Merge pull request #31306 from null-dreams/mnt

QuLogic · web-flow · commit 7c94eb063956 · 2026-03-20T18:37:04.000-04:00
[MNT]: Implement `Scale.val_in_range` and refactor `_point_in_data_domain`
diff --git a/lib/matplotlib/axes/_base.py b/lib/matplotlib/axes/_base.py
@@ -2479,8 +2479,7 @@ def _point_in_data_domain(self, x, y):
         (e.g. negative coordinates with a log scale).
         """
         for val, axis in zip([x, y], self._axis_map.values()):
-            vmin, vmax = axis.limit_range_for_scale(val, val)
-            if vmin != val or vmax != val:
+            if not axis._scale.val_in_range(val):
                 return False
         return True
 
diff --git a/lib/matplotlib/scale.py b/lib/matplotlib/scale.py
@@ -30,6 +30,7 @@
 """  # noqa: E501
 
 import inspect
+import math
 import textwrap
 from functools import wraps
 
@@ -114,6 +115,22 @@ def limit_range_for_scale(self, vmin, vmax, minpos):
         """
         return vmin, vmax
 
+    def val_in_range(self, val):
+        """
+        Return whether the value(s) are within the valid range for this scale.
+
+        This method is a generic implementation. Subclasses may implement more
+        efficient solutions for their domain.
+        """
+        try:
+            if not math.isfinite(val):
+                return False
+            else:
+                vmin, vmax = self.limit_range_for_scale(val, val, minpos=1e-300)
+                return vmin == val and vmax == val
+        except (TypeError, ValueError):
+            return False
+
 
 def _make_axis_parameter_optional(init_func):
     """
@@ -196,6 +213,14 @@ def get_transform(self):
         """
         return IdentityTransform()
 
+    def val_in_range(self, val):
+        """
+        Return whether the value is within the valid range for this scale.
+
+        This is True for all values, except +-inf and NaN.
+        """
+        return math.isfinite(val)
+
 
 class FuncTransform(Transform):
     """
@@ -400,6 +425,14 @@ def limit_range_for_scale(self, vmin, vmax, minpos):
         return (minpos if vmin <= 0 else vmin,
                 minpos if vmax <= 0 else vmax)
 
+    def val_in_range(self, val):
+        """
+        Return whether the value is within the valid range for this scale.
+
+        This is True for value(s) > 0 except +inf and NaN.
+        """
+        return math.isfinite(val) and val > 0
+
 
 class FuncScaleLog(LogScale):
     """
@@ -581,6 +614,14 @@ def get_transform(self):
         """Return the `.SymmetricalLogTransform` associated with this scale."""
         return self._transform
 
+    def val_in_range(self, val):
+        """
+        Return whether the value is within the valid range for this scale.
+
+        This is True for all values, except +-inf and NaN.
+        """
+        return math.isfinite(val)
+
 
 class AsinhTransform(Transform):
     """Inverse hyperbolic-sine transformation used by `.AsinhScale`"""
@@ -707,6 +748,14 @@ def set_default_locators_and_formatters(self, axis):
         else:
             axis.set_major_formatter('{x:.3g}')
 
+    def val_in_range(self, val):
+        """
+        Return whether the value is within the valid range for this scale.
+
+        This is True for all values, except +-inf and NaN.
+        """
+        return math.isfinite(val)
+
 
 class LogitTransform(Transform):
     input_dims = output_dims = 1
@@ -820,6 +869,14 @@ def limit_range_for_scale(self, vmin, vmax, minpos):
         return (minpos if vmin <= 0 else vmin,
                 1 - minpos if vmax >= 1 else vmax)
 
+    def val_in_range(self, val):
+        """
+        Return whether the value is within the valid range for this scale.
+
+        This is True for value(s) which are between 0 and 1 (excluded).
+        """
+        return 0 < val < 1
+
 
 _scale_mapping = {
     'linear': LinearScale,
diff --git a/lib/matplotlib/scale.pyi b/lib/matplotlib/scale.pyi
@@ -12,6 +12,7 @@ class ScaleBase:
     def limit_range_for_scale(
         self, vmin: float, vmax: float, minpos: float
     ) -> tuple[float, float]: ...
+    def val_in_range(self, val: float) -> bool: ...
 
 class LinearScale(ScaleBase):
     name: str
diff --git a/lib/matplotlib/tests/test_scale.py b/lib/matplotlib/tests/test_scale.py
@@ -371,3 +371,65 @@ def set_default_locators_and_formatters(self, axis):
         # cleanup - there's no public unregister_scale()
         del mscale._scale_mapping["custom"]
         del mscale._scale_has_axis_parameter["custom"]
+
+
+def test_val_in_range():
+
+    test_cases = [
+        # LinearScale: Always True (even for Inf/NaN)
+        ('linear', 10.0, True),
+        ('linear', -10.0, True),
+        ('linear', 0.0, True),
+        ('linear', np.inf, False),
+        ('linear', np.nan, False),
+
+        # LogScale: Only positive values (> 0)
+        ('log', 1.0, True),
+        ('log', 1e-300, True),
+        ('log', 0.0, False),
+        ('log', -1.0, False),
+        ('log', np.inf, False),
+        ('log', np.nan, False),
+
+        # LogitScale: Strictly between 0 and 1
+        ('logit', 0.5, True),
+        ('logit', 0.0, False),
+        ('logit', 1.0, False),
+        ('logit', -0.1, False),
+        ('logit', 1.1, False),
+        ('logit', np.inf, False),
+        ('logit', np.nan, False),
+
+        # SymmetricalLogScale: Valid for all real numbers
+        # Uses ScaleBase fallback. NaN returns False since NaN != NaN
+        ('symlog', 10.0, True),
+        ('symlog', -10.0, True),
+        ('symlog', 0.0, True),
+        ('symlog', np.inf, False),
+        ('symlog', np.nan, False),
+    ]
+
+    for name, val, expected in test_cases:
+        scale_cls = mscale._scale_mapping[name]
+        s = scale_cls(axis=None)
+
+        result = s.val_in_range(val)
+        assert result is expected, (
+            f"Failed {name}.val_in_range({val})."
+            f"Expected {expected}, got {result}"
+        )
+
+
+def test_val_in_range_base_fallback():
+    # Directly test the ScaleBase fallback for custom scales.
+    # ScaleBase.limit_range_for_scale returns values unchanged by default
+    s = mscale.ScaleBase(axis=None)
+
+    # Normal values should be True
+    assert s.val_in_range(1.0) is True
+    assert s.val_in_range(-5.5) is True
+
+    # NaN and Inf returns False since they cannot be drawn in a plot
+    assert s.val_in_range(np.nan) is False
+    assert s.val_in_range(np.inf) is False
+    assert s.val_in_range(-np.inf) is False
