Fixed issue with scorer test

tests/test_pipelines/test_scorer.py

@@ -46,14 +46,15 @@ def test_scorer_calls_final_agg(self):
         mock_final_agg = Mock()
         scorer = Scorer(mock_score_func, final_aggregator=mock_final_agg)
         pipe = DummyOptimizablePipeline()
-        scorer(pipeline=pipe, dataset=DummyDataset())
+        ds = DummyDataset()
+        scorer(pipeline=pipe, dataset=ds)
 
         assert mock_final_agg.call_count == 1
         # final aggregator gets called with the previous agg values, the single values, the pipeline and the dataset
         assert mock_final_agg.call_args[0][0] == 1
         assert mock_final_agg.call_args[0][1] == [1] * len(DummyDataset())
         assert mock_final_agg.call_args[0][2] == pipe
-        assert mock_final_agg.call_args[0][3] == DummyDataset()
+        assert mock_final_agg.call_args[0][3] == ds
 
 
 class TestScorer:

tpcp/_dataset.py

@@ -354,6 +354,10 @@ def __repr__(self) -> str:
     def __eq__(self, other):
         if not isinstance(other, type(self)):
             return False
+        # Note: This fails, if you compare a dataset, for which the index was already generated with an identical one
+        #       for which `index` was not called yet.
+        #       This might be a bit unexpected, but I consider this correct behaviour and I don't want to trigger the
+        #       index creation here, as this might be costly.
         return custom_hash(self.get_params()) == custom_hash(other.get_params())
 
     def _repr_html_(self) -> str: