LocalFeatureTransformer¶

The LocalFeatureTransformer class is a subclass of BaseFeatureTransformer and provides methods to fit and transform data using a feature transformer. Your local transformer should implement the following functions:

fit(self, config, X_data, y_data, *args, **kwargs)
transform(self, config, data, *args, **kwargs)

Note that the component configuration is passed into each method.

Configuration¶

Required Configuration¶

LocalFeatureTransformer contains the following required configuration:

feature_transformer_path: The path to the python module to use to transform features.
transformer_class: The class in the python module to use as the feature transformer.

Optional Configuration¶

LocalFeatureTransformer contains no optional configuration.

Default Configuration¶

LocalFeatureTransformer contains the following default configuration:

fit_func: The name of the function in feature_transformer_path to use to fit the transformer to the data. Defaults to "fit".
fit_params: Optional parameters to use to pass into the fit_func. Defaults to {}.
transform_func: The name of the function to use to transform data using the fit feature transformer crated from fit_func. Defaults to "transform".
transform_params: Optional parameters to use to pass into the transform_func when transforming data. Defaults to {}

Methods¶

fit¶

Fits the feature transformer with the given data.

def fit(data, *args, **kwargs)

Arguments:

data: The input data to fit the transformer.
*args: Variable length argument list.
**kwargs: Arbitrary keyword arguments.

Returns:

The fitted feature transformer.

transform¶

Transforms the given data using the fitted feature transformer.

def transform(data, *args, **kwargs)

Arguments:

data: The input data to transform.
*args: Variable length argument list.
**kwargs: Arbitrary keyword arguments.

Returns:

The transformed data.

fit_transform¶

Fits the feature transformer with the given data and then transforms the data.

def fit_transform(data, *args, **kwargs)

Arguments:

data: The input data to fit and transform.
*args: Variable length argument list.
**kwargs: Arbitrary keyword arguments.

Returns:

The transformed data.

Usage¶

from lolpop.component import LocalFeatureTransformer

conf = {
    #insert component configuration 
    config: {
        "feature_transformer_path": "/path/to/my_transformer.py"
    }
}

# Create an instance of LocalFeatureTransformer
transformer = LocalFeatureTransformer(conf=conf)

# Fit the transformer with data
data = ...
transformer.fit(data)

# Transform the data
transformed_data = transformer.transform(data)

print(transformed_data)

Sample Local Transformer¶

from sklearn.preprocessing import OneHotEncoder
import pandas as pd 

class MyTransformer(): 

    def fit(self, config, X_data, y_data, *args, **kwargs): 
        categorical_cols = config.get("categorical_columns")
        #encode categories
        self.ft = OneHotEncoder(handle_unknown="ignore")
        self.ft.fit(X_data[categorical_cols])
        return self.ft

    def transform(self, config, data, *args, **kwargs):
        categorical_cols = config.get("categorical_columns")
        transformed_data = self.ft.transform(data[categorical_cols]) 
        transformed_df = pd.DataFrame(transformed_data.toarray(), columns=self.ft.get_feature_names_out())
        data_out = pd.concat([transformed_df, data.drop(columns=categorical_cols)], axis=1)
        return data_out