Array interface

Note

This section characterizes protocols for backends. The purpose of a protocol is to provide support for nominal and structural subtyping.

Next to backends, dataset, and axes, nata provides objects to have an array interface. Inside nata, a interface is determined by HasArrayInterface protocol. It is especially important to allow dispatching for numpy.

class nata.types.HasArrayInterface(*args, **kwds)

Bases: typing_extensions.Protocol

Base protocol for an object to be ‘characterized’ as an array.

__array__(dtype: Optional[numpy.dtype] = None) → numpy.ndarray

Array interface of an object.

Method which will be called when numpy.array or similar function called with the object provided as an input.

data: numpy.ndarray

Represents the stored data for an object. It is of type numpy.ndarray.

dtype: numpy.dtype

Data type object of the stored data. It is of type numpy.dtype.

shape: Tuple[int, …]

Shape of the underlaying stored data. Similar to numpy.ndarray.shape

ndim: int

Dimensionality of the underlaying stored data. Similar to numpy.ndarray.ndim