Add abstractmethod for BaseWrangler to explicitly enforce

implementation of all required methods before wrangler instantiation.
Add `concretize_abstract_wrangler` helper function for testing classes
which do not implement all abstract methods.

src/pywrangler/base.py

@@ -3,11 +3,13 @@
 
 """
 
+from abc import ABC, abstractmethod
+
 from pywrangler.util import _pprint
 from pywrangler.util.helper import get_param_names
 
 
-class BaseWrangler:
+class BaseWrangler(ABC):
     """Defines the basic interface common to all data wranglers.
 
     In analogy to sklearn transformers (see link below), all wranglers have to
@@ -42,10 +44,12 @@ class BaseWrangler:
     """
 
     @property
+    @abstractmethod
     def preserves_sample_size(self) -> bool:
         raise NotImplementedError
 
     @property
+    @abstractmethod
     def computation_engine(self) -> str:
         raise NotImplementedError
 
@@ -91,12 +95,15 @@ def set_params(self, **params):
 
         return self
 
+    @abstractmethod
     def fit(self, *args, **kwargs):
         raise NotImplementedError
 
+    @abstractmethod
     def transform(self, *args, **kwargs):
         raise NotImplementedError
 
+    @abstractmethod
     def fit_transform(self, *args, **kwargs):
         raise NotImplementedError
 

src/pywrangler/util/testing.py

@@ -0,0 +1,43 @@
+"""This module contains testing utility.
+
+"""
+
+from typing import Type
+
+
+def concretize_abstract_wrangler(wrangler_class: Type) -> Type:
+    """Makes abstract wrangler classes instantiable for testing purposes by
+    implementing abstract methods of `BaseWrangler`.
+
+    Parameters
+    ----------
+    wrangler_class: Type
+        Class object to inherit from while overriding abstract methods.
+
+    Returns
+    -------
+    concrete_class: Type
+        Concrete class usable for testing.
+
+    """
+
+    class ConcreteWrangler(wrangler_class):
+
+        @property
+        def preserves_sample_size(self):
+            return super().preserves_sample_size
+
+        @property
+        def computation_engine(self):
+            return super().computation_engine
+
+        def fit(self, *args, **kwargs):
+            return super().fit(*args, **kwargs)
+
+        def fit_transform(self, *args, **kwargs):
+            return super().fit_transform(*args, **kwargs)
+
+        def transform(self, *args, **kwargs):
+            return super().transform(*args, **kwargs)
+
+    return ConcreteWrangler

tests/dask/test_base.py

@@ -9,9 +9,10 @@
 dask = pytest.importorskip("dask")  # noqa: E402
 
 from pywrangler.dask.base import DaskWrangler
+from pywrangler.util.testing import concretize_abstract_wrangler
 
 
 def test_dask_base_wrangler_engine():
-    wrangler = DaskWrangler()
+    wrangler = concretize_abstract_wrangler(DaskWrangler)()
 
     assert wrangler.computation_engine == "dask"

tests/dask/test_benchmark.py

@@ -21,6 +21,7 @@
     DaskBaseProfiler
 )
 from pywrangler.dask.base import DaskSingleNoFit
+from pywrangler.util.testing import concretize_abstract_wrangler
 
 
 @pytest.fixture
@@ -29,7 +30,7 @@ class DummyWrangler(DaskSingleNoFit):
         def transform(self, df):
             return df.mean()
 
-    return DummyWrangler()
+    return concretize_abstract_wrangler(DummyWrangler)()
 
 
 @pytest.fixture
@@ -66,7 +67,7 @@ def transform(self, df):
                 time.sleep(sleep)
                 return df_out
 
-        return DummyWrangler()
+        return concretize_abstract_wrangler(DummyWrangler)()
 
     return create_wrangler
 

tests/pandas/test_base.py

@@ -7,12 +7,13 @@
 import pandas as pd
 
 from pywrangler.pandas.base import PandasWrangler
+from pywrangler.util.testing import concretize_abstract_wrangler
 
 pytestmark = pytest.mark.pandas
 
 
 def test_pandas_base_wrangler_engine():
-    wrangler = PandasWrangler()
+    wrangler = concretize_abstract_wrangler(PandasWrangler)()
 
     assert wrangler.computation_engine == "pandas"
 
@@ -24,7 +25,7 @@ class DummyWrangler(PandasWrangler):
         def preserves_sample_size(self):
             return preserves_sample_size
 
-    wrangler = DummyWrangler()
+    wrangler = concretize_abstract_wrangler(DummyWrangler)()
 
     df1 = pd.DataFrame([0] * 10)
     df2 = pd.DataFrame([0] * 20)

tests/pandas/test_benchmark.py

@@ -15,6 +15,7 @@
     PandasMemoryProfiler,
     PandasTimeProfiler
 )
+from pywrangler.util.testing import concretize_abstract_wrangler
 
 pytestmark = pytest.mark.pandas
 
@@ -52,7 +53,7 @@ def transform(self, df):
                 time.sleep(sleep)
                 return df_out
 
