version 0.0.3

changelog.txt

@@ -1,3 +1,7 @@
+# 0.0.3
+* Fixed crash when a boolean column contains only one unique value (e.g., all True) and the table is rendered with the native backend.
+* Fixed object columns containing decimal.Decimal values (e.g., from Amazon Redshift) being incorrectly treated as categorical instead of numerical.
+
 # 0.0.2
 * Introduced narwhals as dependency, adds support for pandas 3, polars,
 pyarrow

pysummaries/__init__.py

@@ -30,4 +30,4 @@
            'pandas_to_report_html', 'get_styles', 'Pandas2HTMLSummaryTable',
            'PySummariesException']
 
-__version__ = '0.0.2'
+__version__ = '0.0.3'

pysummaries/reportable/intermediate_representation.py

@@ -180,6 +180,8 @@ def extract_multibodyblocks(indexes, df, value_styles, styles, tabid):
         #row labels
         *rowgroup_labels, currow_label = rowlabel
         if (rowgroup_labels != lastblock_index and lastblock_index is not None) or cnt==len(df):
+            if lastblock_index is None:
+                lastblock_index = rowgroup_labels
             rowgroupstyle = styles.get("rowgrouplabel")
             titles = [RowGroupLabel(x, tabid, level, y, rowgroup_cnts, style=rowgroupstyle) for level,(x,y) in enumerate(zip(lastblock_index, labels_paddings))]
             # prune 

pysummaries/table_summary/table_summary.py

@@ -179,6 +179,11 @@ def calculate_table_summary(df, strata=None, show_overall=True, columns_labels=N
     coltypes = detect_df_col_types(df)
     if not isinstance(df, pd.DataFrame):
         df = nw.from_native(df).to_pandas()
+    # Convert object columns classified as numerical (e.g. decimal.Decimal) to
+    # a numeric dtype so that pandas can compute stats on them.
+    for col_name, col_type in coltypes.items():
+        if col_type == "numerical" and df[col_name].dtype == object:
+            df[col_name] = pd.to_numeric(df[col_name], errors="coerce")
     colnames = df.columns.to_list()
     strat_cats = list()
     if strata is not None:

pysummaries/table_summary/utils.py

@@ -13,6 +13,8 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 # #############################################################################
+from decimal import Decimal
+
 import narwhals as nw
 import pandas as pd
 import numpy as np
@@ -41,7 +43,7 @@ def _classify_object_col(df, col_name):
         return "categorical"
     if curtype == str:
         return "categorical"
-    if np.issubdtype(type(col.iloc[0]), np.number) or isinstance(col.iloc[0], (int, float)):
+    if np.issubdtype(type(col.iloc[0]), np.number) or isinstance(col.iloc[0], (int, float, Decimal)):
         return "numerical"
     return "categorical"
 

setup.py

@@ -28,7 +28,7 @@
 
 setup(
     name="pysummaries",
-    version='0.0.2',
+    version='0.0.3',
     author="Otto Fajardo",
     author_email="pleasecontactviagithub@notvalid.com",
     description="Produce table summaries from pandas, polars or PyArrow dataframes",

tests/test_basic.py

@@ -437,6 +437,77 @@ def test_simple_table_summary_gt(self):
 # ---------------------------------------------------------------------------
 # HTML content tests using BeautifulSoup
 # ---------------------------------------------------------------------------
+class TestEdgeCases(unittest.TestCase):
+    """Tests for edge cases and bug fixes."""
+
+    def test_bool_column_single_value(self):
+        """Boolean column with only one unique value should not crash.
+
+        Regression test: when the summary DataFrame has a single row,
+        cnt==len(df) on the first iteration while lastblock_index is still
+        None, causing a TypeError in extract_multibodyblocks.
+        """
+        df = pd.DataFrame({
+            "flag": [True, True, True, True],
+            "group": ["A", "A", "B", "B"],
+        })
+        result = pysummaries.get_table_summary(
+            df, strata="group", columns_include=["flag", "group"],
+            show_overall=False
+        )
+        self.assertIsNotNone(result)
+
+    def test_decimal_column_treated_as_numeric(self):
+        """Columns containing decimal.Decimal values should be treated as numeric."""
+        from decimal import Decimal
+        df = pd.DataFrame({
+            "value": [Decimal("12.5"), Decimal("24.3"), Decimal("6.1"), Decimal("18.7")],
+            "group": ["A", "A", "B", "B"],
+        })
+        col_types = detect_df_col_types(df)
+        self.assertEqual(col_types["value"], "numerical")
+
+    def test_decimal_column_summary(self):
+        """Decimal columns should produce numeric summaries (Mean, Median, etc.)."""
+        from decimal import Decimal
+        df = pd.DataFrame({
+            "value": [Decimal("12.5"), Decimal("24.3"), Decimal("6.1"), Decimal("18.7")],
+            "group": ["A", "A", "B", "B"],
+        })
+        result, _ = pysummaries.calculate_table_summary(
+            df, strata="group", columns_include=["value", "group"],
+            show_overall=False
+        )
+        # Numeric summaries have "Mean (SD)" as a row label
+        self.assertIn("Mean (SD)", result.index.get_level_values(-1))
+
+    def test_decimal_column_polars(self):
+        """Polars Decimal columns should produce numeric summaries."""
+        from decimal import Decimal
+        df = pl.DataFrame({
+            "value": [Decimal("12.5"), Decimal("24.3"), Decimal("6.1"), Decimal("18.7")],
+            "group": ["A", "A", "B", "B"],
+        })
+        result, _ = pysummaries.calculate_table_summary(
+            df, strata="group", columns_include=["value", "group"],
+            show_overall=False
+        )
+        self.assertIn("Mean (SD)", result.index.get_level_values(-1))
+
+    def test_decimal_column_pyarrow(self):
+        """PyArrow Decimal columns should produce numeric summaries."""
+        from decimal import Decimal
+        table = pa.table({
+            "value": pa.array([Decimal("12.5"), Decimal("24.3"), Decimal("6.1"), Decimal("18.7")]),
+            "group": ["A", "A", "B", "B"],
+        })
+        result, _ = pysummaries.calculate_table_summary(
+            table, strata="group", columns_include=["value", "group"],
+            show_overall=False
+        )
+        self.assertIn("Mean (SD)", result.index.get_level_values(-1))
+
+
 class TestHTMLContent(unittest.TestCase):
 
     def setUp(self):