Development notes#

Clone and set up pre-commit hooks#

We use pre-commit to ensure clean commits in the repository. To set up pre-commit in a fresh repository clone, please run:

pre-commit install

This only needs to do once and ensures code formatting & checks are performed on each commit.

Set up a development environment via conda#

If you use Conda, you can install requirements into a conda environment using the environment.yml file included in the dev subfolder of the source repository.

Using the package has the following dependencies which will be installed automatically via pip:

torch: Consider pre-installing if you have specific system requirements (CPU / GPU / CUDA version).
scipy: We use some statistical helper functions to calculate metrics.
torchmetrics: We use some statistical helper functions to calculate metrics.

To run the tests and example notebooks, you need to install the following additional packages:

To use conda to install all package dependencies locally and start development, run the following commands in the root directory of the repository.

Create the environment:

conda create -y -n torchsurv python=3.10
conda env update -n torchsurv -f dev/environment.yml

Activate the environment:

conda activate torchsurv

Test and develop the package#

To run all unit tests either use dev/run-unittests.sh or run the following command from the repository root directory:

PYTHONPATH=src python -m unittest discover -s tests -v

To run other scripts within the repo, you can run the following from the top level directory of the repository clone, and then run scripts from there that import TorchSurv. Alternatively you may give an absolute path to the source directory.

export PYTHONPATH=src

To test building the package, run:

python -m build # builds the package as a tarball -> dist

To get a local installation for testing e.g. other tools that depend on TorchSurv with a local development version of the code:

pip install -e <repo-clone-directory>  # install for development ("editable")

Installing a local package build#

To install a specific package build e.g. into a local conda environment / virtualenv:

python -mbuild
# ... activate virtualenv
# install the first tarball in dist
pip install "$(set -- dist/*.tar.gz; echo "$1")"

Code formatting#

We use black for code formatting. To format the code, run the following command:

./dev/codeqc.sh

Code should be properly formatted before merging with dev or main.

Build and test the documentation locally#

We use Sphinx to build the documentation. To build and view the documentation, run the following script:

# just run ./dev/build-docs.sh to build but not start a webserver
./dev/build-docs.sh serve

You can then access the documentation at http://localhost:8000 in your web browser.

When updating toctrees / adding content you may need to clean your local documentation build:

cd docs
make clean

To build a single PDF of the documentation, run:

cd docs
make latexpdf LATEXMKOPTS="-silent -f"

There are a few errors in the PDF build due to layout issues, but the PDF can still be used to summarize the package in a single file.

Steps to create a new release#

1. On pypi#

Follow the steps below, ensure each one is successful.

Update the version number in pyproject.toml.
Ensure CHANGELOG.md is up-to-date with the new version number and changes.
Ensure all PRs to be included are merged with main, pushed to Github and that all tests & checks have run successfully.
Build the release from the latest main branch: git checkout main && git pull && rm -rf dist && python -m build
Check the built package: python -m twine check dist/*
Upload the package to testPyPI: python -m twine upload -r testpypi dist/*

Check if the package can be installed (also check this installs the correct version):

rm -rf testenv # ensure a new test env is created
python -m virtualenv testenv
. ./testenv/bin/activate
# these don't simply install from testpypi
pip install torch torchmetrics scipy numpy
pip install -i https://test.pypi.org/simple/ torchsurv
python
>>> from torchsurv.loss import cox
>>> from torchsurv.metrics.cindex import ConcordanceIndex

Upload to pypi: python -m twine upload -r pypi dist/*

Check that the package can be installed:

rm -rf testenv # ensure a new test env is created
python -m virtualenv testenv
. ./testenv/bin/activate
pip install torchsurv
python
>>> from torchsurv.loss import cox
>>> from torchsurv.metrics.cindex import ConcordanceIndex

Create a new tag for the release, e.g. git tag -a v0.1.3 -m "Version 0.1.3". Push the tag to Github: git push origin v0.1.3. Create a release on Github from the tag via Novartis/torchsurv.

2. Update on conda-forge#

Create a fork of conda-forge/torchsurv-feedstock
Create a branch on this fork, and update the version number in recipe/meta.yaml as well as the sha sum (obtained via shasum -a 256 dist/torchsurv-<version>.tar.gz from the file we uploaded to pypi)
Finish the checklist in the PR, ensure it builds.
Double-check and merge the PR.

The new release should become available some time after this.

Development notes

Contents