It creates conda environments for Linux and Mac systems and provides these environments as environment.yml file and installers. Every new commit to the main branch results in a new release with the date as name. The contributors can use the environment.yml files to update their local environments in case the requirements changed during the process of developing the workflow.
Only MacOs and Linux are supported and ARM based systems are not yet supported.
Only one release per day is supported, if the users want to release more than one version per day, then it is necessary to remove the previous release before.
Dependabot is raising updates based on the pip requirements.txt file but does not update the conda environment.
So while I would not claim the tool is production ready, I already use it to keep my conda environment synchronised over multiple systems.
With the conda perspective in mind, I thought about the challenge of having the same packages and dependencies installed on the hardware of all contributors. So based on our work on https://github.com/pyiron/docker-stacks and https://github.com/pyiron/pyiron-installer I created: https://github.com/jan-janssen/shared-conda-environment
It creates conda environments for Linux and Mac systems and provides these environments as
environment.ymlfile and installers. Every new commit to the main branch results in a new release with the date as name. The contributors can use theenvironment.ymlfiles to update their local environments in case the requirements changed during the process of developing the workflow.There are a number of current limitations, the primary ones are:
requirements.txtfile but does not update the conda environment.So while I would not claim the tool is production ready, I already use it to keep my conda environment synchronised over multiple systems.