Release 4.0.2 (spotify#159)

* Fixed categorical_order_by used with array_like (spotify#157) * Fix categorical_order_by check for scatter plot * Fix categorical_order_by check for _construct_source * Refactor category sorting in _construct_source * Add tests for categorical_order_by * Correct scatter test (spotify#158) * Update version in init * Update HISTORY.rst --------- Co-authored-by: Quoc Duong Bui <[email protected]>
ego · Mar 30, 2023 · 364361f · 364361f
1 parent ca828ab
commit 364361f
Show file tree

Hide file tree

Showing 4 changed files with 127 additions and 36 deletions.
diff --git a/HISTORY.rst b/HISTORY.rst
@@ -2,6 +2,15 @@
 History
 =======
 
+4.0.2 (2023-03-30)
+------------------
+
+* Fix categorical_order_by check for scatter plot
+* Fix categorical_order_by check for _construct_source
+* Refactor category sorting in _construct_source
+* Add tests for categorical_order_by
+* Fix scatter plot tests that used line plots
+
 4.0.1 (2023-03-24)
 ------------------
 

diff --git a/chartify/__init__.py b/chartify/__init__.py
@@ -23,7 +23,7 @@
 
 __author__ = """Chris Halpert"""
 __email__ = "[email protected]"
-__version__ = "4.0.1"
+__version__ = "4.0.2"
 
 _IPYTHON_INSTANCE = False
 

diff --git a/chartify/_core/plot.py b/chartify/_core/plot.py
@@ -947,6 +947,49 @@ def _get_bar_width(factors):
         else:
             return 0.9
 
+    @staticmethod
+    def _sort_categories_by_value(source, categorical_columns, categorical_order_ascending):
+        # Recursively sort values within each level of the index.
+        row_totals = source.sum(axis=1, numeric_only=True)
+        row_totals.name = "sum"
+        old_index = row_totals.index
+        row_totals = row_totals.reset_index()
+        row_totals.columns = ["_%s" % col for col in row_totals.columns]
+        row_totals.index = old_index
+
+        hierarchical_sort_cols = categorical_columns[:]
+        for i, _ in enumerate(hierarchical_sort_cols):
+            row_totals["level_%s" % i] = row_totals.groupby(hierarchical_sort_cols[: i + 1])["_sum"].transform(
+                "sum"
+            )
+        row_totals = row_totals.sort_values(
+            by=["level_%s" % i for i, _ in enumerate(hierarchical_sort_cols)],
+            ascending=categorical_order_ascending,
+        )
+        return source.reindex(row_totals.index)
+
+    @staticmethod
+    def _sort_categories(
+            source,
+            categorical_columns,
+            categorical_order_by,
+            categorical_order_ascending
+    ):
+
+        is_string = isinstance(categorical_order_by, str)
+        order_length = getattr(categorical_order_by, "__len__", None)
+        # Sort the categories
+        if is_string and categorical_order_by == "values":
+            return PlotMixedTypeXY._sort_categories_by_value(
+                source, categorical_columns, categorical_order_ascending)
+        elif is_string and categorical_order_by == "labels":
+            return source.sort_index(axis=0, ascending=categorical_order_ascending)
+        # Manual sort
+        elif not is_string and order_length is not None:
+            return source.reindex(categorical_order_by, axis="index")
+
+        raise ValueError("""Must be 'values', 'labels', or a list of values.""")
+
     def _construct_source(
         self,
         data_frame,
@@ -1014,34 +1057,7 @@ def _construct_source(
         if normalize:
             source = source.div(source.sum(axis=1), axis=0)
 
