Skip to content

Conversation

@mikemhenry
Copy link
Collaborator

No description provided.

@mattwthompson
Copy link
Collaborator

The Python 3.8 failure is cryptic but the 3.10 failures look "correct" to me, and they fail in 10 minutes compared to 30-40, which implies to me OpenEye Toolkits are actually being used

@mikemhenry
Copy link
Collaborator Author

Supersedes #225

@mattwthompson
Copy link
Collaborator

I'm looking to find a more stable and portable way to ship the XML files. I can't for the life of me figure out why the symlinks are going two directories up? It looks like these paths are in the root of the repo.

os.symlink('../../amber/ffxml/', 'openmmforcefields/ffxml/amber')
os.symlink('../../charmm/ffxml/', 'openmmforcefields/ffxml/charmm')
os.symlink('../../amber/gaff', 'openmmforcefields/ffxml/amber/gaff')

@mikemhenry
Copy link
Collaborator Author

I do think the root of a lot of issues is not packing data files correctly, we should probably switch to a modern approach
https://setuptools.pypa.io/en/latest/userguide/datafiles.html

@mattwthompson
Copy link
Collaborator

The thing is, I'm pretty sure the package data is configured properly, or at least uses an approach that's on the better end of current best practices:

include_package_data=True,
# Allows `setup.py test` to work correctly with pytest
setup_requires=[] + pytest_runner,
entry_points={
'console_scripts': [],
'openmm.forcefielddir' : [
'ffxml = openmmforcefields.utils:get_ffxml_path',
],
},
package_data={
'openmmforcefields': ['ffxml/amber/*.xml', 'ffxml/amber/gaff/*.{xml,dat}', 'ffxml/charmm/*.xml']
},

Exactly why symlinks should be used is not clear to me; they're not portable and it saves no space on disk. I've attempted to remove them altogether with #229

Add boilerplate `.gitignore` file

Do not pull from non-standard labels

Install instead of update

Try to figure out why if: {{ false }} still triggers

Syntax?

Tinker

Bring in (full) mamba

Bash thing everywhere

Bring down old toolkit

Do not present un-loadable ff14SB port as available

More pins
@mattwthompson
Copy link
Collaborator

mattwthompson commented Sep 15, 2022

After excessive and inefficient iterations on my part, I think this is ready for review. It includes at least the following changes:

There's more I'd like to do, but this is already borderline unreviewable due to my lack of restraint. I can't do much more than point to that the automation has been improved and tests are passing.

@mattwthompson
Copy link
Collaborator

I can't request a review via the UI since you opened it, so a ping will have to do @mikemhenry

Copy link
Collaborator

@mattwthompson mattwthompson left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I approve (I think?)

@mikemhenry
Copy link
Collaborator Author

I'll review this PR after lunch to just double check it, but it should be good to merge in, like we don't have others we want to merge in first, right? @mattwthompson

@mattwthompson
Copy link
Collaborator

Nope, the other PRs have either been merged into this one or were experiments that didn't work out. It's just this one (for now)!

- name: CodeCov
if: ${{ github.repository == 'openmm/openmmforcefields'
&& github.event != 'schedule' }}
- name: Upload coverage report to CodeCov
Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I will add a quick guard to not upload coverage reports when it is scheduled CI

@mikemhenry
Copy link
Collaborator Author

@mattwthompson I will take over from here! I just have a few things I want to tweak before I get it merged in, again, thanks for all your help!

@mikemhenry mikemhenry enabled auto-merge (squash) September 15, 2022 20:25
@mattwthompson
Copy link
Collaborator

Great!!! Thank you!

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

6 participants