[FIX] Ignore non-significant kappa elbow when no non-significant kappa values exist

[tsalo]

Closes #752. In the ICA decision tree, there is a step where two elbows are calculated: one from all kappa values and one from non-significant kappa values. When there are no non-significant kappa values, this step fails.

Changes proposed in this pull request:

• Ignore the non-significant kappa value-based elbow when there are no non-significant kappa values. Log a warning about this when it comes up.
• For any other cases where getelbow or getelbow_cons fail, raise a more informative exception and include in that message a request to open an issue here.
• Add new smoke test for getelbow and getelbow_cons.

[codecov]

Codecov Report

Merging #760 (e57dcc7) into main (12f16eb) will increase coverage by 0.08%.
The diff coverage is 84.61%.

@@            Coverage Diff             @@
##             main     #760      +/-   ##
==========================================
+ Coverage   92.83%   92.91%   +0.08%     
==========================================
  Files          27       27              
  Lines        2192     2203      +11     
==========================================
+ Hits         2035     2047      +12     
+ Misses        157      156       -1     

Impacted Files	Coverage Δ
tedana/selection/tedica.py	91.61% <77.77%> (-0.96%)	⬇️
tedana/selection/_utils.py	100.00% <100.00%> (+7.14%)	⬆️

---

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 12f16eb...e57dcc7. Read the comment docs.

[jbteves]

LGTM, thanks @tsalo ! Thanks in particular for adding tests as well so that this doesn't regress.

[handwerkerd]

The warning message in the getelbow functions seems a bit more dramatic than necessary, but otherwise fine.
If you feel like changing it, I'd change to "This error happens when a getelbow function is incorrectly called on no components. If you see this message, please open an issue at https://github.com/ME-ICA/tedana/issues with the full traceback and any data necessary to reproduce this error, so that we create additional data checks to prevent this from happening."

[jbteves]

LGTM