Further transf

Author: xjules

src/ert/config/scalar_parameter.py

@@ -1,7 +1,9 @@
+import warnings
 from dataclasses import dataclass
 from pathlib import Path
 from typing import Literal
 
+import numpy as np
 import xarray as xr
 from scipy.stats import norm
 
@@ -11,7 +13,7 @@
 
 
 @dataclass
-class TransUniformfSettings:
+class TransUniformSettings:
     name: Literal["uniform"] = "uniform"
     min: float = 0.0
     max: float = 1.0
@@ -22,7 +24,7 @@ def trans(self, x: float) -> float:
 
 
 @dataclass
-class TransNormalfSettings:
+class TransNormalSettings:
     name: Literal["normal"] = "normal"
     mean: float = 0.0
     std: float = 1.0
@@ -48,6 +50,53 @@ def trans(self, _: float) -> float:
         return self.value
 
 
+@dataclass
+class TransErrfSettings:
+    name: Literal["errf"] = "errf"
+    min: float = 0.0
+    max: float = 1.0
+    skew: float = 0.0
+    width: float = 1.0
+
+    def trans(self, x: float) -> float:
+        y = norm(loc=0, scale=self.width).cdf(x + self.skew)
+        if np.isnan(y):
+            raise ValueError(
+                "Output is nan, likely from triplet (x, skewness, width) "
+                "leading to low/high-probability in normal CDF."
+            )
+        return self.min + y * (self.max - self.min)
+
+
+@dataclass
+class TransDerrfSettings:
+    name: Literal["derrf"] = "derrf"
+    steps: int = 1000
+    min: float = 0.0
+    max: float = 1.0
+    skew: float = 0.0
+    width: float = 1.0
+
+    def trans(self, x: float) -> float:
+        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 = TransErrfSettings(min=0, max=1, skew=self.skew, width=self.width).trans(x)
+        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 float(result)
+
+
 @dataclass
 class PolarsData:
     name: Literal["polars"]
@@ -59,10 +108,12 @@ class ScalarParameter(ParameterConfig):
     name: str
     group: str
     distribution: (
-        TransUniformfSettings
+        TransUniformSettings
         | TransRawSettings
         | TransConstSettings
-        | TransNormalfSettings
+        | TransNormalSettings
+        | TransErrfSettings
+        | TransDerrfSettings
     )
     active: bool
     input_source: Literal["design_matrix", "sampled"]