|
1
|
|
|
import os |
|
2
|
|
|
import pytest |
|
3
|
|
|
import sys |
|
4
|
|
|
import tempfile |
|
5
|
|
|
import string |
|
6
|
|
|
|
|
7
|
|
|
import numpy as np |
|
8
|
|
|
import os.path as op |
|
9
|
|
|
import diff_classifier.imagej as ij |
|
10
|
|
|
from diff_classifier.utils import csv_to_pd |
|
11
|
|
|
|
|
12
|
|
|
from urllib.request import urlretrieve |
|
13
|
|
|
|
|
14
|
|
|
is_travis = "CI" in os.environ.keys() |
|
15
|
|
|
is_mac = sys.platform == "darwin" |
|
16
|
|
|
|
|
17
|
|
|
|
|
18
|
|
|
#@pytest.mark.skipif(is_travis, reason="We're running this on Travis") |
|
19
|
|
|
#@pytest.mark.skipif(is_mac, reason="This doesn't work on Macs yet") |
|
20
|
|
|
@pytest.mark.xfail |
|
21
|
|
|
def test_run_tracking(): |
|
22
|
|
|
tempf = tempfile.NamedTemporaryFile(suffix='.csv') |
|
23
|
|
|
ij.track('http://fiji.sc/samples/FakeTracks.tif', tempf.name) |
|
24
|
|
|
assert op.exists(tempf.name) |
|
25
|
|
|
|
|
26
|
|
|
df = csv_to_pd(tempf.name) |
|
27
|
|
|
assert df.shape == (84, 8) |
|
28
|
|
|
|
|
29
|
|
|
|
|
30
|
|
|
def test_mean_intensity(): |
|
31
|
|
|
fname = 'FakeTracks.tif' |
|
32
|
|
|
cwd = os.getcwd() |
|
33
|
|
|
fullname = os.path.join(cwd, fname) |
|
34
|
|
|
urlretrieve('http://fiji.sc/samples/FakeTracks.tif', filename=fullname) |
|
35
|
|
|
|
|
36
|
|
|
test = np.round(ij.mean_intensity(fname, frame=0), 1) |
|
37
|
|
|
assert test == 20.0 |
|
38
|
|
|
|
|
39
|
|
|
|
|
40
|
|
|
def test_partition_im(): |
|
41
|
|
|
fname = 'FakeTracks.tif' |
|
42
|
|
|
cwd = os.getcwd() |
|
43
|
|
|
fullname = os.path.join(cwd, fname) |
|
44
|
|
|
urlretrieve('http://fiji.sc/samples/FakeTracks.tif', filename=fullname) |
|
45
|
|
|
|
|
46
|
|
|
rows = 2 |
|
47
|
|
|
cols = 2 |
|
48
|
|
|
names = ij.partition_im(fullname, irows=rows, icols=2, ores=(128, 128), |
|
49
|
|
|
ires=(64, 64)) |
|
50
|
|
|
for name in names: |
|
51
|
|
|
assert os.path.isfile(name) |
|
52
|
|
|
|
|
53
|
|
|
|
|
54
|
|
|
def test_regress_sys(): |
|
55
|
|
|
fname = 'FakeTracks.tif' |
|
56
|
|
|
cwd = os.getcwd() |
|
57
|
|
|
fullname = os.path.join(cwd, fname) |
|
58
|
|
|
urlretrieve('http://fiji.sc/samples/FakeTracks.tif', filename=fullname) |
|
59
|
|
|
|
|
60
|
|
|
all_videos = ['FakeTracks']*10 |
|
61
|
|
|
yfit = [10, 9] |
|
62
|
|
|
training_size = 2 |
|
63
|
|
|
|
|
64
|
|
|
tracks = ij.regress_sys(cwd, all_videos, yfit, training_size, |
|
65
|
|
|
have_output=False, download=False) |
|
66
|
|
|
for track in tracks: |
|
67
|
|
|
assert track == 'FakeTracks' |
|
68
|
|
|
|
|
69
|
|
|
