Address inf_as_na pandas deprecation (#3424)

* Address inf_as_na pandas deprecation

* Add -np.inf, add import

* flake8

* Make copy

* Use mask instead of replace

seaborn/_base.py

@@ -1121,13 +1121,13 @@ def comp_data(self):
                 parts = []
                 grouped = self.plot_data[var].groupby(self.converters[var], sort=False)
                 for converter, orig in grouped:
-                    with pd.option_context('mode.use_inf_as_na', True):
-                        orig = orig.dropna()
-                        if var in self.var_levels:
-                            # TODO this should happen in some centralized location
-                            # it is similar to GH2419, but more complicated because
-                            # supporting `order` in categorical plots is tricky
-                            orig = orig[orig.isin(self.var_levels[var])]
+                    orig = orig.mask(orig.isin([np.inf, -np.inf]), np.nan)
+                    orig = orig.dropna()
+                    if var in self.var_levels:
+                        # TODO this should happen in some centralized location
+                        # it is similar to GH2419, but more complicated because
+                        # supporting `order` in categorical plots is tricky
+                        orig = orig[orig.isin(self.var_levels[var])]
                     comp = pd.to_numeric(converter.convert_units(orig)).astype(float)
                     if converter.get_scale() == "log":
                         comp = np.log10(comp)

seaborn/_core/plot.py

@@ -20,6 +20,7 @@
 from matplotlib.axes import Axes
 from matplotlib.artist import Artist
 from matplotlib.figure import Figure
+import numpy as np
 from PIL import Image
 
 from seaborn._marks.base import Mark
@@ -1587,21 +1588,24 @@ def split_generator(keep_na=False) -> Generator:
 
                 axes_df = self._filter_subplot_data(df, view)
 
-                with pd.option_context("mode.use_inf_as_na", True):
-                    if keep_na:
-                        # The simpler thing to do would be x.dropna().reindex(x.index).
-                        # But that doesn't work with the way that the subset iteration
-                        # is written below, which assumes data for grouping vars.
-                        # Matplotlib (usually?) masks nan data, so this should "work".
-                        # Downstream code can also drop these rows, at some speed cost.
-                        present = axes_df.notna().all(axis=1)
-                        nulled = {}
-                        for axis in "xy":
-                            if axis in axes_df:
-                                nulled[axis] = axes_df[axis].where(present)
-                        axes_df = axes_df.assign(**nulled)
-                    else:
-                        axes_df = axes_df.dropna()
+                axes_df_inf_as_nan = axes_df.copy()
+                axes_df_inf_as_nan = axes_df_inf_as_nan.mask(
+                    axes_df_inf_as_nan.isin([np.inf, -np.inf]), np.nan
+                )
+                if keep_na:
+                    # The simpler thing to do would be x.dropna().reindex(x.index).
+                    # But that doesn't work with the way that the subset iteration
+                    # is written below, which assumes data for grouping vars.
+                    # Matplotlib (usually?) masks nan data, so this should "work".
+                    # Downstream code can also drop these rows, at some speed cost.
+                    present = axes_df_inf_as_nan.notna().all(axis=1)
+                    nulled = {}
+                    for axis in "xy":
+                        if axis in axes_df:
+                            nulled[axis] = axes_df[axis].where(present)
+                    axes_df = axes_df_inf_as_nan.assign(**nulled)
+                else:
+                    axes_df = axes_df_inf_as_nan.dropna()
 
                 subplot_keys = {}
                 for dim in ["col", "row"]: