Merge pull request #455 from plotly/distplot_update

added the option for multiple binsizes

Author: Kevin Yan

plotly/tests/test_optional/test_figure_factory.py

@@ -151,6 +151,51 @@ def test_distplot_more_args(self):
                                    'yaxis': 'y1'}
         self.assertEqual(dp['data'][1], expected_dp_data_hist_2)
 
+        def test_distplot_binsize_array(self):
+            hist1_x = [0.8, 1.2, 0.2, 0.6, 1.6, -0.9, -0.07, 1.95, 0.9, -0.2,
+                       -0.5, 0.3, 0.4, -0.37, 0.6]
+            hist2_x = [0.8, 1.5, 1.5, 0.6, 0.59, 1.0, 0.8, 1.7, 0.5, 0.8, -0.3,
+                       1.2, 0.56, 0.3, 2.2]
+
+            hist_data = [hist1_x, hist2_x]
+            group_labels = ['2012', '2013']
+
+            dp = tls.FigureFactory.create_distplot(hist_data, group_labels,
+                                                   show_rug=False,
+                                                   bin_size=[.2, .2])
+
+            expected_dp_data_hist_1 = {'autobinx': False,
+                                       'histnorm': 'probability',
+                                       'legendgroup': '2012',
+                                       'marker': {'color': 'rgb(31, 119, 180)'},
+                                       'name': '2012',
+                                       'opacity': 0.7,
+                                       'type': 'histogram',
+                                       'x': [0.8, 1.2, 0.2, 0.6, 1.6, -0.9, -0.07,
+                                             1.95, 0.9, -0.2, -0.5, 0.3, 0.4,
+                                             -0.37, 0.6],
+                                       'xaxis': 'x1',
+                                       'xbins': {'end': 1.95, 'size': 0.2,
+                                                 'start': -0.9},
+                                       'yaxis': 'y1'}
+            self.assertEqual(dp['data'][0], expected_dp_data_hist_1)
+
+            expected_dp_data_hist_2 = {'autobinx': False,
+                                       'histnorm': 'probability',
+                                       'legendgroup': '2013',
+                                       'marker': {'color': 'rgb(255, 127, 14)'},
+                                       'name': '2013',
+                                       'opacity': 0.7,
+                                       'type': 'histogram',
+                                       'x': [0.8, 1.5, 1.5, 0.6, 0.59, 1.0, 0.8,
+                                             1.7, 0.5, 0.8, -0.3, 1.2, 0.56, 0.3,
+                                             2.2],
+                                       'xaxis': 'x1',
+                                       'xbins': {'end': 2.2, 'size': 0.2,
+                                                 'start': -0.3},
+                                       'yaxis': 'y1'}
+            self.assertEqual(dp['data'][1], expected_dp_data_hist_2)
+
 
 class TestStreamline(TestCase):
 
@@ -424,7 +469,7 @@ def test_dendrogram_random_matrix(self):
         self.assert_dict_equal(dendro['layout'], expected_dendro['layout'])
 
     def test_dendrogram_orientation(self):
-        X = np.random.rand(5, 5) 
+        X = np.random.rand(5, 5)
 
         dendro_left = tls.FigureFactory.create_dendrogram(
                        X, orientation='left')

plotly/tools.py

@@ -4033,7 +4033,8 @@ def create_distplot(hist_data, group_labels,
         :param (list[list]) hist_data: Use list of lists to plot multiple data
             sets on the same plot.
         :param (list[str]) group_labels: Names for each data set.
-        :param (float) bin_size: Size of histogram bins. Default = 1.
+        :param (list[float]|float) bin_size: Size of histogram bins.
+            Default = 1.
         :param (str) curve_type: 'kde' or 'normal'. Default = 'kde'
         :param (bool) show_hist: Add histogram to distplot? Default = True
         :param (bool) show_curve: Add curve to distplot? Default = True
@@ -4139,6 +4140,9 @@ def create_distplot(hist_data, group_labels,
         FigureFactory._validate_distplot(hist_data, curve_type)
         FigureFactory._validate_equal_length(hist_data, group_labels)
 
+        if isinstance(bin_size, (float, int)):
+            bin_size = [bin_size]*len(hist_data)
+
         hist = _Distplot(
             hist_data, group_labels, bin_size,
             curve_type, colors, rug_text,
@@ -5084,7 +5088,7 @@ def make_hist(self):
                                autobinx=False,
                                xbins=dict(start=self.start[index],
                                           end=self.end[index],
-                                          size=self.bin_size),
+                                          size=self.bin_size[index]),
                                opacity=.7)
         return hist
 
@@ -5104,7 +5108,7 @@ def make_kde(self):
             self.curve_y[index] = (scipy.stats.gaussian_kde
                                    (self.hist_data[index])
                                    (self.curve_x[index]))
-            self.curve_y[index] *= self.bin_size
+            self.curve_y[index] *= self.bin_size[index]
 
         for index in range(self.trace_number):
             curve[index] = dict(type='scatter',
@@ -5139,7 +5143,7 @@ def make_normal(self):
                                    / 500 for x in range(500)]
             self.curve_y[index] = scipy.stats.norm.pdf(
                 self.curve_x[index], loc=mean[index], scale=sd[index])
-            self.curve_y[index] *= self.bin_size
+            self.curve_y[index] *= self.bin_size[index]
 
         for index in range(self.trace_number):
             curve[index] = dict(type='scatter',