Merge pull request #24 from georgianpartners/issue_7

Fixed OneHotEncoder to use version from sklearn-contrib/categorey enc…
georgian-io-archive · Nov 30, 2018 · cd7c5ad · cd7c5ad
2 parents a9b2059 + 2da32a9
commit cd7c5ad
Show file tree

Hide file tree

Showing 5 changed files with 24 additions and 27 deletions.
diff --git a/foreshadow/tests/test_transformers/test_internal.py b/foreshadow/tests/test_transformers/test_internal.py
@@ -73,28 +73,34 @@ def test_transformer_fancy_impute_invalid_params():
 
 def test_transformer_onehotencoder_fit_transform():
     import pandas as pd
-    from foreshadow.transformers.internals import OneHotEncoder
+    from foreshadow.transformers.externals import OneHotEncoder
 
     df = pd.DataFrame({"neat": ["apple", "apple", "orange", "apple", "orange"]})
-    ohe = OneHotEncoder()
+    ohe = OneHotEncoder(use_cat_names=True, cols=["neat"], handle_unknown="ignore")
     assert ohe.fit(df) == ohe
     assert list(ohe.transform(df)) == [
-        "neat_OneHotEncoder_apple",
-        "neat_OneHotEncoder_orange",
+        "neat_OneHotEncoder_neat_apple",
+        "neat_OneHotEncoder_neat_orange",
     ]
 
 
 def test_transformer_onehotencoder_fit_transform_keep_cols():
     import pandas as pd
-    from foreshadow.transformers.internals import OneHotEncoder
+    from foreshadow.transformers.externals import OneHotEncoder
 
     df = pd.DataFrame({"neat": ["apple", "apple", "orange", "apple", "orange"]})
-    ohe = OneHotEncoder(keep_columns=True, name="encoder")
+    ohe = OneHotEncoder(
+        keep_columns=True,
+        name="encoder",
+        use_cat_names=True,
+        cols=["neat"],
+        handle_unknown="ignore",
+    )
     assert ohe.fit(df) == ohe
     assert list(ohe.transform(df)) == [
         "neat_encoder_origin_0",
-        "neat_encoder_apple",
-        "neat_encoder_orange",
+        "neat_encoder_neat_apple",
+        "neat_encoder_neat_orange",
     ]
 
 

diff --git a/foreshadow/tests/test_transformers/test_smart.py b/foreshadow/tests/test_transformers/test_smart.py
@@ -54,7 +54,7 @@ def test_smart_encoder_less_than_30_levels():
     import scipy.stats as ss
 
     from foreshadow.transformers.smart import Encoder
-    from foreshadow.transformers.internals import OneHotEncoder
+    from foreshadow.transformers.externals import OneHotEncoder
 
     np.random.seed(0)
     leq_30_random_data = np.random.choice(30, size=500)

diff --git a/foreshadow/transformers/externals.py b/foreshadow/transformers/externals.py
@@ -9,7 +9,7 @@
 
 from sklearn.preprocessing import StandardScaler, RobustScaler, MinMaxScaler, Imputer
 from sklearn.decomposition import PCA
-from category_encoders import HashingEncoder
+from category_encoders import HashingEncoder, OneHotEncoder
 
 from .transformers import _get_modules
 

diff --git a/foreshadow/transformers/internals/encoders.py b/foreshadow/transformers/internals/encoders.py
diff --git a/foreshadow/transformers/smart.py b/foreshadow/transformers/smart.py
@@ -11,12 +11,13 @@
 from sklearn.pipeline import Pipeline
 
 from ..transformers.base import SmartTransformer
-from ..transformers.internals import BoxCox, FancyImputer, OneHotEncoder
+from ..transformers.internals import BoxCox, FancyImputer
 from ..transformers.externals import (
     MinMaxScaler,
     StandardScaler,
     RobustScaler,
     HashingEncoder,
+    OneHotEncoder,
 )
 
 
@@ -59,7 +60,12 @@ def _get_transformer(self, X, y=None, unique_num_cutoff=30, **fit_params):
         col_name = X.columns[0]
         unique_count = len(data.value_counts())
         if unique_count <= unique_num_cutoff:
-            return OneHotEncoder()
+            return OneHotEncoder(
+                cols=[col_name],
+                return_df=True,
+                use_cat_names=True,
+                handle_unknown="ignore",
+            )
         else:
             return HashingEncoder(n_components=30, cols=[col_name])