modified lower_case.py

gators/encoders/count_encoder.py

@@ -63,7 +63,7 @@ class CountEncoder(_BaseEncoder):
 
     def __init__(self, columns=None, inplace=True):
         _BaseEncoder.__init__(self, columns=columns, inplace=inplace)
-        self.suffix = "frequency"
+        self.suffix = "count"
 
     def generate_mapping(
         self, X: DataFrame, y: Series = None

gators/encoders/tests/test_count_encoder_pd.py

@@ -55,9 +55,9 @@ def data_not_inplace():
             "B": ["Q", "W", "W"],
             "C": ["W", "Q", "W"],
             "D": [1.0, 2.0, 3.0],
-            "A__frequency": [2.0, 2.0, 1.0],
-            "B__frequency": [1.0, 2.0, 2.0],
-            "C__frequency": [2.0, 1.0, 2.0],
+            "A__count": [2.0, 2.0, 1.0],
+            "B__count": [1.0, 2.0, 2.0],
+            "C__count": [2.0, 1.0, 2.0],
         }
     )
     X_expected_numpy = pd.DataFrame(
@@ -66,9 +66,9 @@ def data_not_inplace():
             "B": ["Q", "W", "W"],
             "C": ["W", "Q", "W"],
             "D": [1.0, 2.0, 3.0],
-            "A__frequency": [2.0, 2.0, 1.0],
-            "B__frequency": [1.0, 2.0, 2.0],
-            "C__frequency": [2.0, 1.0, 2.0],
+            "A__count": [2.0, 2.0, 1.0],
+            "B__count": [1.0, 2.0, 2.0],
+            "C__count": [2.0, 1.0, 2.0],
         }
     ).astype(object)
     obj = CountEncoder(inplace=False).fit(X)

gators/feature_generation/__init__.py

@@ -5,8 +5,8 @@
 from .is_null import IsNull
 from .one_hot import OneHot
 from .plan_rotation import PlanRotation
-from .larger_than_bool import LargerThanBool
-from .smaller_than_bool import SmallerThanBool
+from .is_larger_than import IsLargerThan
+from .is_smaller_than import IsSmallerThan
 from .negation_bool import NegationBool
 from .polynomial_features import PolynomialFeatures
 from .polynomial_object_features import PolynomialObjectFeatures
@@ -20,8 +20,8 @@
     "ClusterStatistics",
     "PlanRotation",
     "PolynomialFeatures",
-    "LargerThanBool",
-    "SmallerThanBool",
+    "IsLargerThan",
+    "IsSmallerThan",
     "NegationBool",
     "PolynomialObjectFeatures",
 ]

gators/feature_generation/larger_than_bool.py → gators/feature_generation/is_larger_than.py

@@ -11,7 +11,7 @@
 from gators import DataFrame, Series
 
 
-class LargerThanBool(_BaseFeatureGeneration):
+class IsLargerThan(_BaseFeatureGeneration):
     """Create new columns based on the following definition: X[binarized_col] = X[col] >= bound.
 
     Parameters

gators/feature_generation/smaller_than_bool.py → gators/feature_generation/is_smaller_than.py

@@ -5,12 +5,12 @@
 
 from ..util import util
 
-from .larger_than_bool import LargerThanBool
+from .is_larger_than import IsLargerThan
 
 from gators import DataFrame
 
 
-class SmallerThanBool(LargerThanBool):
+class IsSmallerThan(IsLargerThan):
     """Create new columns based on the following definition: X[binarized_col] = X[col] >= bound.
 
     Parameters
@@ -64,9 +64,7 @@ class SmallerThanBool(LargerThanBool):
     def __init__(
         self, bounds_dict: Dict[str, List[float]], column_names: List[str] = None
     ):
-        LargerThanBool.__init__(
-            self, bounds_dict=bounds_dict, column_names=column_names
-        )
+        IsLargerThan.__init__(self, bounds_dict=bounds_dict, column_names=column_names)
         if not column_names:
             self.column_names = [
                 f"{col}__-inf_{bound}"

gators/feature_generation_str/split_extract.py

@@ -122,7 +122,7 @@ def transform(self, X: DataFrame) -> DataFrame:
         for col, idx, str_split, name in zip(
             self.columns, self.idx_split_vec, self.str_split_vec, self.column_names
         ):
-            X[name] = X[col].str.split(pad=str_split).str.get(idx).fillna("")
+            X[name] = X[col].str.split(str_split).str.get(idx).fillna("")
 
         return X
 

tests.rtf

@@ -627,7 +627,7 @@ gators/feature_generation_str/tests/test_string_contains_ks.py: 12 warnings
 gators/feature_selection/tests/test_select_from_model_dd.py::test_dd_np
   /Users/cpoli/gators311/lib/python3.11/site-packages/distributed/node.py:182: UserWarning: Port 8787 is already in use.
   Perhaps you already have a cluster running?
-  Hosting the HTTP server on port 53864 instead
+  Hosting the HTTP server on port 57890 instead
     warnings.warn(
 
 gators/feature_selection/tests/test_select_from_model_ks.py::test_ks
@@ -640,25 +640,25 @@ gators/feature_selection/tests/test_select_from_models_ks.py::test_ks_np
 gators/feature_selection/tests/test_select_from_models_dd.py::test_dd
   /Users/cpoli/gators311/lib/python3.11/site-packages/distributed/node.py:182: UserWarning: Port 8787 is already in use.
   Perhaps you already have a cluster running?
-  Hosting the HTTP server on port 53912 instead
+  Hosting the HTTP server on port 57937 instead
     warnings.warn(
 
 gators/feature_selection/tests/test_select_from_models_dd.py::test_dd
   /Users/cpoli/gators311/lib/python3.11/site-packages/distributed/node.py:182: UserWarning: Port 8787 is already in use.
   Perhaps you already have a cluster running?
-  Hosting the HTTP server on port 53939 instead
+  Hosting the HTTP server on port 57964 instead
     warnings.warn(
 
 gators/feature_selection/tests/test_select_from_models_dd.py::test_dd_np
   /Users/cpoli/gators311/lib/python3.11/site-packages/distributed/node.py:182: UserWarning: Port 8787 is already in use.
   Perhaps you already have a cluster running?
-  Hosting the HTTP server on port 53983 instead
+  Hosting the HTTP server on port 58005 instead
     warnings.warn(
 
 gators/feature_selection/tests/test_select_from_models_dd.py::test_dd_np
   /Users/cpoli/gators311/lib/python3.11/site-packages/distributed/node.py:182: UserWarning: Port 8787 is already in use.
   Perhaps you already have a cluster running?
-  Hosting the HTTP server on port 54010 instead
+  Hosting the HTTP server on port 58032 instead
     warnings.warn(
 
 gators/sampling/tests/test_supervised_sampling_dd.py::test_dd
@@ -699,4 +699,4 @@ gators/scalers/tests/test_yeo_johnson_pd.py::test_not_inplace_pd
     result = getattr(ufunc, method)(*inputs, **kwargs)
 
 -- Docs: https://docs.pytest.org/en/stable/how-to/capture-warnings.html
-=============== 830 passed, 215040 warnings in 434.00s (0:07:14) ===============
+=============== 830 passed, 215040 warnings in 437.66s (0:07:17) ===============