Line opacity parallelization

[jvshields]

This expands PR #178 to allow for parallelization of the slowest part of the code. It should not be merged before that pull request.

[codecov]

Codecov Report

Attention: Patch coverage is 48.07692% with 27 lines in your changes are missing coverage. Please review.

Project coverage is 67.91%. Comparing base (ed2a670) to head (e3ab84a).

Files	Patch %	Lines
...adiation_field/opacities/opacities_solvers/base.py	41.37%	17 Missing ⚠️
stardis/base.py	57.89%	8 Missing ⚠️
...is/radiation_field/radiation_field_solvers/base.py	50.00%	2 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #179      +/-   ##
==========================================
- Coverage   69.28%   67.91%   -1.37%     
==========================================
  Files          31       31              
  Lines        1172     1203      +31     
==========================================
+ Hits          812      817       +5     
- Misses        360      386      +26     

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

[tardis-bot]

*beep* *bop*
Hi human,
I ran benchmarks as you asked comparing main (ed2a670) and the latest commit (e3ab84a).
Here are the logs produced by ASV.
Results can also be downloaded as artifacts here.
Significantly changed benchmarks:

       before           after         ratio
   [ed2a6705]       [e3ab84a5]
-      1.16±0.01s          973±4ms     0.84  run_stardis.Sim100AA.time_calc_alpha_line_at_nu
-         241±4ms          216±5ms     0.90  run_stardis.Sim10AA.time_calc_alpha_line_at_nu

All benchmarks:

All benchmarks:

     before           after         ratio
   [ed2a6705]       [e3ab84a5]
    1.12±0.01ms         1.10±0ms     0.99  run_stardis.Sim100AA.time_calc_alpha_file
-      1.16±0.01s          973±4ms     0.84  run_stardis.Sim100AA.time_calc_alpha_line_at_nu
        825±4ms         828±10ms     1.00  run_stardis.Sim100AA.time_raytrace
        4.59±0s       4.46±0.05s     0.97  run_stardis.Sim100AA.time_run_stardis
       989±10μs      1.03±0.03ms     1.04  run_stardis.Sim10AA.time_calc_alpha_file
-         241±4ms          216±5ms     0.90  run_stardis.Sim10AA.time_calc_alpha_line_at_nu
        683±2ms          675±4ms     0.99  run_stardis.Sim10AA.time_create_plasma
       56.3±3ms         55.0±2ms     0.98  run_stardis.Sim10AA.time_raytrace
        1.88±0s       1.89±0.01s     1.01  run_stardis.Sim10AA.time_run_stardis

[DeerWhale]

Would it be better to make the tracing_lambas a parameter initializing the class so you don't need to repeat the rest of code in a seperate class?

[jvshields]

I considered it, but I think having a specific class be a different simulation is easier to understand and make changes to for somebody that knows less about the code in the future. Like you could just build a different RadiationField class with different tracing_lambdas arrays, without resolving the plasma or the model, but exactly where and how to do that and what parts feed in to what requires a little bit of care.

In this way, if somebody wants to change the benchmarks in the future, there's a pretty clear template, and it's more straightforward to follow the logic of the code.

[andrewfullard]

Looks great!