Tranquilo Refactoring

[timmens]

• Replace np.squeeze with operation that is more precise, unless an axis argument is given
• Clip sampled points at bounds
• Adaptively sample points given trustregion shape

[codecov]

Codecov Report

Merging #445 (fca4c94) into main (51fd8b6) will increase coverage by 0.00%.
The diff coverage is 98.55%.

@@           Coverage Diff           @@
##             main     #445   +/-   ##
=======================================
  Coverage   92.97%   92.98%           
=======================================
  Files         248      247    -1     
  Lines       18526    18479   -47     
=======================================
- Hits        17225    17182   -43     
+ Misses       1301     1297    -4     

Impacted Files	Coverage Δ
src/estimagic/optimization/tranquilo/geometry.py	100.00% <ø> (+3.44%)	⬆️
...sts/optimization/subsolvers/test_gqtpar_lambdas.py	100.00% <ø> (ø)
tests/visualization/test_visualize_tranquilo.py	100.00% <ø> (ø)
src/estimagic/parameters/block_trees.py	90.34% <87.50%> (+0.05%)	⬆️
...timagic/optimization/tranquilo/aggregate_models.py	87.50% <100.00%> (+0.22%)	⬆️
src/estimagic/optimization/tranquilo/region.py	97.22% <100.00%> (+0.07%)	⬆️
.../estimagic/optimization/tranquilo/sample_points.py	98.24% <100.00%> (+1.78%)	⬆️
src/estimagic/optimization/tranquilo/tranquilo.py	97.03% <100.00%> (-0.02%)	⬇️
tests/optimization/tranquilo/test_fit_models.py	100.00% <100.00%> (ø)
tests/optimization/tranquilo/test_sample_points.py	100.00% <100.00%> (ø)

... and 1 file with indirect coverage changes

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

[janosg]

We can merge after:

• Test coverage is higher
• We confirmed that the benchmark is unchanged