prevent float overflow when calculating He I population

[DeerWhale]

📝 Description

When the T_radiative in the shell is below 404 K, the highlighted part in the screenshot below will return Inf due to float overflow, which cause the nan values in the calculation of electron density.

This PR initially aims to prevent the overflow by capping the value to be the max float value allowed in the program. However a better way to do this maybe raise an error when the input model T_radiative is below certain threshold (1000K is chosen in this PR), so the user can double check the input model.

Fun fact, the low T_rad also cause kernel crash when using the macroatom setting.

📌 Resources

Examples, notebooks, and links to useful references.

🚦 Testing

How did you test these changes?

• Testing pipeline
• Other method (describe)
• My changes can't be tested (explain why)

☑️ Checklist

• I requested two reviewers for this pull request
• I updated the documentation according to my changes
• I built the documentation by applying the build_docs label

Note: If you are not allowed to perform any of these actions, ping (@) a contributor.

[codecov]

Codecov Report

Attention: Patch coverage is 50.00000% with 1 lines in your changes are missing coverage. Please review.

Project coverage is 68.73%. Comparing base (64df98b) to head (5706282).
Report is 1 commits behind head on master.

Files	Patch %	Lines
tardis/model/parse_input.py	50.00%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #2522      +/-   ##
==========================================
- Coverage   68.74%   68.73%   -0.01%     
==========================================
  Files         165      165              
  Lines       13999    14001       +2     
==========================================
+ Hits         9623     9624       +1     
- Misses       4376     4377       +1     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[andrewfullard]

Good find. My issue with this solution is that nothing in a TARDIS simulation should ever be that cold!

[DeerWhale]

Good find. My issue with this solution is that nothing in a TARDIS simulation should ever be that cold!

The issue only occurs in the most outer layer in some of the cases (in some of the models there are shells >40k km/s with low T_rad ).

Alternative is printing out a warning state why the overflow is encountered there, so the user can modify their models accordingly with that Info?

[andrewfullard]

I would suggest a combination of both. A warning that the n_electron calculation has overflowed, telling the user that their temperatures are too low (with a printout of the problem temperature value and possibly the shells that are affected)

[tardis-bot]

*beep* *bop*

Hi, human.

The docs workflow has succeeded ✔️

Click here to see your results.

[andrewfullard]

A warning is perhaps optimistic but we shall see how it goes