forked from NifTK/NiftyNet
-
Notifications
You must be signed in to change notification settings - Fork 0
/
.gitlab-ci.yml
244 lines (220 loc) · 10.3 KB
/
.gitlab-ci.yml
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
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
stages:
- dev_test
- pip_test
- pip_publish
testjob:
stage: dev_test
only:
- master
- dev
- dev-staging
- tags
- supports-axbxc-patch
script:
# !!kill coverage in case of hanging processes
- if pgrep coverage; then pkill -f coverage; fi
# print system info
- which nvidia-smi
- nvidia-smi
- pwd
- python -c "import tensorflow as tf; print tf.__version__"
- python -c "import tensorflow as tf; from tensorflow.python.client import device_lib; print device_lib.list_local_devices()"
- ls -la /dev | grep nvidia
- echo $(python tests/get_gpu_index.py)
- export CUDA_VISIBLE_DEVICES=$(python tests/get_gpu_index.py)
# download data
# - wget -q https://www.dropbox.com/s/y7mdh4m9ptkibax/example_volumes.tar.gz
# - tar -xzvf example_volumes.tar.gz
- wget -q https://www.dropbox.com/s/lioecnpv82r5n6e/example_volumes_v0_2.tar.gz
- tar -xzvf example_volumes_v0_2.tar.gz
# - wget -q https://www.dropbox.com/s/94wa4fl8f8k3aie/testing_data.tar.gz
# - tar -xzvf testing_data.tar.gz
- wget -q https://www.dropbox.com/s/p7b3t2c3mewtree/testing_data_v0_2.tar.gz
- tar -xzvf testing_data_v0_2.tar.gz
# run python code with coverage wrapper
- coverage erase
- coverage run -a --source . net_segment.py train -c config/highres3dnet_config.ini --batch_size=1 --num_threads=2 --queue_length=40 --max_iter=10
- coverage run -a --source . net_segment.py inference -c config/highres3dnet_config.ini --batch_size 8 --spatial_window_size 64,64,64 --queue_length 64
- coverage run -a --source . net_segment.py train -c config/scalenet_config.ini --batch_size 1 --queue_length 5 --num_threads 2
- coverage run -a --source . net_segment.py inference -c config/scalenet_config.ini --batch_size 16 --spatial_window_size 64,64,64 --queue_length 32
- coverage run -a --source . net_segment.py train -c config/vnet_config.ini --batch_size 1 --queue_length 5 --num_threads 2 --activation_function relu
- coverage run -a --source . net_segment.py inference -c config/vnet_config.ini --batch_size 16 --spatial_window_size 64,64,64 --queue_length 32 --activation_function relu
# need a large GPU to run
#- coverage run -a --source . net_segment.py train -c config/unet_config.ini --batch_size 1 --queue_length 5 --num_threads 2
#- coverage run -a --source . net_segment.py inference -c config/unet_config.ini --batch_size 1 --spatial_window_size 96,96,96 --queue_length 5
#- coverage run -a --source . net_segment.py train -c config/deepmedic_config.ini --batch_size 128 --queue_length 48 --num_threads 4
#- coverage run -a --source . net_segment.py inference -c config/deepmedic_config.ini --batch_size 12 --spatial_window_size 135,135,135 --queue_length 128
- coverage run -a --source . net_segment.py train -c config/default_segmentation.ini --batch_size 3 --queue_length 6
- coverage run -a --source . net_segment.py train -c config/default_segmentation.ini --batch_size 3 --queue_length 6 --starting_iter 0 --max_iter 15
- coverage run -a --source . net_segment.py inference -c config/default_segmentation.ini --spatial_window_size 84,84,84 --batch_size 7 --queue_length 14 --inference_iter 14
- coverage run -a --source . net_segment.py train -c config/default_multimodal_segmentation.ini --batch_size 3
- coverage run -a --source . net_segment.py inference -c config/default_multimodal_segmentation.ini --spatial_window_size 64,64 --batch_size 7
- coverage run -a --source . net_gan.py train -c config/GAN_demo_train_config.ini --max_iter 5
- coverage run -a --source . net_gan.py inference -c config/GAN_demo_train_config.ini
- coverage run -a --source . net_autoencoder.py train -c config/vae_config.ini --max_iter 5
- coverage run -a --source . net_autoencoder.py inference -c config/vae_config.ini --inference_type sample
- coverage run -a --source . net_autoencoder.py inference -c config/vae_config.ini --inference_type encode --save_seg_dir output/vae_demo_features
- coverage run -a --source . net_autoencoder.py inference -c config/vae_config.ini --inference_type encode-decode
- coverage run -a --source . -m unittest discover -s "tests" -p "*_test.py"
- coverage report -m
- echo 'finished test'
tags:
- gift-linux
quicktest:
stage: dev_test
except:
- master
- dev
- dev-staging
- tags
- supports-axbxc-patch
- 112-publish-api-docs-online
- 147-revise-contribution-guidelines-to-include-github
script:
# print system info
- which nvidia-smi
- nvidia-smi
- pwd
- python -c "import tensorflow as tf; print tf.__version__"
- python -c "import tensorflow as tf; from tensorflow.python.client import device_lib; print device_lib.list_local_devices()"
- ls -la /dev | grep nvidia
- echo $(python tests/get_gpu_index.py)
- export CUDA_VISIBLE_DEVICES=$(python tests/get_gpu_index.py)
