Feedback

Author: fealho

sdgym/benchmark.py

@@ -38,7 +38,6 @@
 )
 from sdmetrics.single_table import DCRBaselineProtection
 
-# from sdgym import __version__ as SDGYM_VERSION
 from sdgym.datasets import get_dataset_paths, load_dataset
 from sdgym.errors import BenchmarkError, SDGymError
 from sdgym.metrics import get_metrics

sdgym/result_explorer/result_explorer.py

@@ -10,6 +10,7 @@
     S3ResultsHandler,
 )
 from sdgym.s3 import _get_s3_client, is_s3_path
+from sdgym.synthesizers.base import _validate_modality
 
 
 def _validate_local_path(path):
@@ -34,9 +35,7 @@ def _resolve_effective_path(path, modality):
     if not modality:
         return path
 
-    if modality not in ('single_table', 'multi_table'):
-        valid = ', '.join(sorted(('single_table', 'multi_table')))
-        raise ValueError(f'Invalid modality "{modality}". Valid options are: {valid}.')
+    _validate_modality(modality)
 
     # Avoid double-appending if already included
     if str(path).rstrip('/').endswith(('/' + modality, modality)):

sdgym/result_explorer/result_handler.py

@@ -401,7 +401,7 @@ def load_synthetic_data(self, file_path):
                     if name.endswith('.csv'):
                         table_name = os.path.splitext(os.path.basename(name))[0]
                         with zf.open(name) as csv_file:
-                            tables[table_name] = pd.read_csv(csv_file, low_memory=False)
+                            tables[table_name] = pd.read_csv(csv_file)
 
             return tables
 

tests/unit/result_explorer/test_result_handler.py

@@ -261,6 +261,24 @@ def test_load_metainfo(self):
 class TestLocalResultsHandler:
     """Unit tests for the LocalResultsHandler class."""
 
+    def test__init__sets_base_path_and_default_baseline(self, tmp_path):
+        """Test it initializes base_path and default baseline."""
+        # Run
+        handler = LocalResultsHandler(str(tmp_path))
+
+        # Assert
+        assert handler.base_path == str(tmp_path)
+        assert handler.baseline_synthesizer == 'GaussianCopulaSynthesizer'
+
+    def test__init__supports_baseline_override(self, tmp_path):
+        """Test it allows overriding baseline synthesizer."""
+        # Run
+        handler = LocalResultsHandler(str(tmp_path), baseline_synthesizer='CustomBaseline')
+
+        # Assert
+        assert handler.base_path == str(tmp_path)
+        assert handler.baseline_synthesizer == 'CustomBaseline'
+
     def test_list(self, tmp_path):
         """Test the `list` method"""
         # Setup
@@ -418,9 +436,7 @@ def test_get_file_path_local_error(self, mock_isfile, mock_exists):
 class TestS3ResultsHandler:
     """Unit tests for the S3ResultsHandler class."""
 
-    def test__init__(
-        self,
-    ):
+    def test__init__(self):
         """Test the `__init__` method."""
         # Setup
         path = 's3://my-bucket/prefix'
@@ -432,6 +448,21 @@ def test__init__(
         assert result_handler.s3_client == 's3_client'
         assert result_handler.bucket_name == 'my-bucket'
         assert result_handler.prefix == 'prefix/'
+        assert result_handler.baseline_synthesizer == 'GaussianCopulaSynthesizer'
+
+    def test__init__supports_baseline_override(self):
+        """Test it allows overriding baseline synthesizer."""
+        # Run
+        s3_client = Mock()
+        handler = S3ResultsHandler(
+            's3://bkt/prefix', s3_client, baseline_synthesizer='CustomBaseline'
+        )
+
+        # Assert
+        assert handler.baseline_synthesizer == 'CustomBaseline'
+        assert handler.s3_client == s3_client
+        assert handler.bucket_name == 'bkt'
+        assert handler.prefix == 'prefix/'
 
     def test_list(self):
         """Test the `list` method."""

tests/unit/test_benchmark.py

@@ -729,7 +729,6 @@ def test__setup_output_destination_multi_table(tmp_path):
 @patch('sdgym.benchmark.datetime')
 def test__write_metainfo_file(mock_datetime, tmp_path):
     """Test the `_write_metainfo_file` method."""
-    pytest.importorskip('realtabformer')
     # Setup
     output_destination = tmp_path / 'SDGym_results_06_26_2025'
     output_destination.mkdir()