Fix Profile1DExport storing identical profiles by appending view inst…

[hhy2022]

Proposed changes

Fix Profile1DExport storing identical profiles by appending view instead of copy

This PR fixes a bug in Profile1DExport where profile data were stored as views of the same underlying array rather than independent copies. As a result, all exported profiles at different timesteps ended up containing identical values, which corresponding to the last computed timestep, regardless of the requested export times.

This bug was discussed in Issue #1046
Fixes #1046.

Types of changes

What types of changes does your code introduce to FESTIM?

• Bugfix (non-breaking change which fixes an issue)
• New tests

Checklist

• Black formatted
• Unit tests pass locally with my changes
• I have added tests that prove my fix is effective or that my feature works

Further comments

This fix is small but crucial for all workflows relying on Profile1DExport, including:

• Transient diffusion problems
• Post-processing using stored spatial profiles
• Validation tasks requiring comparison between profiles at different times

[RemDelaporteMathurin]

@hhy2022 there seem to be some conflicts with your PR. Let me know if you need help resolving them

[RemDelaporteMathurin]

Thanks for this quick fix @hhy2022 !

You need to resolve the conflicts with the fenicsx branch and then make sure your test is 1) really testing something and catching the bug (ie. failing without your fix, with an assert) 2) encapsulated inside a python function

[RemDelaporteMathurin]

Tests need to be inside functions:

def test_profile1dexport():
    ...
    ...
    assert profile.data[1] != profile.data[0]

[hhy2022]

Thanks for this quick fix @hhy2022 !

You need to resolve the conflicts with the fenicsx branch and then make sure your test is 1) really testing something and catching the bug (ie. failing without your fix, with an assert) 2) encapsulated inside a python function

Thank you for your guidance. This issue first appeared when I was working with the v2.0-beta release for the V-V book. I updated to the latest fenicsx branch and confirmed that the issue has already been fixed.
I will try to use the latest version from the fenicsx branch in V-V report.

[hhy2022]

Closing since it was fixed by #1042