# download data
# - wget -q https://www.dropbox.com/s/y7mdh4m9ptkibax/example_volumes.tar.gz
# - tar -xzvf example_volumes.tar.gz
- wget -q https://www.dropbox.com/s/lioecnpv82r5n6e/example_volumes_v0_2.tar.gz
- tar -xzvf example_volumes_v0_2.tar.gz
#- wget -q https://www.dropbox.com/s/94wa4fl8f8k3aie/testing_data.tar.gz
#- tar -xzvf testing_data.tar.gz
- wget -q https://www.dropbox.com/s/p7b3t2c3mewtree/testing_data_v0_2.tar.gz
- tar -xzvf testing_data_v0_2.tar.gz
- coverage erase
# run only fast tests
- QUICKTEST=True coverage run -a --source . -m unittest discover -s "tests" -p "*_test.py"
- coverage report -m
- echo 'finished quick tests'
tags:
- gift-linux
pip-installer:
stage: pip_test
only:
- master
- dev
- dev-staging
- tags
- supports-axbxc-patch
script:
# source utils
- source ci/utils.sh
# following three lines copied over from dev script:
- ls -la /dev | grep nvidia
- echo $(python tests/get_gpu_index.py)
- export CUDA_VISIBLE_DEVICES=$(python tests/get_gpu_index.py)
# create a Python file that will import all available packages from the pip installer
- package_importer="$(pwd)/import_niftynet_packages.py"
# traverse the file hierarchy recursively to discover all packages
- find niftynet -type f \( ! -name . \) -print | grep '.py$' | grep -v __init__ | sed 's/\.\.\///g;s/\//\./g;s/\.py//g;s/^niftynet/import niftynet/g' > $package_importer
# save NiftyNet folder path just in case
- export niftynet_dir=$(pwd)
# create the NiftyNet wheel
- rm -rf dist # remove dist directory, just in case
- sh ci/bundlewheel.sh
- source ci/findwheel.sh
- echo $niftynet_wheel
# ============= Python 2 ============================
# create a virtual env to test pip installer
- venv="niftynet-pip-installer-venv-py2"
- mypython=$(which python2)
- virtualenv -p $mypython $venv
- cd $venv
- venv_dir=$(pwd)
- source bin/activate
# print Python version to CI output
- which python
- python --version
# NiftyNet console entries should fail gracefully if TF not installed
# i.e. check that the warning displays the TF website
- cd $niftynet_dir
- set +e
- python -c "import niftynet" 2>&1 | grep "https://www.tensorflow.org/"
- set -e
- cd $venv_dir
# install TF
- pip install tensorflow-gpu==1.1
# install using built NiftyNet wheel
- pip install $niftynet_wheel
# install SimpleITK for package importer test to work properly
- pip install simpleitk
# check whether all packages are importable
- cat $package_importer
- python $package_importer
# test niftynet command
- ln -s /home/gitlab-runner/environments/niftynet/data/example_volumes ./example_volumes
- net_segment train -c $niftynet_dir/config/default_segmentation.ini --name toynet --batch_size 3 --max_iter 5
- net_segment inference -c $niftynet_dir/config/default_segmentation.ini --name toynet --spatial_window_size 80,80,80 --batch_size 8
# deactivate virtual environment
- deactivate
- cd $niftynet_dir
# ============= Python 3 ============================
# create a virtual env to test pip installer
- venv="niftynet-pip-installer-venv-py3"
- mypython=$(which python3)
- virtualenv -p $mypython $venv
- cd $venv
- venv_dir=$(pwd)
- source bin/activate
# print Python version to CI output
- which python
- python --version
# NiftyNet console entries should fail gracefully if TF not installed
# i.e. check that the warning displays the TF website
- cd $niftynet_dir
- set +e
- python -c "import niftynet" 2>&1 | grep "https://www.tensorflow.org/"
- set -e
- cd $venv_dir
# install TF
- pip install tensorflow-gpu==1.1
# install using built NiftyNet wheel
- pip install $niftynet_wheel
# install SimpleITK for package importer test to work properly
- pip install simpleitk
# check whether all packages are importable
- cat $package_importer
- python $package_importer
# test niftynet command
- ln -s /home/gitlab-runner/environments/niftynet/data/example_volumes ./example_volumes
- net_segment train -c $niftynet_dir/config/default_segmentation.ini --name toynet --batch_size 3 --max_iter 5
- net_segment inference -c $niftynet_dir/config/default_segmentation.ini --name toynet --spatial_window_size 80,80,80 --batch_size 8
# deactivate virtual environment
- deactivate
- cd $niftynet_dir
tags:
- gift-adelie
pip-camera-ready:
stage: pip_publish
only:
- tags
script:
# Copy wheel created in previous stage to a specific location on GIFT-Adelie
- export niftynet_dir=$(pwd)
# create the NiftyNet wheel
- rm -rf dist # remove dist directory, just in case
- sh ci/bundlewheel.sh
- source ci/findwheel.sh
- echo $niftynet_wheel
# Creat camera-ready folder if doesn't exist
- camera_ready_dir=/home/gitlab-runner/environments/niftynet/pip/camera-ready
- mkdir -p $camera_ready_dir
- ls -lrtha $camera_ready_dir
# Clean up the camera-ready folder if already there
- rm -rf $camera_ready_dir/*.whl
# Finally do copy
- cp $niftynet_wheel $camera_ready_dir
- ls -lrtha $camera_ready_dir
# Instruct developer which file to publish
- echo "Camera-ready pip installer bundle (wheel) created:"
- echo "$(ls $camera_ready_dir/*.whl)"
tags:
- gift-adelie