Merge branch 'mr-cell-props-156264181' into rel-5.0.64

src/cr/cube/crunch_cube.py

@@ -357,7 +357,13 @@ def _mr_margin(self, axis, weighted, adjusted):
         margin = table[self.ind_selected] + table[self.ind_non_selected]
         margin = margin[np.ix_(*self.valid_indices)]
 
-        return np.sum(margin, 1 - self.mr_dim_ind) if axis is None else margin
+        if axis is None:
+            axis = tuple([
+                i for i, _ in enumerate(self.dimensions) if i != self.mr_dim_ind
+            ])
+            return np.sum(margin, axis)
+
+        return margin
 
     def _inflate_dim(self, array, axis):
         # Explicitly check if axis is tuple (which could be the case for doing
@@ -433,14 +439,15 @@ def _mr_proportions(self, axis, weighted, include_transforms_for_dims=None):
                            (table[:, 0, :, :] + table[:, 1, :, :]))
 
             # The following are normal MR x something (not CA)
-            if axis == 1:
+            if axis == 0:
+                res = table[:, 0, :] / (table[:, 0, :] + table[:, 1, :])
+            else:
                 num = self.as_array(
                     include_transforms_for_dims=include_transforms_for_dims
                 )
-                den = self.margin(axis=1)[:, np.newaxis]
+                den = self.margin(axis=axis)[:, np.newaxis]
                 return num / den
-            else:
-                res = table[:, 0, :] / (table[:, 0, :] + table[:, 1, :])
+
         elif self.mr_dim_ind == 1:
             num = table[self.ind_selected][np.ix_(*self.valid_indices)]
             non_selected = table[self.ind_non_selected][np.ix_(*self.valid_indices)]

tests/integration/test_crunch_cube.py

@@ -885,7 +885,6 @@ def test_z_scores_admit_by_gender_weighted_rows(self):
         actual = cube.standardized_residuals
         np.testing.assert_almost_equal(actual, expected)
 
-
     def test_selected_crosstab_dim_names(self):
         cube = CrunchCube(FIXT_SELECTED_CROSSTAB_4)
         expected = ['Statements agreed with about Climate', 'Gender']
@@ -976,15 +975,15 @@ def test_selected_crosstab_proportions_by_cols(self):
         actual = cube.proportions(axis=0)
         np.testing.assert_almost_equal(actual, expected)
 
-    def test_selected_crosstab_proportions_total(self):
+    def test_selected_crosstab_proportions_by_cell(self):
         cube = CrunchCube(FIXT_SELECTED_CROSSTAB_4)
         expected = np.array([
-            [.6815894,  .73842499],
-            [.6632888,  .63436033],
-            [.81146893, .85004262],
-            [.85174892, .88409986],
-            [.37871497, .28475016],
-            [.53511338, .5426064],
+            [0.329036700375595, 0.381950958618156],
+            [0.320620717695708, 0.327723062528721],
+            [0.392254504152701, 0.439142047148397],
+            [0.418103897371069, 0.450115632023491],
+            [0.185809278853744, 0.14504292098248],
+            [0.262843097025161, 0.27608279697761],
         ])
         actual = cube.proportions()
         np.testing.assert_almost_equal(actual, expected)
@@ -1190,9 +1189,9 @@ def test_fruit_x_pets_proportions_by_col(self):
     def test_pets_x_fruit_proportions_by_cell(self):
         cube = CrunchCube(FIXT_PETS_X_FRUIT)
         expected = np.array([
-            [.4285714, .5384615],
-            [.48, .4074074],
-            [.5217391, .5531915],
+            [0.15, 0.35],
+            [0.15189873, 0.27848101],
+            [0.17142857, 0.37142857],
         ])
         actual = cube.proportions(axis=None)
         np.testing.assert_almost_equal(actual, expected)
@@ -1346,6 +1345,12 @@ def test_pets_array_x_pets_cell(self):
         actual = cube.proportions(axis=None)[0]
         np.testing.assert_almost_equal(actual, expected)
 
+    def test_pets_x_pets_array_margin_by_cell(self):
+        cube = CrunchCube(FIXT_PETS_X_PETS_ARRAY)
+        expected = np.array([229, 229, 229])
+        actual = cube.margin()
+        np.testing.assert_almost_equal(actual, expected)
+
     def test_pets_x_pets_array_by_col(self):
         cube = CrunchCube(FIXT_PETS_X_PETS_ARRAY)
         expected = np.array([
@@ -1360,23 +1365,29 @@ def test_pets_x_pets_array_by_col(self):
     def test_pets_x_pets_array_by_row(self):
         cube = CrunchCube(FIXT_PETS_X_PETS_ARRAY)
         expected = np.array([
-            [0.44444444, 0.41176471],
-            [0., 1.],
-            [0.5, 0.47368421],
+            [0.55555556, 0.19444444],
+            [0., 0.55555556],
+            [0.44444444, 0.25],
         ])
-        actual = cube.proportions(axis=2)[0]
-        # Since cube is 3D, col dim is 1 (instead of 0)
+        actual = cube.proportions(axis=1)[0]
         np.testing.assert_almost_equal(actual, expected)
 
     def test_pets_x_pets_array_by_cell(self):
         cube = CrunchCube(FIXT_PETS_X_PETS_ARRAY)
         expected = np.array([
-            [0.44444444, 0.41176471],
-            [0., 1.],
-            [0.5, 0.47368421],
+            [[0.08733624, 0.06113537],
+             [0., 0.17467249],
+             [0.069869, 0.07860262]],
+
+            [[0., 0.14847162],
+             [0.06550218, 0.06113537],
+             [0.05676856, 0.069869]],
+
+            [[0.09606987, 0.069869],
+             [0.08733624, 0.07860262],
+             [0., 0.16593886]]
         ])
-        actual = cube.proportions(axis=None)[0]
-        # Since cube is 3D, col dim is 1 (instead of 0)
+        actual = cube.proportions()
         np.testing.assert_almost_equal(actual, expected)
 
     def test_mr_x_cat_x_cat_by_row(self):