add python 3.11 support #3756
add python 3.11 support #3756
Conversation
|
numba <0.58 is not supported on python 3.11. |
|
Hi there, first things first: Thanks a stack for conceiving and maintaining PyCaret, and for working on this patch. We have been tripped by missing support for Python 3.11 when trying to bring in support for MLflow on our CI. In this spirit, we wanted to probe if corresponding support would be around the corner, or if you would like to see us supporting you on this matter, when possible. With kind regards, |
|
Hi @amotl, we plan to support 3.11 with next release but we have some issues with dependencies. All help is welcome. |
|
Dear Mavs, we have been able to work on a few patches to support you on this matter. With kind regards, Python 3.11 on LinuxGH-3813 might support this patch a bit. Apparently, more work needs to be done to fully support Python 3.11 on all platorms, as subsequent PRs will demonstrate. But at least, it's a start, and may already satisfy a good share of users, when combined with this patch, and published with an upcoming release. Python 3.11 on all workflowsCI job #6763789876 is currently running all workflows with corresponding adjustments to add Python 3.11 on behalf of crate-workbench/pycaret@python311-linux...crate-workbench:pycaret:python311-more. |
It wants to install `numpy==1.21.3`, for whatever reason, but that version does not exist. ERROR: Ignored the following versions that require a different python version: 1.21.2 Requires-Python >=3.7,<3.11; 1.21.3 Requires-Python >=3.7,<3.11; 1.21.4 Requires-Python >=3.7,<3.11; 1.21.5 Requires-Python >=3.7,<3.11; 1.21.6 Requires-Python >=3.7,<3.11 ERROR: Could not find a version that satisfies the requirement numpy==1.21.3
Fuse GHA/pytest configurations into single workflow recipe.
|
Hi again, we cleaned up our act, and submitted two followup patches to conclude Python 3.11 support. With kind regards, |
More Python 3.11 support
|
Thank you for integrating our patches. 💯 |
.github/workflows/test.yml
Outdated
| # remove-item tests/benchmarks/* | ||
| # - name: Test with pytest | ||
| # run: pytest --durations=0 | ||
| run: pytest --durations=0 -m "not (benchmark or plotting or tuning_random)" |
There was a problem hiding this comment.
Maybe rename tuning_random to ts_tuning as it is specific to time series only.
There was a problem hiding this comment.
OK, I see that you have separated out random and grid. So maybe add grid in addition to random here also.
| @@ -19,6 +20,9 @@ | |||
| pytestmark = pytest.mark.filterwarnings("ignore::UserWarning") | |||
| os.environ["PYCARET_TESTING"] = "1" | |||
|
|
|||
| if sys.platform == "win32": | |||
There was a problem hiding this comment.
Why are we skipping these on Windows?
There was a problem hiding this comment.
I think those are the corresponding references.
Describe the changes you've made
Add support for python 3.11.
Additionnaly, fix a bug in
test_classification_custom_metric.Type of change
Please delete options that are not relevant.