MESMER-X: use params instead of distr

[mathause]

• Closes frozen distribution considerably slower #532
• Tests added
• Fully documented, including CHANGELOG.rst

This avoids Expression.evaluate(...) in favor of Expression.evaluate_params(...), which leads to a approximately 2x-4x speedup of calibrating mesmer-x. Instead of adding many methods to Expression it calls the method on Expression.distr (e.g. Expression.distr.pdf(x, **params).

cc @l-pierini, @yquilcaille

[codecov]

Codecov Report

Attention: Patch coverage is 81.81818% with 8 lines in your changes missing coverage. Please review.

Project coverage is 80.46%. Comparing base (b46f590) to head (bf52404).
Report is 1 commits behind head on main.

Files with missing lines	Patch %	Lines
mesmer/mesmer_x/train_l_distrib_mesmerx.py	80.00%	8 Missing ⚠️

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #611   +/-   ##
=======================================
  Coverage   80.45%   80.46%           
=======================================
  Files          49       49           
  Lines        3080     3081    +1     
=======================================
+ Hits         2478     2479    +1     
  Misses        602      602           

Flag	Coverage Δ
unittests	80.46% <81.81%> (+<0.01%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

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

[mathause]

Speedup

py3.13 ubuntu-latest

Before (6m 38s total)

169.00s call   tests/integration/test_calibrate_mesmer_x.py::test_calibrate_mesmer_x[ssp585-tasmax-norm(loc=c1 + c2 * __tas__, scale=c3)-expr1_2ndfit-True-False]
85.83s call    tests/integration/test_calibrate_mesmer_x.py::test_calibrate_mesmer_x[ssp585-tasmax-norm(loc=c1 + c2 * __tas__, scale=c3)-expr1-False-False]
8.70s call     tests/unit/test_mesmer_x_first_guess.py::test_first_guess_truncnorm

After (3m 43s total)

56.41s call    tests/integration/test_calibrate_mesmer_x.py::test_calibrate_mesmer_x[ssp585-tasmax-norm(loc=c1 + c2 * __tas__, scale=c3)-expr1_2ndfit-True-False]
29.68s call    tests/integration/test_calibrate_mesmer_x.py::test_calibrate_mesmer_x[ssp585-tasmax-norm(loc=c1 + c2 * __tas__, scale=c3)-expr1-False-False]
2.75s call     tests/unit/test_mesmer_x_first_guess.py::test_first_guess_truncnorm

---

py3.13 windows-latest

Before (11m 40s)

229.83s call     tests/integration/test_calibrate_mesmer_x.py::test_calibrate_mesmer_x[ssp585-tasmax-norm(loc=c1 + c2 * __tas__, scale=c3)-expr1_2ndfit-True-False]
116.79s call     tests/integration/test_calibrate_mesmer_x.py::test_calibrate_mesmer_x[ssp585-tasmax-norm(loc=c1 + c2 * __tas__, scale=c3)-expr1-False-False]
11.97s call     tests/unit/test_mesmer_x_first_guess.py::test_first_guess_truncnorm

After (9m 52s)

75.44s call     tests/integration/test_calibrate_mesmer_x.py::test_calibrate_mesmer_x[ssp585-tasmax-norm(loc=c1 + c2 * __tas__, scale=c3)-expr1_2ndfit-True-False]
39.01s call     tests/integration/test_calibrate_mesmer_x.py::test_calibrate_mesmer_x[ssp585-tasmax-norm(loc=c1 + c2 * __tas__, scale=c3)-expr1-False-False]
3.44s call     tests/unit/test_mesmer_x_first_guess.py::test_first_guess_truncnorm

[veni-vidi-vici-dormivi]

Really cool!!