Make scale a required parameter of mapping setup

seaborn/_core/mappings.py

@@ -73,7 +73,7 @@ def _homogenize_values(self, values):
     def setup(
         self,
         data: Series,
-        scale: Scale | None = None,
+        scale: Scale,
     ) -> SemanticMapping:
 
         raise NotImplementedError()
@@ -117,7 +117,7 @@ def _default_values(self, n: int) -> list:
     def setup(
         self,
         data: Series,
-        scale: Scale | None = None,
+        scale: Scale,
     ) -> LookupMapping:
 
         values = self._values
@@ -186,7 +186,7 @@ def _infer_map_type(
     def setup(
         self,
         data: Series,
-        scale: Scale | None = None,
+        scale: Scale,
     ) -> NormedMapping | LookupMapping:
 
         values = self.default_range if self._values is None else self._values
@@ -265,7 +265,7 @@ def __init__(self, palette: PaletteSpec = None, variable: str = "color"):
     def setup(
         self,
         data: Series,
-        scale: Scale | None = None,
+        scale: Scale,
     ) -> LookupMapping | NormedMapping:
         """Infer the type of mapping to use and define it using this vector of data."""
         mapping: LookupMapping | NormedMapping

seaborn/_core/plot.py

@@ -10,7 +10,7 @@
 import matplotlib as mpl
 import matplotlib.pyplot as plt  # TODO defer import into Plot.show()
 
-from seaborn._core.rules import variable_type, categorical_order
+from seaborn._core.rules import categorical_order
 from seaborn._core.data import PlotData
 from seaborn._core.subplots import Subplots
 from seaborn._core.mappings import (
@@ -575,10 +575,9 @@ def _setup_scales(self) -> None:
                 # Because we only want to concat if a variable was *added* here
                 *(y.data.frame.get(var) for y in self._layers if var in y.variables)
             ], ignore_index=True)
-            var_type = variable_type(all_values)
 
             # TODO eventually this will be updating a different dictionary
-            self._scales[var] = ScaleWrapper.from_inferred_type(var_type)
+            self._scales[var] = ScaleWrapper.from_inferred_type(all_values)
 
         # TODO Think about how this is going to handle situations where we have
         # e.g. ymin and ymax but no y specified. I think in that situation one
@@ -828,7 +827,7 @@ def _scale_coords_single(
             # TODO FIXME:feedback wrap this in a try/except and reraise with
             # more information about what variable caused the problem
             values = scale.cast(values)
-            axis_obj.update_units(categorical_order(values))
+            axis_obj.update_units(categorical_order(values))  # TODO think carefully
 
             # TODO it seems wrong that we need to cast to float here,
             # but convert_units sometimes outputs an object array (e.g. w/Int64 values)

seaborn/_core/scales.py

@@ -52,15 +52,19 @@ def __deepcopy__(self, memo=None):
             return copy(self)
 
     @classmethod
-    def from_inferred_type(cls, var_type: VarType):
+    def from_inferred_type(cls, data: Series) -> ScaleWrapper:
 
+        var_type = variable_type(data)
+        axis = data.name
         if var_type == "numeric":
-            return cls(LinearScale(), "numeric", None)
+            scale = cls(LinearScale(axis), "numeric", None)
         elif var_type == "categorical":
-            return cls(CategoricalScale(), "categorical", None)
+            scale = cls(CategoricalScale(axis), "categorical", None)
         elif var_type == "datetime":
             # TODO add DateTimeNorm that converts to numeric first
-            return cls(DatetimeScale(), "datetime", None)
+            scale = cls(DatetimeScale(axis), "datetime", None)
+        scale.type_declared = False
+        return scale
 
     @property
     def order(self):
@@ -90,7 +94,7 @@ class CategoricalScale(LinearScale):
 
     def __init__(
         self,
-        axis: str | None = None,
+        axis: str,
         order: list | None = None,
         formatter: Any = None
     ):