Add pounders filtering to tranquilo

[segsell]

• Fix bug in Pounders fitting
• Make Tranquilo more flexible
• Add pounders filtering
• Refactor
• Write more tests

[codecov]

Codecov Report

Merging #389 (74aa852) into main (0cd0b18) will increase coverage by 0.03%.
The diff coverage is 87.03%.

@@            Coverage Diff             @@
##             main     #389      +/-   ##
==========================================
+ Coverage   92.60%   92.63%   +0.03%     
==========================================
  Files         207      208       +1     
  Lines       16100    16372     +272     
==========================================
+ Hits        14909    15167     +258     
- Misses       1191     1205      +14     

Impacted Files	Coverage Δ
tests/optimization/test_pounders_integration.py	92.45% <ø> (ø)
tests/optimization/tranquilo/test_fit_models.py	78.89% <58.92%> (-21.11%)	⬇️
src/estimagic/optimization/tranquilo/tranquilo.py	90.47% <79.54%> (-1.38%)	⬇️
.../estimagic/optimization/tranquilo/filter_points.py	80.00% <82.35%> (+13.33%)	⬆️
src/estimagic/optimization/tranquilo/fit_models.py	83.53% <92.30%> (+5.24%)	⬆️
...stimagic/benchmarking/process_benchmark_results.py	95.55% <100.00%> (ø)
...timagic/optimization/tranquilo/aggregate_models.py	87.71% <100.00%> (+4.11%)	⬆️
src/estimagic/optimization/tranquilo/models.py	100.00% <100.00%> (ø)
...ts/optimization/tranquilo/test_aggregate_models.py	100.00% <100.00%> (ø)
tests/optimization/tranquilo/test_filter_points.py	100.00% <100.00%> (ø)
... and 6 more

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

[janosg]

Thanks! I ran some benchmarks on the different fitting methods and settled on fit_pounders_experimental. The differences are small but it is better in some cases.