debug: Check windows issue of not legal

Author: eddiebergman

pyproject.toml

@@ -157,8 +157,6 @@ select = [
 ignore = [
   "T201",    # TODO: Remove
   "D100",
-  "D101",    # Missing docstring in public class
-  "D102",    # Missing docstring in public method   (Should be active but it's too much for right now)
   "D104",    # Missing docstring in public package
   "D105",    # Missing docstring in magic mthod
   "D203",    # 1 blank line required before class docstring
@@ -208,7 +206,6 @@ exclude = [
 "test/*.py" = [
   "S101",
   "D102",
-  "D103",
   "D404",
   "ARG001",
   "A002",

src/ConfigSpace/forbidden.py

@@ -289,7 +289,7 @@ def __init__(self, hyperparameter: Hyperparameter, values: Iterable[Any]) -> Non
         >>> cs.add(forbidden_clause_a)
         Forbidden: a in {2, 3}
 
-        Note:
+        Note
         ----
         The forbidden values have to be a subset of the hyperparameter's values.
 

src/ConfigSpace/hyperparameters/_hp_components.py

@@ -129,7 +129,7 @@ class UnitScaler(_Transformer[DType]):
 
     lower_vectorized: f64 = field(init=False)
     upper_vectorized: f64 = field(init=False)
-    _scale_vec_to_int: DType = field(init=False)
+    _size: DType = field(init=False)
 
     def __post_init__(self) -> None:
         if self.lower_value >= self.upper_value:
@@ -147,7 +147,7 @@ def __post_init__(self) -> None:
         self.lower_vectorized = f64(0)
         self.upper_vectorized = f64(1)
 
-        self._scale_vec_to_int = self.upper_value - self.lower_value
+        self._size = self.upper_value - self.lower_value
 
     def to_value(self, vector: Array[f64]) -> Array[DType]:
         """Transform a value from the unit interval to the range."""
@@ -169,9 +169,7 @@ def _unsafe_to_value_single(self, vector: f64) -> f64:
             scaled = vector * (_u - _l) + _l
             return np.exp(scaled)  # type: ignore
 
-        _l = self.lower_value
-        _u = self.upper_value
-        return vector * (_u - _l) + _l  # type: ignore
+        return vector * self._size + _l  # type: ignore
 
     def _unsafe_to_value(self, vector: Array[f64]) -> Array[f64]:
         # NOTE: Unsafe as it does not check boundaries, clip or integer'ness
@@ -203,7 +201,7 @@ def vectorize_size(self, size: f64) -> f64:
                 / (np.log(self.upper_value) - np.log(self.lower_value)),
             )
 
-        return f64(size / (self.upper_value - self.lower_value))
+        return f64(size / self._size)
 
     def legal_value(self, value: Array[Any]) -> Mask:
         # If we have a non numeric dtype, we have to unfortunatly go through but by bit
@@ -253,7 +251,10 @@ def legal_vector_single(self, vector: np.number) -> bool:
 
         if not self.log:
             inbounds = bool(self.lower_vectorized <= vector <= self.upper_vectorized)
-            scaled = vector * self._scale_vec_to_int
+            scaled = vector * self._size
+            print("SCALED", scaled)
+            print("is_close", is_close_to_integer_single(scaled, atol=ATOL))
+            print("inbounds", inbounds)
             return bool(is_close_to_integer_single(scaled, atol=ATOL) and inbounds)
 
         value = self._unsafe_to_value_single(vector)  # type: ignore