* fix issue #1042 and also test openfecli for API breaks
* just check against main instead of last tag
* change to run on PR so I can test the changes
* test cli api change
* undo change that didn't break api
* change default that will trigger API change
* undo api change
* remove comments if api change check passes
* fix reference error
* break api
* add a message if there is no API break
* fix api break
* Switch back to pull request target now that testing is done
* Update gpu-runner.yaml
* test on a g4dn.xlarge @ 53 cents an hour
* testing on a p2.xlarge @ 90 cents an hour
* Update gpu-runner.yaml
* Update gpu-runner.yaml
* update AMI to correct region (still the 'Deep Learning Base OSS Nvidia Driver GPU AMI' AMI)
* fix default shell, yolo that there is no syntax issues
* use the _REAL_ correct default shell
* Add a timeout of 12 hrs instead of the default 6
* Added oe toolkit + espaloma_charge to GPU env for testing
* fix channel name
* got a download error on the cudatoolkit shim which we don't want anyway
---------
Co-authored-by: Mike Henry <11765982+mikemhenry@users.noreply.github.com>
* conda cron job canary
* add conda cron job
* Add dispatch trigger
* 12 not 13 for macos
* Update conda_cron.yaml
* Update conda_cron.yaml
* make sure we install latest version of openfe
* install latest version found on conda-forge
* bash syntax? don't know her
* openfe latest not openmm
* use some bash PE tricks
* fix typo in micromamba cmd
* update codecov action
* removed run on each PR
---------
Co-authored-by: Mike Henry <11765982+mikemhenry@users.noreply.github.com>
* see if this is enough to test the removal of gaff
* see if now we get the right error message
* rerun ci
* rerun ci
* see what happens if we the solver do its thing
* add omff version to matrix
* Added note about how to use gaff
* depending on the ommff version, we make sure the test does what we expect
* add an osx arm test since we should be able to do that now
* see if we can install missing stubs
* add missing stubs directly
* we return a sampler and not a system for the rfe
* Apply suggestions from code review
Co-authored-by: Hannah Baumann <43765638+hannahbaumann@users.noreply.github.com>
* added link to give some context
* testing on python 3.12 now
* unpin numpy
* no openeye toolkits for 3.12
---------
Co-authored-by: Hannah Baumann <43765638+hannahbaumann@users.noreply.github.com>
* debug script for people to test openmm
* Added news entry
* added checking which platforms can do mixed p
* fix typo in debug script
* added how to run script
* point to new showcase
* give full path to assests to avoid copying notebooking into build env
* rerun ci
* fix mypy env
* looks like we still need to pin mypy but can relax python
---------
Co-authored-by: Irfan Alibay <IAlibay@users.noreply.github.com>
