Montecarlo configuration refactor, Numba developer proposed version

[andrewfullard]

📝 Description

Type: 🪲 bugfix

This is the version of the refactor that uses the methods proposed here: https://hackmd.io/Wmy-e2RsQCSqjJsP0VAmbg?both which allow us to send config objects to different functions with smart compilation that accounts for changes in the config while retaining the performance benefits of skipping paths (this effectively compiles JIT functions to each config).

🚦 Testing

How did you test these changes?

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

[codecov]

Codecov Report

Attention: Patch coverage is 60.59113% with 80 lines in your changes are missing coverage. Please review.

Project coverage is 67.99%. Comparing base (328ec77) to head (8224fdc).
Report is 2 commits behind head on master.

❗ Current head 8224fdc differs from pull request most recent head 6d06b34. Consider uploading reports for the commit 6d06b34 to get more accurate results

Files	Patch %	Lines
.../montecarlo/estimators/radfield_estimator_calcs.py	39.28%	17 Missing ⚠️
.../montecarlo/montecarlo_numba/single_packet_loop.py	5.55%	17 Missing ⚠️
tardis/montecarlo/montecarlo_numba/base.py	0.00%	9 Missing ⚠️
tardis/montecarlo/montecarlo_numba/interaction.py	38.46%	8 Missing ⚠️
tardis/montecarlo/montecarlo_numba/vpacket.py	0.00%	8 Missing ⚠️
tardis/transport/r_packet_transport.py	28.57%	5 Missing ⚠️
...is/montecarlo/estimators/radfield_mc_estimators.py	42.85%	4 Missing ⚠️
tardis/montecarlo/montecarlo_configuration.py	95.16%	3 Missing ⚠️
tardis/montecarlo/tests/test_montecarlo.py	0.00%	3 Missing ⚠️
tardis/simulation/base.py	60.00%	2 Missing ⚠️
... and 3 more

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #2568      +/-   ##
==========================================
+ Coverage   67.18%   67.99%   +0.80%     
==========================================
  Files         173      171       -2     
  Lines       14570    14259     -311     
==========================================
- Hits         9789     9695      -94     
+ Misses       4781     4564     -217     

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

[andrewfullard]

Total test duration (Moria): 683s
Total test duration #2525 (Moria): 622s

These times are quite reliable and reproducible.

This is likely because the obj2strdict is being called in the single packet loop. I believe to get the best benefit of this method would require splitting our physics functions up wherever they have if-statements based on the configuration.

[andrewfullard]

need to rebase