-
Notifications
You must be signed in to change notification settings - Fork 1
/
Data.py
162 lines (144 loc) · 5.13 KB
/
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
from __future__ import print_function
import numpy as np
import pandas as pd
import os
from PIL import Image
iris_path="Data/IRIS/iris.data"
earthquake_path="Data/Earthquake/Earthquakes"
osuleaf_path="Data/OSULeaf/OSULeaf"
face_path="/media/parthosarothi/OHWR/Dataset/UCI_CMU_Faces"
def load_iris_data(iris_csv):
f=open(iris_csv)
line=f.readline()
iris_df=pd.DataFrame()
while line:
info=line.strip("\n").split(",")
if(len(info)==5):
label=info[-1]
features=info[:4]
features.append(label)
#print(features)
df = pd.DataFrame([features], columns=['SL', 'SW', 'PL', 'PW', 'class'])
iris_df=iris_df.append(df)
line=f.readline()
return iris_df
def load_and_split_iris(iris_path,split_ratio):
iris_df = load_iris_data(iris_path)
class_labels = list(set(iris_df['class']))
print(class_labels)
train_set=pd.DataFrame()
test_set=pd.DataFrame()
for cl in class_labels:
class_loc=iris_df.loc[iris_df['class']==cl]
nb_samples=len(class_loc)
class_loc=class_loc.sample(nb_samples)
#print(class_loc)
test_volume=int(nb_samples*split_ratio)
train_volume=nb_samples-test_volume
#print(class_loc.iloc[train_volume:,:])
train_set=train_set.append(class_loc.iloc[:train_volume,:])
test_set=test_set.append(class_loc.iloc[train_volume:,:])
nbtests=len(test_set)
test_set=test_set.sample(nbtests)
nbtrain = len(train_set)
train_set = train_set.sample(nbtrain)
print("Train samples=%d Test samples=%d"%(nbtrain,nbtests))
return train_set,test_set,class_labels
def make_one_hot(batch_labels,all_labels):
total_classes=len(all_labels)
total_samples=len(batch_labels)
one_hot=np.zeros([total_samples,total_classes],dtype=float)
for ts in range(total_samples):
one_hot[ts][all_labels.index(batch_labels[ts])]=1
return one_hot
def load_earthquake_data(earthquake_csv):
f=open(earthquake_csv)
line=f.readline()
sequence_length=[]
labels=[]
features=[]
while line:
info=line.strip("\n").split(",")
label=info[0]
label_one_hot=[0,0]
label_one_hot[int(label)]=1
feat=info[1:]
nbfeatures=len(feat)
sequence_length.append(nbfeatures)
labels.append(label_one_hot)
features.append(feat)
line=f.readline()
max_length=max(sequence_length)
min_length=min(sequence_length)
print("Maximum sequence length: ",max_length," minimum sequence length: ",min_length)
return np.asarray(features),np.asarray(labels),sequence_length
def load_timeseries_data(timeseries_csv,nbclass,class_label_from_0=False):
f=open(timeseries_csv)
line=f.readline()
sequence_length=[]
labels=[]
features=[]
while line:
info=line.strip("\n").split(",")
label=info[0]
label_one_hot=np.zeros([nbclass])
if(class_label_from_0):
label_index=int(label)
else:
label_index=int(label)-1
label_one_hot[label_index]=1
feat=info[1:]
nbfeatures=len(feat)
sequence_length.append(nbfeatures)
labels.append(label_one_hot)
features.append(feat)
line=f.readline()
max_length=max(sequence_length)
min_length=min(sequence_length)
print("Maximum sequence length: ",max_length," minimum sequence length: ",min_length)
return np.asarray(features),np.asarray(labels),sequence_length
def split_cmu_face_images(face_path,resolution,test_split):
all_files=[]
for root,sd,files in os.walk(face_path):
for fn in files:
# print(fn)
if(fn[-3:]=='pgm'):
filename = fn[:-4]
attributes=filename.split("_")
if(attributes[-1]==resolution):
all_files.append(os.path.join(root,fn))
total=len(all_files)
nbtests=int(total*test_split)
nbtrain=total-nbtests
train_data=all_files[:nbtrain]
test_data=all_files[nbtrain:]
print("Directory scan complete: %d samples Train %d Test %d" % (total,len(train_data),len(test_data)))
return train_data,test_data
def load_image(imfile):
print("Loading %s"%imfile)
img=Image.open(imfile).convert('L')
w_c,h_r=img.size
pixels=img.getdata()
imagemat=np.reshape(pixels,[h_r,w_c])
return imagemat
def load_x_y_cmu(files,class_labels):
x_ = []
y_ = []
for fl in files:
imgmat = load_image(fl)
x_.append(imgmat)
info = fl.split(".")[0].split("_")[-2]
one_hot_label = [0, 0]
one_hot_label[class_labels.index(info)] = 1.0
y_.append(one_hot_label)
print("\t Label %s" % info)
x_ = np.asarray(x_)
y_ = np.asarray(y_)
return x_, y_
def load_cmu_face_images(face_path,resolution,test_split,class_labels):
train,test=split_cmu_face_images(face_path,resolution,test_split)
x_train,y_train=load_x_y_cmu(train,class_labels)
x_test,y_test=load_x_y_cmu(test,class_labels)
print("Train: X ",x_train.shape," Y ",y_train.shape)
print("Test: X ", x_test.shape, " Y ", y_test.shape)
return x_train,y_train,x_test,y_test