Merge pull request #7126 from efiring/logauto

tacaswell · web-flow · commit 6741ddf692ea · 2016-09-18T21:43:48.000-04:00
ENH: LogLocator can auto-select number of ticks based on axis
diff --git a/lib/matplotlib/tests/baseline_images/test_axes/auto_numticks_log.png b/lib/matplotlib/tests/baseline_images/test_axes/auto_numticks_log.png
diff --git a/lib/matplotlib/tests/test_axes.py b/lib/matplotlib/tests/test_axes.py
@@ -4418,9 +4418,19 @@ def test_adjust_numtick_aspect():
 @image_comparison(baseline_images=["auto_numticks"], style='default',
                   extensions=['png'])
 def test_auto_numticks():
+    # Make tiny, empty subplots, verify that there are only 3 ticks.
     fig, axes = plt.subplots(4, 4)
 
 
+@image_comparison(baseline_images=["auto_numticks_log"], style='default',
+                  extensions=['png'])
+def test_auto_numticks_log():
+    # Verify that there are not too many ticks with a large log range.
+    fig, ax = plt.subplots()
+    matplotlib.rcParams['axes.autolimit_mode'] = 'round_numbers'
+    ax.loglog([1e-20, 1e5], [1e-16, 10])
+
+
 @cleanup
 def test_broken_barh_empty():
     fig, ax = plt.subplots()
diff --git a/lib/matplotlib/ticker.py b/lib/matplotlib/ticker.py
@@ -1639,8 +1639,11 @@ def set_params(self, **kwargs):
 
     def _raw_ticks(self, vmin, vmax):
         if self._nbins == 'auto':
-            nbins = max(min(self.axis.get_tick_space(), 9),
-                        max(1, self._min_n_ticks - 1))
+            if self.axis is not None:
+                nbins = max(min(self.axis.get_tick_space(), 9),
+                            max(1, self._min_n_ticks - 1))
+            else:
+                nbins = 9
         else:
             nbins = self._nbins
 
@@ -1757,15 +1760,19 @@ class LogLocator(Locator):
     Determine the tick locations for log axes
     """
 
-    def __init__(self, base=10.0, subs=(1.0,), numdecs=4, numticks=15):
+    def __init__(self, base=10.0, subs=(1.0,), numdecs=4, numticks=None):
         """
         place ticks on the location= base**i*subs[j]
         """
+        if numticks is None:
+            if rcParams['_internal.classic_mode']:
+                numticks = 15
+            else:
+                numticks = 'auto'
         self.base(base)
         self.subs(subs)
-        # this needs to be validated > 1 with traitlets
-        self.numticks = numticks
         self.numdecs = numdecs
+        self.numticks = numticks
 
     def set_params(self, base=None, subs=None, numdecs=None, numticks=None):
         """Set parameters within this locator."""
@@ -1786,7 +1793,7 @@ def base(self, base):
 
     def subs(self, subs):
         """
-        set the minor ticks the log scaling every base**i*subs[j]
+        set the minor ticks for the log scaling every base**i*subs[j]
         """
         if subs is None:
             self._subs = None  # autosub
@@ -1799,6 +1806,14 @@ def __call__(self):
         return self.tick_values(vmin, vmax)
 
     def tick_values(self, vmin, vmax):
+        if self.numticks == 'auto':
+            if self.axis is not None:
+                numticks = max(min(self.axis.get_tick_space(), 9), 2)
+            else:
+                numticks = 9
+        else:
+            numticks = self.numticks
+
         b = self._base
         # dummy axis has no axes attribute
         if hasattr(self.axis, 'axes') and self.axis.axes.name == 'polar':
@@ -1836,12 +1851,14 @@ def tick_values(self, vmin, vmax):
             subs = self._subs
 
         stride = 1
-        if not self.numticks > 1:
-            raise RuntimeError('The number of ticks must be greater than 1 '
-                               'for LogLocator.')
-        # FIXME: The following was designed for integer division in py2.
-        while numdec / stride + 1 > self.numticks:
-            stride += 1
+
+        if rcParams['_internal.classic_mode']:
+            # Leave the bug left over from the PY2-PY3 transition.
+            while numdec / stride + 1 > numticks:
+                stride += 1
+        else:
+            while numdec // stride + 1 > numticks:
+                stride += 1
 
         decades = np.arange(math.floor(vmin) - stride,
                             math.ceil(vmax) + 2 * stride, stride)
