dedupe tests

Signed-off-by: nina-xu <19981858+nina-xu@users.noreply.github.com>

Author: nina-xu

tests/data_processing/test_validation.py

@@ -16,9 +16,23 @@ def test_validate_groupby_column_noop_when_groupby_is_none() -> None:
     validate_groupby_column(df, None)
 
 
+def test_validate_groupby_column_passes_when_column_exists() -> None:
+    df = pd.DataFrame(
+        {
+            "col1": [1, 2, 3, 4, 5],
+            "col2": ["a", "b", "c", "d", "e"],
+            "group_col": ["g1", "g1", "g2", "g2", "g3"],
+        }
+    )
+    validate_groupby_column(df, "group_col")
+
+
 def test_validate_groupby_column_raises_for_missing_column() -> None:
     df = pd.DataFrame({"a": [1, 2], "b": [3, 4]})
-    with pytest.raises(ParameterError, match="disable grouping"):
+    with pytest.raises(
+        ParameterError,
+        match=r"Group by column 'missing_group' not found in input dataset columns.*disable grouping",
+    ):
         validate_groupby_column(df, "missing_group")
 
 

tests/training/test_huggingface_backend.py

@@ -18,8 +18,7 @@
     SafeSynthesizerParameters,
     TrainingHyperparams,
 )
-from nemo_safe_synthesizer.data_processing.validation import validate_groupby_column
-from nemo_safe_synthesizer.errors import DataError, ParameterError
+from nemo_safe_synthesizer.errors import ParameterError
 from nemo_safe_synthesizer.training.huggingface_backend import (
     HuggingFaceBackend,
     compute_metrics,
@@ -209,17 +208,6 @@ def sample_dataframe():
     )
 
 
-@pytest.fixture
-def dataframe_with_null_group():
-    """Create a DataFrame with null values in the group column."""
-    return pd.DataFrame(
-        {
-            "col1": [1, 2, 3],
-            "group_col": ["g1", None, "g2"],
-        }
-    )
-
-
 class TestFilterModelKwargs:
     def test_filters_trainer_specific_keys(self, backend):
         """Test that trainer-specific keys are filtered out."""
@@ -561,31 +549,6 @@ def test_uses_provided_data_collator(self, backend):
         assert "data_collator" not in training_args
 
 
-class TestValidateGroupbyColumn:
-    def test_does_nothing_when_no_groupby(self, sample_dataframe):
-        """Test that nothing happens when groupby is None."""
-        validate_groupby_column(sample_dataframe, None)  # Should not raise
-
-    def test_passes_when_column_exists(self, sample_dataframe):
-        """Test that validation passes when column exists."""
-        validate_groupby_column(sample_dataframe, "group_col")  # Should not raise
-
-    def test_raises_when_column_missing(self, sample_dataframe):
-        """Test that ParameterError is raised when column is missing."""
-        with pytest.raises(ParameterError, match="Group by column 'nonexistent_col' not found"):
-            validate_groupby_column(sample_dataframe, "nonexistent_col")
-
-    def test_raises_with_comma_hint_when_column_has_comma(self, sample_dataframe):
-        """Test that ParameterError is raised when column name has a comma."""
-        with pytest.raises(ParameterError, match="multi-column grouping is not supported"):
-            validate_groupby_column(sample_dataframe, "patient_id,event_id")
-
-    def test_raises_when_column_has_nulls(self, dataframe_with_null_group):
-        """Test that DataError is raised when column has null values."""
-        with pytest.raises(DataError, match="has missing values"):
-            validate_groupby_column(dataframe_with_null_group, "group_col")
-
-
 class TestValidateOrderbyColumn:
     def test_does_nothing_when_no_orderby(self, backend, sample_dataframe):
         """Test that nothing happens when orderby is None."""