ""
if type == 1:
ruta = self.path_data + "train/" + name
elif type == 2:
ruta = self.path_data + "test/" + name
elif type == 3:
ruta = self.path_data + "first3/" + name
reuters = et.parse(ruta, et.XMLParser(encoding='ISO-8859-1')).getroot()
extract_labels = False
#print reuters
#for reuters in xml.findall('REUTERS'):
# print reuters
matrix = []
for text in reuters.findall("TEXT"):
body = utils.extract_body(text)
if body != "" and body != None:
extract_labels = True
#if extract_labels == True:
labels_temp = np.zeros(config.label_size)
all_labels = 0
for a_topic in reuters.findall("TOPICS"):
for a_d in a_topic.findall("D"):
try:
label_index = utils.find_label_index(a_d.text)
labels_temp[label_index] = 1.0
self.label_examples[label_index] += 1
all_labels += 1
except ValueError:
extract_labels = True
for a_topic in reuters.findall("PLACES"):
for a_d in a_topic.findall("D"):
try:
label_index = utils.find_label_index(a_d.text)
labels_temp[label_index] = 1.0
self.label_examples[label_index] += 1
all_labels += 1
except ValueError:
extract_labels = True
for a_topic in reuters.findall("PEOPLE"):
for a_d in a_topic.findall("D"):
try:
label_index = utils.find_label_index(a_d.text)
labels_temp[label_index] = 1.0
self.label_examples[label_index] += 1
all_labels += 1
except ValueError:
extract_labels = True
for a_topic in reuters.findall("ORGS"):
for a_d in a_topic.findall("D"):
try:
label_index = utils.find_label_index(a_d.text)
labels_temp[label_index] = 1.0
self.label_examples[label_index] += 1
all_labels += 1
except ValueError:
extract_labels = True
for a_topic in reuters.findall("EXCHANGES"):
for a_d in a_topic.findall("D"):
try:
label_index = utils.find_label_index(a_d.text)
labels_temp[label_index] = 1.0
self.label_examples[label_index] += 1
all_labels += 1
except ValueError:
extract_labels = True
if all_labels != 0:
#print("READ...")
self.labels_train = np.append(self.labels_train, labels_temp)
self.texts_train = np.append(self.texts_train, utils.stop_characters(body.text))
extract_labels = False
else:
extract_labels = False
def next_batch(self):
if self.end == 0:
self.start = 0
self.end = self.batch
elif self.end + self.batch >= self.total_texts:
self.start = self.end
self.end = self.total_texts
else:
self.start = self.end
self.end = self.end + self.batch
def prev_batch(self):
if self.start == 0:
self.start = 0
self.end = self.end
elif self.start - self.batch <= 0:
self.start = 0
self.end = self.batch
else:
self.end = self.start
self.start = self.start - self.batch
def generate_batch(self, type):
start = self.start
end = self.end
self.texts_train = np.array([])
self.labels_train = np.array([])
for i in range(start, end):
self.read_data(self.names[i], type)
def generate_batch_test(self):
start = self.start_test
end = self.end_test
self.texts_train = np.array([])
self.labels_train = np.array([])
for i in range(start, end):
self.read_data(self.names_test[i], 2)
def next_test(self):
if self.end_test == 0:
self.start_test = 0
self.end_test = self.batch
elif self.end_test + self.batch >= self.total_test:
self.start_test = self.end_test
self.end_test = self.total_test
else:
self.start_test = self.end_test
self.end_test = self.end_test + self.batch
def all_data(self, type):
for i in range(0, 7270):
#print(self.names[i])
self.read_data(self.names[i], type)
def shuffler(self):
print ("shuffling data")
random.shuffle(self.names)
self.end = 0
self.start = 0