To work around #6327

[arafune]

While this PR is not still sufficient to work around #6327, such revision must be required at least.

The essential point of the change is the following part.

•        kdim_len = len(kdim_param.default) if kdims is None else len(kdims)
  

•        if kdims is None:
  

•            kdims = list(data.dims)
  

•        kdim_len = len(kdims)
  

While reading the code, I found the following part difficult to understand:

if not isinstance(data, (xr.Dataset, xr.DataArray)):

To improve readability, I changed it to a more straightforward approach. I hope you will accept this change as well.

[arafune]

Do you need further information? Why is there no response at all?

[ahuang11]

Thanks for creating this PR.

While this PR is not still sufficient to work around #6327, such revision must be required at least.

Can you elaborate a bit on this? What else is needed?

Taking a glance at the code changes, can you explain what the changes aim to fix?

[arafune]

Thanks for creating this PR.

While this PR is not still sufficient to work around #6327, such revision must be required at least.

Can you elaborate a bit on this? What else is needed?

Sorry for the confusion. I have checked against a simple Xarray object and as far as I can tell the problem is solved, just not 100% sure.

Taking a glance at the code changes, can you explain what the changes aim to fix?

I think yes. The change is really tiny.
In the current version, the assignment to the X- and Y-axis is based on the coords object, but coords is not a sequence, which means that the order is not constant.

[arafune]

What is the further information to accept this PR?

[hoxbro]

Currently, tests are failing because of your changes. Some seem trivial fixes (moving order around), while others do not.

Making it so the tests pass will be the next step.

[arafune]

@hoxbro
Thank you for your comment.

Unfortunately, different from the result of GitHub Actions, most of the tests have passed on my local system (Mac).

Only the fault is

, which I believe is not related to my change. Probably I have missed something, but I couldn't find what I should do. Could you let me know what should I do, if you can guess?

[hoxbro]

You are only running the core-test; try running the full test suite with pixi run -e test-312 test-unit.

Also, make sure you have pulled in the latest changes made to this PR, e.g., I merged with the main branch to fix some compatibility issues.

[arafune]

@hoxbro

Thank you for your reply.

Unfortunately, I cannot still go well. (I did pixi install and pixi run download-data, they worked well.)

Can -e test-312 really work on Mac ? Or I have sitll missed something?

[hoxbro]

Yes, they should be able to run on MacOS, but it seems like a bad file. Try rm -rf .pixi pixi.lock and then pixi run again.

Are you running on an ARM Mac? Or is it an Intel?

[arafune]

@hoxbro

Thank you for your help.

After removing .pixi and pixi.lock, pixie run works well.
I'll try to revise the PR. But the current simple PR would not sufficiently good.
Thus, I would like to close this PR. After revision (with test on my system) I'll request the PR once again.

Thank you.

[github-actions]

This pull request has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs.