Fix attributes (#27)

[moritz-h]

Fix #27.

• Fix attributes typo
• Set a valid scalar/vector data key as attribute. Just use the first matching key from the list. The check if data is scalar or vector is the same as used in addData()

[codecov-commenter]

Codecov Report

Merging #28 (123185c) into master (f7b5b77) will increase coverage by 0.17%. The diff coverage is 81.25%.

@@            Coverage Diff             @@
##           master      #28      +/-   ##
==========================================
+ Coverage   70.32%   70.50%   +0.17%     
==========================================
  Files           6        6              
  Lines         829      834       +5     
==========================================
+ Hits          583      588       +5     
  Misses        246      246              

Impacted Files	Coverage Δ
pyevtk/vtk.py	85.50% <70.00%> (+0.07%)	:arrow_up:
pyevtk/hl.py	80.69% <100.00%> (+0.30%)	:arrow_up:

---

Continue to review full report at Codecov.

Legend - Click here to learn more Δ = absolute <relative> (impact), ø = not affected, ? = missing data Powered by Codecov. Last update f7b5b77...123185c. Read the comment docs.

[renefritze]

Thank you for the PR @moritz-h! Could you change the type().__name__ checks to use isinstance instead? Otherwise I think this can be merged directly and also warrants a new release.

[moritz-h]

@renefritze I copied the check from within vtkFile.addData(), which is called directly after: https://github.com/pyscience-projects/pyevtk/blob/v1.2.0/pyevtk/vtk.py#L533 I think it would make sense to keep this consistent. Do you prefer to update update both places then or should I avoid changes within vtk.py?

[renefritze]

@renefritze I copied the check from within vtkFile.addData(), which is called directly after: https://github.com/pyscience-projects/pyevtk/blob/v1.2.0/pyevtk/vtk.py#L533

I think it would make sense to keep this consistent. Agreed. Do you prefer to update update both places then or should I avoid changes within vtk.py? I would prefer updating both places, if you don't mind. Maybe the ndarray check too?

[moritz-h]

@renefritze Should be done

[renefritze]

I'll have to look into the apparently broken GH action before I do the next release, and it might be a few days before I get to that.

[moritz-h]

It looks like an authentication issue to pypi. I think the reason might be that third party pull request do not have access to the secret defined in the GitHub workflow. (Which is generally a very good idea from security perspective.)

[renefritze]

It looks like an authentication issue to pypi. I think the reason might be that third party pull request do not have access to the secret defined in the GitHub workflow. (Which is generally a very good idea from security perspective.)

Yeah, my issue is only that I remembered having that setup so it at least wouldn't fail the PR check.

[moritz-h]

I think this line may is wrong: https://github.com/pyscience-projects/pyevtk/blob/master/.github/workflows/test_deploy.yml#L6 The deploy test may only should run when it is merged into master and not onto pull requests.

While wanting to use a static URL above, I noticed, when switching back to v1.2.0, there this line is exactly like what my idea would have been: https://github.com/pyscience-projects/pyevtk/blob/v1.2.0/.github/workflows/test_deploy.yml

Maybe this change to test_deploy.yml within this commit was unintentional and it has worked before: https://github.com/pyscience-projects/pyevtk/commit/571426c4272a93c756b6aae0b8b27bed5cbde160#diff-6b5ff66c8969e24f75b0e7d96ffd86e0939f2e7ef4ed90658d6c054fafcee1fa

[renefritze]

You are absolutely right, that is an entirely unrelated and weird change in 571426c. Good catch! Thanks!

[renefritze]

1.3.0 is now out with your fixes included @moritz-h. Thanks again!