Skip to content

Latest commit

 

History

History
118 lines (80 loc) · 4.64 KB

DEVELOPER.md

File metadata and controls

118 lines (80 loc) · 4.64 KB
Raw

This document is intended to provide guidelines for software developers either maintaining or contributing to svtools development.

Installing svtools from a git repository

For normal installation, follow the installation instructions in INSTALL.md. Howver, during development it is sometimes desirable to install a version from a git repository directly.

svtools is based on python 2.7. First you will need to prepare your python environment. You should use pyenv virtualenv. The creation of the pyenv virtual environment and activation looks like this:

pyenv virtualenv 2.7.9 svtools_install_from_repo-2.7.9
pyenv activate svtools_install_from_repo-2.7.9

Once you have your python environment set up you will want to check out svtools from the hall-lab github repository. The develop branch contains the bleeding edge version.

git clone https://github.com/hall-lab/svtools.git svtools_test
cd svtools_test

Note: If you've made local changes or master contains differences from the last tagged release, the version identifier will reflect this.

Or you can discover the release tags on the repo and checkout the latest version

git tag -l

when you discover the version tag you wish to install you can switch to that version using:

git checkout tags/v0.2.0b1

Note: You can ignore the warning about "You are in 'detached HEAD' state."

OR, you can just proceed to install from master.

Now install the dependencies:

pip install nose
pip install coverage
pip install statsmodels

Installing statsmodels can take a few minutes, but it satisfies the requirement for numpy, pandas, and scipy.

In older environments, you may need to specify pysam versions greater than 0.8.1 and less than 0.9.0

pip install 'pysam>=0.8.1,<0.9.0'

Now we can use pip to install svtools from within the repo. If you are not already in the directory

cd svtools_test

or just

pip install .

Finally we can spot check our svtools installation and observe the version number.

svtools --version

Contributing

svtools development follows the general development model described here. To contribute code, submit a pull request against the develop branch of https://github.com/hall-lab/svtools.

Releasing a new version

Tagging a release on github

svtools manages its versions using python-versioneer. New versions are derived and generated from the names of tags in git. To release a new version, all that is needed is to tag the correct commit with an annotated tag. Always prepend versions with a 'v' character.

For example, to release version 0.0.1 from the current commit:

git tag -a v0.0.1 -m 'v0.0.1'
git push --tags

Next navigate to the github Releases page, draft a new release using the tag you just generated and add release information (a description of changes made since the last release). This will create an entry on the Github Releases page and upload the release to Zenodo.

Build a pip package and upload to PyPI

Now that you have a new release, upload the package to PyPI. To do so, you'll need a PyPI account, configuration for both the standard and test PyPI servers and necessary permissions on the svtools package.

These instructions assume you have committed no additional changes after tagging the new release.

  1. Build the new release and test by uploading to the PyPI test server.
pip install --user twine
python setup.py sdist bdist_wheel
twine upload --repository-url https://test.pypi.org/legacy/ dist/*

  1. Verify that the package appears and information looks correct at https://test.pypi.org/

  2. Verify that the test package installs correctly.

    pip install --user --extra-index-url https://pypi.org/simple --index-url https://test.pypi.org/simple svtools

  3. Build and upload the package to PyPI itself.

twine upload dist/*
  1. In a fresh virtual environment, verify that the new package installs.
pyenv virtualenv 2.7.9 test_new_package
pyenv activate test_new_package
pip install svtools