Support Python 3.14

[mikemhenry]

Tips

• Comment "pre-commit.ci autofix" to have pre-commit.ci atomically format your PR.
  Since this will create a commit, it is best to make this comment when you are finished with your work.

Checklist

• Added a news entry

Developers certificate of origin

• I certify that this contribution is covered by the MIT License here and the Developer Certificate of Origin at https://developercertificate.org/.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 98.54%. Comparing base (e15ae9b) to head (2908298).
⚠️ Report is 1 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #796   +/-   ##
=======================================
  Coverage   98.54%   98.54%           
=======================================
  Files          43       43           
  Lines        2810     2811    +1     
=======================================
+ Hits         2769     2770    +1     
  Misses         41       41           

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[mikemhenry]

Getting these errors now:

FAILED gufe/tests/test_models.py::TestBoxVectors::test_valid_box_quantity[value0-m_expected0] - pydantic_core._pydantic_core.ValidationError: 1 validation error for BoxSettingsModel
box_vectors
  Extra inputs are not permitted [type=extra_forbidden, input_value=<Quantity([[0. 0. 1.]

We are just missing ConfigDict(extra='ignore') in our dummy settings.

[mikemhenry]

Ah our old friend:

FAILED gufe/tests/test_serialization_json.py::TestPathCodec::test_default - AssertionError: assert {':is_custom:...h': 'foo/bar'} == {':is_custom:...h': 'foo/bar'}
  
  Omitting 3 identical items, use -vv to show
  Differing items:
  {'__module__': 'pathlib'} != {'__module__': 'pathlib._local'}
  
  Full diff:
    {
        ':is_custom:': True,
        '__class__': 'PosixPath',
  -     '__module__': 'pathlib._local',
  ?                           -------
  +     '__module__': 'pathlib',
        'path': 'foo/bar',
    }

[atravitz]

Ah our old friend:

FAILED gufe/tests/test_serialization_json.py::TestPathCodec::test_default - AssertionError: assert {':is_custom:...h': 'foo/bar'} == {':is_custom:...h': 'foo/bar'}
  
  Omitting 3 identical items, use -vv to show
  Differing items:
  {'__module__': 'pathlib'} != {'__module__': 'pathlib._local'}
  
  Full diff:
    {
        ':is_custom:': True,
        '__class__': 'PosixPath',
  -     '__module__': 'pathlib._local',
  ?                           -------
  +     '__module__': 'pathlib',
        'path': 'foo/bar',
    }

First battle with this was here: https://github.com/OpenFreeEnergy/gufe/pull/577/changes

If I understand correctly, pathlib._local is only a python 3.13 thing? In which case we should be able to change this line to if sys.version_info[:2] == (3, 13):

[mikemhenry]

You are correct, we need to fix that test since this is the current situation:

Runtime writing JSON	observed module
Python ≤3.12	pathlib
Python 3.13	pathlib._local
Python 3.14	pathlib

For the test, we can just get the module we need at runtime:

self.dcts = [
    {
        "__class__": "PosixPath",
        "__module__": pathlib.PosixPath.__module__,
        ":is_custom:": True,
        "path": "foo/bar",
    }
]

We can also set up the codec to just use the public module name, something like:

PATH_CODEC = JSONCodec(
    cls=pathlib.PosixPath,
    to_dict=lambda p: {
        "__module__": "pathlib",
        "__class__": "PosixPath",
        ":is_custom:": True,
        "path": str(p),
    },
    from_dict=lambda dct: pathlib.PosixPath(dct["path"]),
    is_my_dict=is_legacy_path_dict,
)

Then the reader will accept ("pathlib", "pathlib._local") but we will always write pathlib even on python 3.13

[atravitz]

Then the reader will accept ("pathlib", "pathlib._local") but we will always write pathlib even on python 3.13

We originally decided not to do this because we anticipated pathlib._local to be the behavior for the future and for pathlib to be legacy, but that turned out not to be the case.

I'm typically inclined to keep the code as-is so that we're not arbitrarily changing things, but since the 3.13 behavior was immediately reverted in 3.14, I'm fine with updating the code to treat 3.13 as an anomaly that we're accounting for and update the codec.

tl;dr: yeah do it 🚀

[mikemhenry]

Sweet, looking at the issues and mailing lists, it looks like it was a mistake that happened during a re-factor which is why 3.13 is an anomaly

tl;dr 🚀

[mikemhenry]

The API check bot is having an issue, will fix later ImportError: With sys.path = ['/tmp/griffe-worktree-gufe-origin-main-e4757g4_/origin-main/src/gufe'], accessing 'gufe' raises ModuleNotFoundError: No module named 'gufe'

[atravitz]

some questions

[atravitz]

just one docstring correction, then good to merge!

[github-actions]

🚨 API breaking changes detected! 🚨
View logs for this step