Source code for featuretools.primitives.standard.transform.binary.or_primitive
import numpy as np
from woodwork.column_schema import ColumnSchema
from woodwork.logical_types import Boolean, BooleanNullable
from featuretools.primitives.base.transform_primitive_base import TransformPrimitive
from featuretools.utils.gen_utils import Library
[docs]class Or(TransformPrimitive):
"""Performs element-wise logical OR of two lists.
Description:
Given a list of booleans X and a list of booleans Y,
determine whether each value in X is `True`, or
whether its corresponding value in Y is `True`.
Examples:
>>> _or = Or()
>>> _or([False, True, False], [True, True, False]).tolist()
[True, True, False]
"""
name = "or"
input_types = [
[
ColumnSchema(logical_type=BooleanNullable),
ColumnSchema(logical_type=BooleanNullable),
],
[ColumnSchema(logical_type=Boolean), ColumnSchema(logical_type=Boolean)],
[
ColumnSchema(logical_type=Boolean),
ColumnSchema(logical_type=BooleanNullable),
],
[
ColumnSchema(logical_type=BooleanNullable),
ColumnSchema(logical_type=Boolean),
],
]
return_type = ColumnSchema(logical_type=BooleanNullable)
commutative = True
compatibility = [Library.PANDAS, Library.DASK, Library.SPARK]
description_template = "whether {} is true or {} is true"
def get_function(self):
return np.logical_or
def generate_name(self, base_feature_names):
return "OR(%s, %s)" % (base_feature_names[0], base_feature_names[1])