Add newest python version 3.13

[ludwiglierhammer]

No description provided.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 82.48%. Comparing base (2a185c8) to head (73a4491).
Report is 32 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #157   +/-   ##
=======================================
  Coverage   82.48%   82.48%           
=======================================
  Files          69       69           
  Lines        2655     2655           
=======================================
  Hits         2190     2190           
  Misses        465      465           

Flag	Coverage Δ
unittests	82.48% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

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

---

🚨 Try these New Features:

• Flaky Tests Detection - Detect and resolve failed and flaky tests

[jtsiddons]

@ludwiglierhammer: Is it numba causing problems?

[github-actions]

Warning

This Pull Request modifies GitHub Workflows and is coming from a fork.
It is very important for the reviewer to ensure that the workflow changes are appropriate.

[ludwiglierhammer]

@ludwiglierhammer: Is it numba causing problems?

I think some packages aren't compatible with the newest python version 3.13. This will be fixed in a couple of days/weeks. We just have to wait.

[jtsiddons]

Testing locally, the only dependency that fails to install with pip on python 3.13 is numba.

Building numba wheels from source allows all dependencies to be installed, and the unit tests to pass.

It looks like we're waiting for numba 0.61. It looks like they're close to a release candidate, so hopefully not too long.

[jtsiddons]

Hi @ludwiglierhammer: numba is now at version v0.61 which supports python 3.13. Can now build cdm_reader_mapper for python 3.13. Merging main allows for the unit tests to run. I get two failing tests (I have trimmed most of the output):

==================================================================================================================================================== test session starts ====================================================================================================================================================
platform darwin -- Python 3.13.1, pytest-8.3.4, pluggy-1.5.0
rootdir: /private/var/folders/vf/pskk3w4j38l8kk7bc9xm07j00000gp/T/cdm
configfile: pyproject.toml
collected 68 items

tests/test_databundle.py .......                                                                                                                                                                                                                                                                                      [ 10%]
tests/test_duplicates.py ...........                                                                                                                                                                                                                                                                                  [ 26%]
tests/test_operations.py .........                                                                                                                                                                                                                                                                                    [ 39%]
tests/test_workflow.py ...................F........F.....                                                                                                                                                                                                                                                             [ 89%]
tests/test_writers.py .......                                                                                                                                                                                                                                                                                         [100%]

****** TRIMMED by jtsiddons ******
================================================================================================================================================== short test summary info ==================================================================================================================================================
FAILED tests/test_workflow.py::test_read_data[gcc-None-None-True-None-mdf_kwargs19] - KeyError: ('header', 'record_timestamp')
FAILED tests/test_workflow.py::test_read_data[gcc-None-None-True-drops28-mdf_kwargs28] - KeyError: ('header', 'record_timestamp')
=================================================================================================================================== 2 failed, 66 passed, 25 warnings in 193.74s (0:03:13) ===================================================================================================================================

I can send you the full pytest output if you want it.

Note: I also get a number of warnings indicating checksums do not match:

UserWarning: TRIMMED/cdm-testdata/icoads/r300/d781/output/mask_icoads_r300_d781_1987-09-01_subset.csv and md5 checksum do not match.

[github-actions]

Warning

This Pull Request modifies GitHub Workflows and is coming from a fork.
It is very important for the reviewer to ensure that the workflow changes are appropriate.

[ludwiglierhammer]

@jtsiddons: thanks for the info. I'll merge this PR into the main branch. Do you think we need support for py9 anymore?
The warnings indicating checksums do not match originate from my last changes in the cdm-testdata (v2025.01.16 and v2025.01.17). No worries 😉

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 83.19%. Comparing base (67c9426) to head (fdd269a).
Report is 3 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #157   +/-   ##
=======================================
  Coverage   83.19%   83.19%           
=======================================
  Files          79       79           
  Lines        2958     2958           
=======================================
  Hits         2461     2461           
  Misses        497      497           

Flag	Coverage Δ
unittests	83.19% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

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

[jtsiddons]

Do you think we need support for py9 anymore?

Python 3.9 is end of support in October 2025 I believe. We don't use 3.9 for any of our work here.

Ok. Then I think I wait till the official end of support.

[ludwiglierhammer]

@jtsiddons: There are still some issues with py13 creating a conda environment: https://github.com/glamod/cdm_reader_mapper/actions/runs/12924665583/job/36044081420?pr=157

[github-actions]

Warning

This Pull Request modifies GitHub Workflows and is coming from a fork.
It is very important for the reviewer to ensure that the workflow changes are appropriate.

[jtsiddons]

I wonder if numba 0.61 is not available through conda yet then. I have similar issues building with uv - it fails to install a numba version compatitble with datashader. Can you try pinning "numba>=0.61" in pyproject.toml?

[github-actions]

Warning

This Pull Request modifies GitHub Workflows and is coming from a fork.
It is very important for the reviewer to ensure that the workflow changes are appropriate.

[github-actions]

Warning

This Pull Request modifies GitHub Workflows and is coming from a fork.
It is very important for the reviewer to ensure that the workflow changes are appropriate.

[jtsiddons]

   error    libmamba Could not solve for environment specs
      The following package could not be installed
      └─ numba >=0.61 * does not exist (perhaps a typo or a missing channel).

Updated version not available through microconda then.

[jtsiddons]

I've also noticed that numba 0.61 does not support python 3.9

[ludwiglierhammer]

I think we still have to wait until other packages allow numby>=0.61, don't we?

[jtsiddons]

I think we still have to wait until other packages allow numby>=0.61, don't we?

I think the other dependencies in the tree are all fine with 0.61. This is the version that is installed if you build cdm_reader_mapper with pip for python 3.13.

I can build with uv if I pin numba to 0.61 (and drop 3.9 support) and have no dependency clashes.

[jtsiddons]

We should wait until 0.61 is available through microconda

[jtsiddons]

numba 0.61 is not yet on conda-forge (https://anaconda.org/conda-forge/numba/), however it is available on their own channel (https://anaconda.org/numba/numba) and pypi (https://pypi.org/project/numba).

Can you try specifying numba>=0.61 in the pip dependency section of the two ci yaml files?

[jtsiddons]

@ludwiglierhammer: numba 0.61 is now on conda-forge: https://anaconda.org/conda-forge/numba/files, I would hope the tests should pass (except py3.9...) - hopefully you won't need to specifically pin the numba version (since 0.61 doesn't support python 3.9)

[github-actions]

Warning

This Pull Request modifies GitHub Workflows and is coming from a fork.
It is very important for the reviewer to ensure that the workflow changes are appropriate.

[github-actions]

Warning

This Pull Request modifies GitHub Workflows and is coming from a fork.
It is very important for the reviewer to ensure that the workflow changes are appropriate.

[ludwiglierhammer]

Tests are not running for py13.

[ludwiglierhammer]

This issue is beeing dealt with in #228.