forked from tensorflow/gan
-
Notifications
You must be signed in to change notification settings - Fork 0
/
release_pypi_package.sh
executable file
·71 lines (60 loc) · 2.1 KB
/
release_pypi_package.sh
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
#!/bin/bash
#
# A script for creating and releasing the TensorFlow GAN PyPi package.
#
# Usage:
# ./release_pypi_package.sh {test|live}
#
# Order of operations of this script:
# 1) Build and test the whls (one for each desired version of Python).
# 2) Publish the whl (in test or live mode). Contact a TF-GAN team member if you
# think you have a good reason to be building PyPi packages.
#
# To test that the package was uploaded, you can run the following:
# $ pip install --index-url https://test.pypi.org/simple/ tensorflow-gan --upgrade
# $ pip install tensorflow-gan --upgrade
#
# The resulting packages will be visible here:
# $ https://pypi.org/project/tensorflow-gan/
# $ https://test.pypi.org/project/tensorflow-gan/
set -e # fail and exit on any command erroring
# Bring in useful functions.
source gbash.sh || exit 1
source module pypi_utils.sh
mode=$1
if [ "${mode}" = "live" ]; then
TWINE_ARGS=""
elif [ "${mode}" = "test" ]; then
TWINE_ARGS="--repository-url https://test.pypi.org/legacy/"
else
echo "Need to pass 'live' or 'test' as first arg to script."
exit 1
fi
echo "Building whl and test it..."
WHL_TMP_DIR=$(mktemp -d)
# TODO(joelshor): Add support for python3.x.
test_build_and_install_whl "python3.6" "TF2.x" "${WHL_TMP_DIR}"
# Publish to PyPI
read -p "Publish to ${mode}? (y/n) " -r
echo
if [[ $REPLY =~ ^[Yy]$ ]]; then
make_virtual_env "python3.6" "TF2.x"
# TODO(joelshor): Check that pip exists and, if not, install it.
# Probably install using `sudo apt-get install python-pip`.
# Check that twine exists. Note that on some systems, we must install it with
# `pip install twine` and not `apt-get install twine`.
pip install twine
echo "Publishing to PyPI"
# Since we must install using `pip install` (TODO(joelshor): Why?), we must
# run twine inside python ex `python -m twine xxx` instead of `twine xxx`.
# Note: Files should be of the form
# ${WHL_TMP_DIR}/wheel/[python2.7|python3.5]/*.whl
python -m twine upload ${TWINE_ARGS} ${WHL_TMP_DIR}/wheel/*/*
# Deactivate virtualenv.
deactivate
else
echo "Skipping upload"
exit 1
fi
rm -rf $TMP_DIR1
rm -rf $TMP_DIR2