Closed eagmon closed 2 years ago
Since this is a breaking change, should it go in a 2.0 release? I think there are other breaking changes we've been wanting to make--should those all go in together?
@U8NWXD -- This could go into a 2.0 release, but it alone would not warrant a 2.0 release. What is the best practice for working on a number of changes and then releasing them all at once as a new version? Maybe this should go into a dev branch?
I looked into this and proposed a workflow in https://github.com/vivarium-collective/vivarium-core/issues/194 for developing v2 while still maintaining v1
@U8NWXD -- I'm in support of your proposal for v2. But to put that off for a bit I brought composition.py
back in this PR with a deprecation warning. So the main change for this PR is that Vivarium-core no longer uses composition.py
, even though it is still available for dependent libraries. This will make it easier to remove when the time comes.
This removes vivarium-core dependence from
composition.py
, and adds a deprecation warning. This file holds a collection of helper functions for bringing processes together. In the last yearcomposite.merge
andEngine
have been made much easier to use directly instead of relying on helper functions. Because of thiscomposition.py
is no longer helpful and in some cases made upkeep harder.directories.py
now has a few standard directories --PROCESS_OUT_DIR
,COMPOSITE_OUT_DIR
, etc.This will be a breaking change and I expect many dependent libraries will have to update their code to remove the helper functions. Some tips for this:
Instead of
simulate_process
:instead of
simulate_composer
:adding a timeseries process:
By creating this pull request, I agree to the Contributor License Agreement, which is available in
CLA.md
at the top level of this repository.