Fix GUI and plasma tests

[epassaro]

Description

Plasma tests are fixed by removing the following lines in tardis/io/atom_data/base.py:

# Sort the dataframe before indexing to improve performance
lines.sort_index(inplace=True)

GUI test fixed by adding an if statement before importing modules that requires QT_API:

if 'QT_API' in os.environ:
    from tardis.gui.widgets import Tardis
    from tardis.gui.datahandler import SimpleTableModel

Then, forced to skip test_gui.py by removing export QT_API=pyqt from simple_test_framework.yml (we should re-activate this test on another PR).

Motivation and Context

Sorting introduced in #1141 makes plasma tests fail, but went unnoticed because a bug introduced in #1068 (fixed by #1162).

Combined with the failing test_gui.py the pipeline became hard to debug.

How Has This Been Tested?

1. Go to the list of merged PRs: https://github.com/tardis-sn/tardis/pulls?q=is%3Apr+is%3Amerged.
2. Run plasma tests for every recent PR and bisect the bug:

git checkout upstream/pr/1141
pytest tardis/plasma --tardis-refdata=/path/to/refdata

3. Apply the above mentioned changes for the GUI and run:

pytest tardis --tardis-refdata=/path/to/refdata

4. All tests must pass now.

Screenshots (if appropriate):

Types of changes

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)

Checklist:

• My code follows the code style of this project.
• My change requires a change to the documentation.
• I have updated the documentation accordingly.
• I have assigned/requested two reviewers for this pull request.

[ebjordi]

Excellent

[epassaro]

Sorry, YAML was formatted by VSCode. The only change here is the removal of the export QT_API=pyqt line.

[Rodot-]

This is fine, could you make a note of it in the PR so no one in the future is confused why this looks like such a big change?

[Rodot-]

These changes look fine, though they seem to really just be masking the issues with the gui, but I suppose fixing those is another problem.

[marxwillia]

Tests pass, new code looks clean. Excellent catch on the bug introduced by sorting the lines in atom_data/base.py. Future optimization attempts should keep this in mind.