Plugin for the Coldfront HPC management platform that provides customisations for Imperial College London.
As this repository is a plugin it must be used in conjunction with a Coldfront instance. Limited development is possible working with this repository in isolation via Unit Tests and QA tooling. For interactive testing in the browser, a dedicated development environment is provided by [GitHub: ImperialCollegeLondon/coldfront_development_environment] that contains a Coldfront environment with the necessary configuration.
This repository is a reusable Django app and so follows the associated conventions. See Django Advanced Tutorial: How to write reusable apps for further information.
This is a Python application that uses [pip-tools
] for packaging and dependency management. It also provides pre-commit
hooks (for for ruff and mypy
) and automated tests using pytest
and GitHub Actions. Pre-commit hooks are automatically kept updated with a dedicated GitHub Action, this can be removed and replace with pre-commit.ci if using an public repo. It was developed by the Imperial College Research Computing Service.
[pip-tools
] is chosen as a lightweight dependency manager that adheres to the latest standards using pyproject.toml
.
To add or remove dependencies:
dependencies
variables in the pyproject.toml
file (aim to keep development tools separate from the project requirements).pip-compile
for requirements.txt
- the project requirements.pip-compile --extra dev -o dev-requirements.txt
for the development requirements.pip-compile --extra doc -o doc-requirements.txt
for the documentation tools.pip-sync *requirements.txt
To upgrade pinned versions, use the --upgrade
flag with pip-compile
.
Versions can be restricted from updating within the pyproject.toml
using standard python package version specifiers, i.e. "black<23"
or "pip-tools!=6.12.2"
To get started:
Note: It is important that this environment is created with Python 3.11
. It cannot be any higher or lower Python version due to coldfront
limitations at the time of writing. You may note we ensure this requirement in the pyproject.toml
file by requires-python = "~=3.11"
, but wish to mention it again explicitly here.
Taking care of the note, we can then create a virtual environment from the command line like so:
python -m venv .venv
source .venv/bin/activate # with Powershell on Windows: `.venv\Scripts\Activate.ps1`
Install development requirements:
pip install -r dev-requirements.txt
Install the git hooks:
pre-commit install
As a standalone Django application this plugin provides models to extend the Coldfront data model. A helper script is provided to generate migrations for these models. To use it, run the following command:
python makemigrations.py
This is equivalent to running python manage.py makemigrations
in a Django
project and all the same options are available. See options with:
python makemigrations.py --help