regress = ij.regress_sys(cwd, all_videos, yfit, training_size, |
|
70
|
|
|
have_output=True, download=False) |
|
71
|
|
|
assert len(regress) == 8 |
|
72
|
|
|
|
|
73
|
|
|
all_videos = list(string.ascii_lowercase) |
|
74
|
|
|
yfinal = ['e', 'b'] |
|
75
|
|
|
|
|
76
|
|
|
tracks = ij.regress_sys(cwd, all_videos, yfit, training_size, |
|
77
|
|
|
have_output=False, download=False) |
|
78
|
|
|
counter = 0 |
|
79
|
|
|
for track in tracks: |
|
80
|
|
|
assert track == yfinal[counter] |
|
81
|
|
|
counter = counter + 1 |
|
82
|
|
|
|
|
83
|
|
|
|
|
84
|
|
|
def test_regress_tracking_params(): |
|
85
|
|
|
fname = 'FakeTracks.tif' |
|
86
|
|
|
cwd = os.getcwd() |
|
87
|
|
|
fullname = os.path.join(cwd, fname) |
|
88
|
|
|
urlretrieve('http://fiji.sc/samples/FakeTracks.tif', filename=fullname) |
|
89
|
|
|
|
|
90
|
|
|
all_videos = ['FakeTracks']*10 |
|
91
|
|
|
yfit = [10, 9] |
|
92
|
|
|
training_size = 2 |
|
93
|
|
|
|
|
94
|
|
|
regress = ij.regress_sys(cwd, all_videos, yfit, training_size, |
|
95
|
|
|
have_output=True, download=False) |
|
96
|
|
|
|
|
97
|
|
|
quality = ij.regress_tracking_params(regress, 'FakeTracks', frame=0) |
|
98
|
|
|
assert quality == 9.5 |
|
99
|
|
|
quality = ij.regress_tracking_params(regress, 'FakeTracks', regmethod='SVR', |
|
100
|
|
|
frame=0) |
|
101
|
|
|
assert quality == 9.5 |
|
102
|
|
|
quality = ij.regress_tracking_params(regress, 'FakeTracks', |
|
103
|
|
|
regmethod='BayesianRidge', frame=0) |
|
104
|
|
|
assert quality == 9.5 |
|
105
|
|
|
quality = ij.regress_tracking_params(regress, 'FakeTracks', |
|
106
|
|
|
regmethod='SGDRegressor', frame=0) |
|
107
|
|
|
assert quality < -10000 |
|
108
|
|
|
quality = ij.regress_tracking_params(regress, 'FakeTracks', |
|
109
|
|
|
regmethod='LassoLars', frame=0) |
|
110
|
|
|
assert quality == 9.5 |
|
111
|
|
|
quality = ij.regress_tracking_params(regress, 'FakeTracks', |
|
112
|
|
|
regmethod='ARDRegression', frame=0) |
|
113
|
|
|
assert quality == 9.5 |
|
114
|
|
|
quality = ij.regress_tracking_params(regress, 'FakeTracks', |
|
115
|
|
|
regmethod='PassiveAggressiveRegressor', |
|
116
|
|
|
frame=0) |
|
117
|
|
|
assert np.round(quality, 1) == 9.1 or np.round(quality, 1) == 9.9 |
|
118
|
|
|
quality = ij.regress_tracking_params(regress, 'FakeTracks', |
|
119
|
|
|
regmethod='TheilSenRegressor', frame=0) |
|
120
|
|
|
assert quality == 9.5 |
|
121
|
|
|
quality = ij.regress_tracking_params(regress, 'FakeTracks', |
|
122
|
|
|
regmethod='None', frame=0) |
|
123
|
|
|
assert quality == 3.0 |
|
124
|
|
|
|
ccurtis7 /
diff_classifier
