updating testing approach for MedianImputer Transformer

[Ritunjai-Sharma]

Connected to #150

Moved testing for MedianImputer Transformer to new approach

[Ritunjai-Sharma]

I have left a few tests unchanged/haven't removed them. These are:

1. TestFit:

• test_learnt_values and test_learnt_values_weighted: Couldn't find an equivalent test for this in tests.base_tests.GenericFitTests

• test_fit_returns_self_weighted: Although test_fit_returns_self in tests.base_tests.GenericFitTests tests for similar behaviour, it doesn't seem test for the case when weight parameter is populated.

• test_fit_not_changing_data and test_fit_not_changing_data_weighted: Couldn't find an equivalent test for this in tests.base_tests.GenericFitTests

2. TestTransform:

• test_learnt_values_not_modified_weights: Although test_impute_value_unchanged in test_BaseImputer.GenericImputerTransformTests tests for similar behavior, it doesn't seem to test for the case when weight parameter is populated.

Let me know if any of these need to be removed :)

[limlam96]

agree with leaving this in until issue is resolved

[limlam96]

Think this one is covered by the GenerateImputerTransform tests?

[limlam96]

Same with the next 2 - though maybe we can parametrize the inherited test to work with and without weights?

[Ritunjai-Sharma]

Thanks for catching this :) I have now removed test_nulls_imputed_correctly_1 and test_nulls_imputed_correctly_2 as well as expected_df_1 and expected_df_2, here: 2d7b045. Regarding testing with and without weights, perhaps it's best to add this functionality as part of #221 so that all transformers with weights can have this functionality? So, for now, I have renamed test_nulls_imputed_correctly_3 to test_nulls_imputed_correctly_weights and not deleted it. Let me know if all this makes sense!

[limlam96]

makes sense, thanks RJ :)