Add (untested/incomplete) prototype of marker mapping

seaborn/_core/mappings.py

@@ -28,14 +28,13 @@ def setup(self, data: Series, scale: Scale | None) -> SemanticMapping:
 
     def __call__(self, x):  # TODO types; will need to overload (wheee)
         # TODO this is a hack to get things working
-        # We are missing numeric maps and lots of other things
         if isinstance(x, pd.Series):
             if x.dtype.name == "category":  # TODO! possible pandas bug
                 x = x.astype(object)
             # TODO where is best place to ensure that LUT values are rgba tuples?
-            return np.stack(x.map(self.lookup_table).map(to_rgb))
+            return np.stack(x.map(self.lookup_table))
         else:
-            return to_rgb(self.lookup_table[x])
+            return self.lookup_table[x]
 
 
 # TODO Currently, the SemanticMapping objects are also the source of the information
@@ -74,6 +73,17 @@ def __init__(self, palette: PaletteSpec = None):
 
         self._input_palette = palette
 
+    def __call__(self, x):  # TODO types; will need to overload (wheee)
+        # TODO this is a hack to get things working
+        # We are missing numeric maps and lots of other things
+        if isinstance(x, pd.Series):
+            if x.dtype.name == "category":  # TODO! possible pandas bug
+                x = x.astype(object)
+            # TODO where is best place to ensure that LUT values are rgba tuples?
+            return np.stack(x.map(self.lookup_table).map(to_rgb))
+        else:
+            return to_rgb(self.lookup_table[x])
+
     def setup(
         self,
         data: Series,  # TODO generally rename Series arguments to distinguish from DF?
@@ -249,3 +259,38 @@ def _setup_numeric(
             lookup_table = dict(zip(levels, cmap(norm(levels))))
 
         return levels, lookup_table, norm, cmap
+
+
+class MarkerMapping(SemanticMapping):
+
+    def __init__(self, shapes: list | dict | None = None):  # TODO full types
+
+        # TODO fill or filled parameter?
+        # allow singletons? e.g. map_marker(shapes="o", filled=[True, False])?
+        # allow full matplotlib fillstyle API?
+
+        self._input_shapes = shapes
+
+    def setup(
+        self,
+        data: Series,  # TODO generally rename Series arguments to distinguish from DF?
+        scale: Scale | None = None,  # TODO or always have a Scale?
+    ) -> MarkerMapping:
+
+        shapes = self._input_shapes
+        order = None if scale is None else scale.order
+        levels = categorical_order(data, order)
+
+        # TODO input checking; generalize across mappings
+
+        # TODO default shapes
+
+        if isinstance(shapes, dict):
+            shapes = {k: mpl.markers.MarkerStyle(s) for k, s in shapes.items()}
+        elif isinstance(shapes, list):
+            shapes = {k: mpl.markers.MarkerStyle(s) for k, s in zip(levels, shapes)}
+
+        self.levels = levels
+        self.lookup_table = shapes
+
+        return self

seaborn/_core/plot.py

@@ -13,7 +13,7 @@
 from seaborn._core.rules import categorical_order, variable_type
 from seaborn._core.data import PlotData
 from seaborn._core.subplots import Subplots
-from seaborn._core.mappings import GroupMapping, ColorMapping
+from seaborn._core.mappings import GroupMapping, ColorMapping, MarkerMapping
 from seaborn._core.scales import (
     ScaleWrapper,
     CategoricalScale,
@@ -67,6 +67,7 @@ def __init__(
             "color": ColorMapping(),
             "facecolor": ColorMapping(),
             "edgecolor": ColorMapping(),
+            "marker": MarkerMapping(),
         }
 
         # TODO is using "unknown" here the best approach?
@@ -283,6 +284,14 @@ def map_edgecolor(
         self._mappings["edgecolor"] = ColorMapping(palette)
         return self
 
+    def map_marker(
+        self,
+        shapes: list | dict | None = None,
+    ) -> Plot:
+
+        self._mappings["marker"] = MarkerMapping(shapes)
+        return self
+
     # TODO have map_gradient?
     # This could be used to add another color-like dimension
     # and also the basis for what mappings like stat.density -> rgba do

seaborn/_marks/basic.py

@@ -61,6 +61,11 @@ def _plot_split(self, keys, data, ax, mappings, kws):
                 func = getattr(points, f"set_{var}")
                 func(mappings[var](data[var]))
 
+        if "marker" in data:
+            markers = mappings["marker"](data["marker"])
+            paths = [m.get_path().transformed(m.get_transform()) for m in markers]
+            points.set_paths(paths)
+
         # TODO note that when scaling this up we'll need to catch the artist
         # and update its attributes (like in scatterplot) to allow marker variation
 

seaborn/relational.py

@@ -565,6 +565,7 @@ def plot(self, ax, kws):
         # See https://github.com/matplotlib/matplotlib/issues/17849 for context
         m = kws.get("marker", mpl.rcParams.get("marker", "o"))
         if not isinstance(m, mpl.markers.MarkerStyle):
+            # TODO in more recent matplotlib (which?) can pass a MarkerStyle here
             m = mpl.markers.MarkerStyle(m)
         if m.is_filled():
             kws.setdefault("edgecolor", "w")