myfm.utils.encoders.DataFrameEncoder

class myfm.utils.encoders.DataFrameEncoder

Bases: object

Encode pandas.DataFrame into concatenated sparse matrices.

__init__() → None

Construct the encoders starting from empty one.

Methods

__init__()	Construct the encoders starting from empty one.
add_column(colname, encoder)	Add a column name to be encoded / encoder pair.
all_names()
encode_df(df)	Encode the dataframe into a concatenated CSR matrix.

Attributes

encoder_shapes

Show how the columns for an encoded CSR matrix are organized.

add_column(colname: str, encoder: myfm.utils.encoders.base.SparseEncoderBase) → myfm.utils.encoders.base.DataFrameEncoder

Add a column name to be encoded / encoder pair.

Parameters

• colname (str) – The column name to be encoded.

• encoder (SparseEncoderBase) – The corresponding encoder.

encode_df(df: pandas.core.frame.DataFrame) → scipy.sparse.csr.csr_matrix

Encode the dataframe into a concatenated CSR matrix.

Parameters

df (pd.DataFrame) – The source.

Returns

The result.

Return type

sps.csr_matrix

property encoder_shapes: List[int]

Show how the columns for an encoded CSR matrix are organized.

Returns

list of length of internal encoders.

Return type

List[int]