Validate derrf distribution parameters on startup

[larsevj]

Issue
Resolves #9523

• PR title captures the intent of the changes, and is fitting for release notes.
• Added appropriate release note label
• Commit history is consistent and clean, in line with the contribution guidelines.
• Make sure unit tests pass locally after every commit (git rebase -i main --exec 'pytest tests/ert/unit_tests -n logical -m "not integration_test"')

When applicable

• When there are user facing changes: Updated documentation
• New behavior or changes to existing untested code: Ensured that unit tests are added (See Ground Rules).
• Large PR: Prepare changes in small commits for more convenient review
• Bug fix: Add regression test for the bug
• Bug fix: Create Backport PR to latest release

[codspeed-hq]

CodSpeed Performance Report

Merging #9599 will not alter performance

_{Comparing larsevj:validate_derrf_function (b51d234) with main (ab373dd)}

Summary

✅ 24 untouched benchmarks

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 91.71%. Comparing base (839610b) to head (b51d234).
Report is 4 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #9599      +/-   ##
==========================================
- Coverage   91.74%   91.71%   -0.03%     
==========================================
  Files         430      430              
  Lines       26639    26661      +22     
==========================================
+ Hits        24439    24452      +13     
- Misses       2200     2209       +9     

Flag	Coverage Δ
cli-tests	39.69% <75.00%> (+<0.01%)	⬆️
everest-models-test	34.16% <0.00%> (-0.03%)	⬇️
gui-tests	71.82% <16.66%> (-0.04%)	⬇️
integration-test	38.26% <0.00%> (-0.04%)	⬇️
performance-tests	51.75% <16.66%> (+0.06%)	⬆️
test	39.58% <0.00%> (-0.03%)	⬇️
unit-tests	73.97% <100.00%> (-0.02%)	⬇️

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.

[oyvindeide]

It might be outside the scope for this PR, but did you consider converting this from a function to a dataclass? Something like:

@pydantic.dataclass
class DERRF:
    name: str
    steps: PositiveInt
    min: float
    max: float
    skewness: float
    width: confloat(gt=0)

    @model_validator(mode="after")
    def validate_min_max(self) -> Self:
        if not self.max > self.min
            raise ValueError(f"Max ({self.max}) must be larger than max ({self.max})
        return self
 
    def transform(self, x: float) -> float:
        """
        Bin the result of `trans_errf` with `min=0` and `max=1` to closest of `nbins`
        linearly spaced values on [0,1]. Finally map [0,1] to [min, max].
        """
        q_values = np.linspace(start=0, stop=1, num=self.steps)
        q_checks = np.linspace(start=0, stop=1, num=self.steps + 1)[1:]
        y = TransformFunction.trans_errf(x, [0, 1, self.skewness, self.width])
        bin_index = np.digitize(y, q_checks, right=True)
        y_binned = q_values[bin_index]
        result = self.min + y_binned * (self.max - self.min)
        if result > self.max or result < self.min:
            warnings.warn(
                "trans_derff suffered from catastrophic loss of precision, clamping to min,max",
                stacklevel=1,
            )
            return np.clip(result, self.min, self.max)
        if np.isnan(result):
            raise ValueError(
                "trans_derrf returns nan, check that input arguments are reasonable"
            )
        return result

[larsevj]

I did not, but will look into it.

[larsevj]

I guess this will require a storage migration?

[oyvindeide]

Yes, I guess it will, though it would be possible to work around it for now, and only migrate once we do something like this for all parameter types. If that is the direction you are heading with rewriting parameter configs?

[eivindjahren]

I suggest we merge this fix for now and then you can do this conversion later.