dimod.as_samples¶

as_samples(samples_like, dtype=None, copy=False, order='C')[source]¶

Convert a samples_like object to a NumPy array and list of labels.

Parameters

samples_like (samples_like) – A collection of raw samples. samples_like is an extension of NumPy’s array_like structure. See examples below.
dtype (data-type, optional) – dtype for the returned samples array. If not provided, it is either derived from samples_like, if that object has a dtype, or set to the smallest dtype that can hold the given values.
copy (bool, optional, default=False) – If true, then samples_like is guaranteed to be copied, otherwise it is only copied if necessary.
order ({'K', 'A', 'C', 'F'}, optional, default='C') – Specify the memory layout of the array. See numpy.array().

Returns

A 2-tuple containing:

numpy.ndarray: Samples.

list: Variable labels

Return type

tuple

Examples

The following examples convert a variety of samples_like objects:

NumPy arrays

>>> import numpy as np
...
>>> dimod.as_samples(np.ones(5, dtype='int8'))
(array([[1, 1, 1, 1, 1]], dtype=int8), [0, 1, 2, 3, 4])
>>> dimod.as_samples(np.zeros((5, 2), dtype='int8'))
(array([[0, 0],
       [0, 0],
       [0, 0],
       [0, 0],
       [0, 0]], dtype=int8), [0, 1])

Lists

>>> dimod.as_samples([-1, +1, -1])
(array([[-1,  1, -1]], dtype=int8), [0, 1, 2])
>>> dimod.as_samples([[-1], [+1], [-1]])
(array([[-1],
       [ 1],
       [-1]], dtype=int8), [0])

Dicts

>>> dimod.as_samples({'a': 0, 'b': 1, 'c': 0}) 
(array([[0, 1, 0]], dtype=int8), ['a', 'b', 'c'])
>>> dimod.as_samples([{'a': -1, 'b': +1}, {'a': 1, 'b': 1}]) 
(array([[-1,  1],
        [ 1,  1]], dtype=int8), ['a', 'b'])

A 2-tuple containing an array_like object and a list of labels

>>> dimod.as_samples(([-1, +1, -1], ['a', 'b', 'c']))
(array([[-1,  1, -1]], dtype=int8), ['a', 'b', 'c'])
>>> dimod.as_samples((np.zeros((5, 2), dtype='int8'), ['in', 'out']))
(array([[0, 0],
       [0, 0],
       [0, 0],
       [0, 0],
       [0, 0]], dtype=int8), ['in', 'out'])