-        return DummyWrangler()
+        return concretize_abstract_wrangler(DummyWrangler)()
 
     return create_wrangler
 

tests/pyspark/test_base.py

@@ -9,9 +9,10 @@
 pyspark = pytest.importorskip("pyspark")  # noqa: E402
 
 from pywrangler.pyspark.base import PySparkWrangler
+from pywrangler.util.testing import concretize_abstract_wrangler
 
 
 def test_spark_base_wrangler_engine():
-    wrangler = PySparkWrangler()
+    wrangler = concretize_abstract_wrangler(PySparkWrangler)()
 
     assert wrangler.computation_engine == "pyspark"

tests/pyspark/test_benchmark.py

@@ -13,6 +13,7 @@
 from pywrangler.pyspark.base import PySparkSingleNoFit
 from pywrangler.pyspark.benchmark import PySparkTimeProfiler, \
     PySparkBaseProfiler
+from pywrangler.util.testing import concretize_abstract_wrangler
 
 SLEEP = 0.0001
 
@@ -24,7 +25,7 @@ def transform(self, df):
             time.sleep(SLEEP)
             return df
 
-    return DummyWrangler
+    return concretize_abstract_wrangler(DummyWrangler)
 
 
 def test_spark_time_profiler_fastest(spark, wrangler_sleeps):

tests/test_base.py

@@ -5,9 +5,10 @@
 import pytest
 
 from pywrangler import base
+from pywrangler.util.testing import concretize_abstract_wrangler
 
 
-@pytest.fixture(scope="module")
+@pytest.fixture(scope="session")
 def dummy_wrangler():
     """Create DummyWrangler for testing BaseWrangler.
 
@@ -26,23 +27,26 @@ def preserves_sample_size(self):
         def computation_engine(self):
             return "DummyEngine"
 
-    return DummyWrangler("arg_val", "kwarg_val")
+    return concretize_abstract_wrangler(DummyWrangler)("arg_val", "kwarg_val")
 
 
 def test_base_wrangler_not_implemented():
 
-    wrangler = base.BaseWrangler()
+    with pytest.raises(TypeError):
+        base.BaseWrangler()
+
+    empty_wrangler = concretize_abstract_wrangler(base.BaseWrangler)()
 
     test_attributes = ("preserves_sample_size", "computation_engine")
     test_methods = ("fit", "transform", "fit_transform")
 
     for test_attribute in test_attributes:
         with pytest.raises(NotImplementedError):
-            getattr(wrangler, test_attribute)
+            getattr(empty_wrangler, test_attribute)
 
     for test_method in test_methods:
         with pytest.raises(NotImplementedError):
-            getattr(wrangler, test_method)()
+            getattr(empty_wrangler, test_method)()
 
 
 def test_base_wrangler_get_params(dummy_wrangler):

tests/test_interfaces.py

@@ -5,10 +5,11 @@
 import pytest
 
 from pywrangler import wranglers
+from pywrangler.util.testing import concretize_abstract_wrangler
 
 
 @pytest.fixture(scope="module")
-def interval_ident_kwargs():
+def ii_kwargs():
 
     return {"marker_column": "marker_col",
             "marker_start": "start",
@@ -19,26 +20,31 @@ def interval_ident_kwargs():
             "target_column_name": "abc"}
 
 
-def test_base_interval_identifier_init(interval_ident_kwargs):
+def test_base_interval_identifier_init(ii_kwargs):
 
-    bii = wranglers.IntervalIdentifier(**interval_ident_kwargs)
+    wrangler = concretize_abstract_wrangler(wranglers.IntervalIdentifier)
+    bii = wrangler(**ii_kwargs)
 
-    assert bii.get_params() == interval_ident_kwargs
+    assert bii.get_params() == ii_kwargs
 
 
-def test_base_interval_identifier_sort_length_exc(interval_ident_kwargs):
+def test_base_interval_identifier_sort_length_exc(ii_kwargs):
 
-    incorrect_length = interval_ident_kwargs.copy()
+    incorrect_length = ii_kwargs.copy()
     incorrect_length["ascending"] = (True, )
 
+    wrangler = concretize_abstract_wrangler(wranglers.IntervalIdentifier)
+
     with pytest.raises(ValueError):
-        wranglers.IntervalIdentifier(**incorrect_length)
+        wrangler(**incorrect_length)
 
 
-def test_base_interval_identifier_sort_keyword_exc(interval_ident_kwargs):
+def test_base_interval_identifier_sort_keyword_exc(ii_kwargs):
 
-    incorrect_keyword = interval_ident_kwargs.copy()
+    incorrect_keyword = ii_kwargs.copy()
     incorrect_keyword["ascending"] = ("wrong keyword", "wrong keyword too")
 
+    wrangler = concretize_abstract_wrangler(wranglers.IntervalIdentifier)
+
     with pytest.raises(ValueError):
-        wranglers.IntervalIdentifier(**incorrect_keyword)
+        wrangler(**incorrect_keyword)

tests/util/test_testing.py

@@ -0,0 +1,24 @@
+"""This module contains tests for testing utilities.
+
+"""
+
+import pytest
+
+from pywrangler.base import BaseWrangler
+from pywrangler.util.testing import concretize_abstract_wrangler
+
+
+def test_concretize_abstract_wrangler():
+
+    class Dummy(BaseWrangler):
+        @property
+        def computation_engine(self) -> str:
+            return "engine"
+
+    concrete_class = concretize_abstract_wrangler(Dummy)
+    instance = concrete_class()
+
+    assert instance.computation_engine == "engine"
+
+    with pytest.raises(NotImplementedError):
+        instance.preserves_sample_size