gloria.cast_series_to_kind

gloria.cast_series_to_kind(series, kind)

Cast a pandas.Series to a canonical NumPy dtype chosen by its one-letter dtype kind code.

This utility is useful for preparing data before calling Gloria.fit(), as each model supports different data types.

Parameters:

• series (pd.Series) – The data to cast.

• kind ({"u", "i", "f", "b"}) –

  One-letter code defined by numpy.dtype.kind. The mapping used here is:

  • "u": unsigned integer → uint64

  • "i": signed integer → int64

  • "f": floating point → float64

  • "b": boolean → bool

Returns:

The input series, cast to the corresponding dtype.

Return type:

pandas.Series

Raises:

ValueError – If kind is not one of the supported codes or the cast fails (e.g. due to incompatible values).

Examples

>>> cast_series_to_kind(pd.Series([1, 2, 3]), "f").dtype
dtype('float64')