-        order_length = getattr(categorical_order_by, "__len__", None)
-        # Sort the categories
-        if categorical_order_by == "values":
-            # Recursively sort values within each level of the index.
-            row_totals = source.sum(axis=1, numeric_only=True)
-            row_totals.name = "sum"
-            old_index = row_totals.index
-            row_totals = row_totals.reset_index()
-            row_totals.columns = ["_%s" % col for col in row_totals.columns]
-            row_totals.index = old_index
-
-            heirarchical_sort_cols = categorical_columns[:]
-            for i, _ in enumerate(heirarchical_sort_cols):
-                row_totals["level_%s" % i] = row_totals.groupby(heirarchical_sort_cols[: i + 1])["_sum"].transform(
-                    "sum"
-                )
-            row_totals = row_totals.sort_values(
-                by=["level_%s" % i for i, _ in enumerate(heirarchical_sort_cols)],
-                ascending=categorical_order_ascending,
-            )
-            source = source.reindex(row_totals.index)
-        elif categorical_order_by == "labels":
-            source = source.sort_index(axis=0, ascending=categorical_order_ascending)
-        # Manual sort
-        elif order_length is not None:
-            source = source.reindex(categorical_order_by, axis="index")
-        else:
-            raise ValueError("""Must be 'values', 'labels', or a list of values.""")
+        source = self._sort_categories(source, categorical_columns, categorical_order_by, categorical_order_ascending)
 
         # Cast all categorical columns to strings
         # Plotting functions will break with non-str types.
@@ -2003,13 +2019,14 @@ def scatter(
 
         axis_factors = data_frame.groupby(categorical_columns).size()
 
+        is_string = isinstance(categorical_order_by, str)
         order_length = getattr(categorical_order_by, "__len__", None)
-        if categorical_order_by == "labels":
+        if is_string and categorical_order_by == "labels":
             axis_factors = axis_factors.sort_index(ascending=categorical_order_ascending).index
-        elif categorical_order_by == "count":
+        elif is_string and categorical_order_by == "count":
             axis_factors = axis_factors.sort_values(ascending=categorical_order_ascending).index
         # User-specified order.
-        elif order_length is not None:
+        elif not is_string and order_length is not None:
             axis_factors = categorical_order_by
         else:
             raise ValueError("""Must be 'count', 'labels', or a list of values.""")

diff --git a/tests/test_plots.py b/tests/test_plots.py
@@ -130,15 +130,15 @@ def setup_method(self):
         })
 
     def test_single_numeric_scatter(self):
-        """Single line test"""
+        """Single scatter test"""
         single_scatter = self.data[self.data['category1'] == 'a']
         ch = chartify.Chart()
-        ch.plot.line(single_scatter, x_column='number1', y_column='number2')
+        ch.plot.scatter(single_scatter, x_column='number1', y_column='number2')
         assert (np.array_equal(chart_data(ch, '')['number1'], [1., 2., 3.]))
         assert (np.array_equal(chart_data(ch, '')['number2'], [5, 10, 0]))
 
     def test_multi_numeric_scatter(self):
-        """Single line test"""
+        """Multi scatter test"""
         ch = chartify.Chart()
         ch.plot.scatter(
             self.data,
@@ -151,7 +151,7 @@ def test_multi_numeric_scatter(self):
         assert (np.array_equal(chart_data(ch, 'b')['number2'], [4, -3, -10]))
 
     def test_single_datetime_scatter(self):
-        """Single line test"""
+        """Single datetime scatter test"""
         data = pd.DataFrame({
             'number': [1, 10, -10, 0],
             'datetimes':
@@ -794,6 +794,71 @@ def test_grouped_histogram(self):
         assert (np.array_equal(chart_data(ch, 'b')['min_edge'], [2., 6.]))
 
