/
tasks.py
169 lines (133 loc) · 5.23 KB
/
tasks.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
#!/usr/bin/env python3
# -*- coding: utf-8; -*-
import cv2
import gzip, pickle
import multiprocessing as mp
import numpy as np
from invoke import task, run
from pathlib import Path
from itertools import product
from sklearn.decomposition import RandomizedPCA
from sklearn.preprocessing import StandardScaler
from sklearn.neighbors import KNeighborsClassifier
from skimage.io import imread, imsave, imread_collection
from src.imageutils import prepare, roi, imgdir_to_array
#-----------------------------------------------------------------------------
datadir = Path('./data/')
training_models = datadir / 'models.pickle.gz'
training_data_npz = datadir / 'training-data.npz'
training_data_dir = datadir / 'training/'
testing_data_dir = datadir / 'testing/'
labeled_data_dir = datadir / 'labeled/'
raw_training_data_dir = datadir / 'raw/'
@task
def prepare_labeled(root='training', parallel=False):
root = {'training': training_data_dir, 'testing': testing_data_dir}[root]
# Create all destination directories.
labels = product('ABC', '0123456789')
for group, label in labels:
path = root/group/label
if not path.exists():
path.mkdir(parents=True)
ic = imread_collection('%s/*/*/*.jpg' % labeled_data_dir)
images = ((ic[n], name, root) for n, name in enumerate(ic.files))
if parallel:
pool = mp.Pool(7)
pool.map(prepare_labeled_inner, images)
else:
for img in images:
prepare_labeled_inner(img)
def prepare_labeled_inner(arg):
img, name, root = arg
name = Path(name)
group, label = name.parts[-3:-1] # e.g. A, 1
img = prepare(img)
A, B, C = roi(img)
img = vars()[group] # A, B or C
dest = root / group / label / name.with_suffix('.png').name
print('Source: %s\nDest: %s' % (name, dest))
if any(i is None for i in (A,B,C)):
print('Invalid input ... skipping.')
return
imsave(str(dest), img)
@task(aliases=['training-data'])
def training_data():
print('Loading images from %s.' % training_data_dir)
images_A, labels_A, paths_A = imgdir_to_array(training_data_dir/'A')
images_B, labels_B, paths_A = imgdir_to_array(training_data_dir/'B')
images_C, labels_C, paths_A = imgdir_to_array(training_data_dir/'C')
def getcounts(arr):
counts = np.bincount(arr)
indexes = np.nonzero(counts)[0]
return zip(indexes, counts[indexes])
print('A: %s' % getcounts(labels_A))
print('B: %s' % getcounts(labels_B))
print('C: %s' % getcounts(labels_C))
print('Serializing images to %s.' % training_data_npz)
np.savez_compressed(str(training_data_npz),
images_A=images_A, labels_A=labels_A,
images_B=images_B, labels_B=labels_B,
images_C=images_C, labels_C=labels_C)
@task
def invert_images(dest='training'):
dest = {'training': training_data_dir, 'testing': testing_data_dir}[dest]
for path in dest.rglob('**/*.png'):
print('%s' % path)
img = imread(str(path), 0)
img = np.invert(img)
imsave(str(path), img)
@task
def train_pca_std():
print('Loading training data %s.' % training_data_npz)
npz = np.load(str(training_data_npz))
images_A, labels_A = npz['images_A'], npz['labels_A']
images_B, labels_B = npz['images_B'], npz['labels_B']
images_C, labels_C = npz['images_C'], npz['labels_C']
def trainset(data, labels):
pca = RandomizedPCA(n_components=10)
std = StandardScaler()
data = np.reshape(data, (data.shape[0], -1))
data = pca.fit_transform(data)
data = std.fit_transform(data)
knn = KNeighborsClassifier()
knn.fit(data, labels)
return pca, std, knn
pca_A, std_A, knn_A = trainset(images_A, labels_A)
pca_B, std_B, knn_B = trainset(images_B, labels_B)
pca_C, std_C, knn_C = trainset(images_C, labels_C)
with gzip.open(str(training_models), 'wb') as fh:
print('Serializing to %s.' % fh.name)
res = [pca_A, std_A, knn_A, pca_B, std_B, knn_B, pca_C, std_C, knn_C]
pickle.dump(res, fh)
@task
def train():
print('Loading training data %s.' % training_data_npz)
npz = np.load(str(training_data_npz))
images_A, labels_A = npz['images_A'], npz['labels_A']
images_B, labels_B = npz['images_B'], npz['labels_B']
images_C, labels_C = npz['images_C'], npz['labels_C']
def trainset(data, labels):
knn = KNeighborsClassifier()
knn.fit(data.reshape(data.shape[0], -1), labels)
return knn
knn_A = trainset(images_A, labels_A)
knn_B = trainset(images_B, labels_B)
knn_C = trainset(images_C, labels_C)
with gzip.open(str(training_models), 'wb') as fh:
print('Serializing to %s.' % fh.name)
res = [knn_A, knn_B, knn_C]
pickle.dump(res, fh)
@task
def clean():
paths = [training_models, training_data_npz]
for path in paths:
if path.exists():
print('Removing %s' % path)
path.unlink()
@task
def deploy():
run('rsync -vr ./src ./tasks.py gasmeter.service pi@192.168.0.20:/home/pi/meter/')
run('rsync -vr ./data/training-data.npz pi@192.168.0.20:/home/pi/meter/data/')
@task
def pullrawimg():
run('rsync -vr pi@192.168.0.20:/home/pi/meter/data/raw ./data/')