Development#
Development Flow#
To manage new features, bug-fix and release we use Git-Flow. External contributions can be submitted through Github Pull Requests.
For maintainers:
- The master
branch is used for production (i.e., features ready for the public to use).
- The develop
branch is used to merge intermediate features during before releasing to the master
branch.
- All feature/$FEATURE_NAME
branches are used to develop each intermediate feature (many can be used in parallel).
For other developers, it is better to use Github Pull Requests so that the submitted code can be reviewed by maintainers of the repository. The first, step is to fork the github repository (i.e., do not forget to “synchronise” your fork regularly with the main repository), then develop new features the develop
branch and once completed or well advanced submit it to the main repository develop
branch.
Developer Installation#
The installation of the development dependencies (e.g., code formatting, linter, documentation) uses the [dev]
marker. Before cloning the repository from Github, git-lfs
needs to be installed (see more on Installing Git Large File System).
Git Large File System (LFS) is used to store all files which are necessary and not quickly retrievable such as figures derived from tutorials or data for testing.
$ git clone -b develop git@github.com:deephyper/deephyper.git
$ cd deephyper/
$ # Generic installation
$ pip install -e ".[default,dev]"
$ # MacOS (arm64) installation
$ conda env create -f install/environment.dev.macOS.arm64.yml
$ pre-commit install
Warning
When installing the develop
branch of DeepHyper the published documentation may be outdated. In case of trouble, do not hesitate to contact the maintainers of the repository on Slack or Github.
Note
More details about DeepHyper’s optional modules can be found in the Install with pip section.
Code linting and formatting#
Python code should be linted and formatted with Ruff. Linter and formatter checks are performed on the examples and deephyper directories. The linter checks are run with the following command:
$ ruff check examples deephyper
While the formatter checks are run with the following command:
$ ruff format --check examples deephyper
See the Ruff documentation for more information.