 
+class TestCategoricalOrderBy:
+    def _assert_order_by_array_like(self, chart):
+        assert (np.array_equal(chart.figure.x_range.factors, ['b', 'd', 'a', 'c']))
+        # check bar data
+        assert (np.array_equal(chart_data(chart, '')['factors'], ['b', 'd', 'a', 'c']))
+        assert (np.array_equal(chart_data(chart, '')['number1'], [3, 1, 4, 2]))
+        # check scatter data
+        assert (np.array_equal(chart_data(chart, 'number1')['factors'], ['a', 'b', 'c', 'd']))
+        assert (np.array_equal(chart_data(chart, 'number1')['number1'], [4, 3, 2, 1]))
+
+    def setup_method(self):
+        self.data1 = pd.DataFrame({
+            'category1': ['a', 'b', 'c', 'd'],
+            'number1': [4, 3, 2, 1],
+        })
+
+        self.data2 = pd.DataFrame({
+            'category2': ['b', 'a', 'b', 'b', 'a', 'c'],
+            'number2': [1, 2, 3, 4, 5, 6]
+        })
+
+    def test_order_by_labels(self):
+        ch = chartify.Chart(x_axis_type='categorical')
+
+        ch.plot.bar(self.data1, ['category1'], 'number1', categorical_order_by='labels')
+        assert (np.array_equal(ch.figure.x_range.factors, ['d', 'c', 'b', 'a']))
+        assert (np.array_equal(chart_data(ch, '')['factors'], ['d', 'c', 'b', 'a']))
+        assert (np.array_equal(chart_data(ch, '')['number1'], [1, 2, 3, 4]))
+
+        ch.plot.scatter(self.data1, ['category1'], 'number1', categorical_order_by='labels')
+        assert (np.array_equal(ch.figure.x_range.factors, ['d', 'c', 'b', 'a']))
+        assert (np.array_equal(chart_data(ch, 'number1')['factors'], ['a', 'b', 'c', 'd']))
+        assert (np.array_equal(chart_data(ch, 'number1')['number1'], [4, 3, 2, 1]))
+
+    def test_order_by_values(self):
+        ch = chartify.Chart(x_axis_type='categorical')
+        ch.plot.bar(self.data1, ['category1'], 'number1', categorical_order_by='values')
+        assert (np.array_equal(chart_data(ch, '')['factors'], ['a', 'b', 'c', 'd']))
+        assert (np.array_equal(chart_data(ch, '')['number1'], [4, 3, 2, 1]))
+
+    def test_order_by_count(self):
+        ch = chartify.Chart(x_axis_type='categorical')
+        ch.plot.scatter(self.data2, ['category2'], 'number2', categorical_order_by='count')
+
+        assert (np.array_equal(ch.figure.x_range.factors, ['b', 'a', 'c']))
+        assert (np.array_equal(chart_data(ch, 'number2')['factors'], ['b', 'a', 'b', 'b', 'a', 'c']))
+        assert (np.array_equal(chart_data(ch, 'number2')['number2'], [1, 2, 3, 4, 5, 6]))
+
+    @pytest.mark.parametrize(
+        'array_like', [['b', 'd', 'a', 'c'], np.array(['b', 'd', 'a', 'c']), pd.Series(['b', 'd', 'a', 'c'])])
+    def test_order_by_array_like(self, array_like):
+        ch = chartify.Chart(x_axis_type='categorical')
+        ch.plot.scatter(self.data1, ['category1'], 'number1', categorical_order_by=array_like)
+        ch.plot.bar(self.data1, ['category1'], 'number1', categorical_order_by=array_like)
+
+        self._assert_order_by_array_like(ch)
+
+    @pytest.mark.parametrize('plot_method,categorical_order_by', [('bar', 'count'), ('scatter', 'values')])
+    def test_error(self, plot_method, categorical_order_by):
+        ch = chartify.Chart(x_axis_type='categorical', y_axis_type='linear')
+        with pytest.raises(ValueError):
+            plot_method = getattr(ch.plot, plot_method)
+            plot_method(self.data1, ['category1'], 'number1', categorical_order_by=categorical_order_by)
+
+
 def test_categorical_axis_type_casting():
     """Categorical axis plotting breaks for non-str types.
     Test that type casting is performed correctly"""