Skip to content

Latest commit

 

History

History
165 lines (115 loc) · 4.47 KB

README.dev.md

File metadata and controls

165 lines (115 loc) · 4.47 KB
Raw

machine_learning developer documentation

If you're looking for user documentation, go here.

Development install

# Create a virtual environment, e.g. with
python3 -m venv env

# activate virtual environment
source env/bin/activate

# make sure to have a recent version of pip and setuptools
python3 -m pip install --upgrade pip setuptools

# (from the project root directory)
# install machine_learning as an editable package
python3 -m pip install --no-cache-dir --editable .
# install development dependencies
python3 -m pip install --no-cache-dir --editable .[dev]

Afterwards check that the install directory is present in the PATH environment variable.

Running the tests

Running the tests requires an activated virtual environment with the development tools installed.

# unit tests
pytest
pytest tests/

Running linters locally

For linting we will use prospector and to sort imports we will use isort. Running the linters requires an activated virtual environment with the development tools installed.

# linter
prospector

# recursively check import style for the machine_learning module only
isort --recursive --check-only machine_learning

# recursively check import style for the machine_learning module only and show
# any proposed changes as a diff
isort --recursive --check-only --diff machine_learning

# recursively fix import style for the machine_learning module only
isort --recursive machine_learning

You can enable automatic linting with prospector and isort on commit by enabling the git hook from .githooks/pre-commit, like so:

git config --local core.hooksPath .githooks

Generating the API docs

cd docs
make html

The documentation will be in docs/_build/

Versioning

Bumping the version across all files is done with bumpversion, e.g.

bumpversion major
bumpversion minor
bumpversion patch

Making a release

This section describes how to make a release in 3 parts:

  1. preparation
  2. making a release on PyPI
  3. making a release on GitHub

(1/3) Preparation

  1. Update the CHANGELOG.md
  2. Verify that the information in CITATION.cff is correct, and that .zenodo.json contains equivalent data
  3. Make sure the version has been updated.
  4. Run the unit tests with pytest tests/

(2/3) PyPI

In a new terminal, without an activated virtual environment or an env directory:

# prepare a new directory
cd $(mktemp -d --tmpdir machine_learning.XXXXXX)

# fresh git clone ensures the release has the state of origin/main branch
git clone https://github.com/online-behaviour/machine-learning .

# prepare a clean virtual environment and activate it
python3 -m venv env
source env/bin/activate

# make sure to have a recent version of pip and setuptools
python3 -m pip install --upgrade pip setuptools

# install runtime dependencies and publishing dependencies
python3 -m pip install --no-cache-dir .
python3 -m pip install --no-cache-dir .[publishing]

# clean up any previously generated artefacts 
rm -rf machine_learning.egg-info
rm -rf dist

# create the source distribution and the wheel
python3 setup.py sdist bdist_wheel

# upload to test pypi instance (requires credentials)
twine upload --repository-url https://test.pypi.org/legacy/ dist/*

Visit https://test.pypi.org/project/machine_learning and verify that your package was uploaded successfully. Keep the terminal open, we'll need it later.

In a new terminal, without an activated virtual environment or an env directory:

cd $(mktemp -d --tmpdir machine_learning-test.XXXXXX)

# prepare a clean virtual environment and activate it
python3 -m venv env
source env/bin/activate

# make sure to have a recent version of pip and setuptools
pip install --upgrade pip setuptools

# install from test pypi instance:
python3 -m pip -v install --no-cache-dir \
--index-url https://test.pypi.org/simple/ \
--extra-index-url https://pypi.org/simple machine_learning

Check that the package works as it should when installed from pypitest.

Then upload to pypi.org with:

# Back to the first terminal,
# FINAL STEP: upload to PyPI (requires credentials)
twine upload dist/*

(3/3) GitHub

Don't forget to also make a release on GitHub. If your repository uses the GitHub-Zenodo integration this will also trigger Zenodo into making a snapshot of your repository and sticking a DOI on it.