docarray.math.ndarray module#

docarray.math.ndarray.unravel(docs, field)[source]#
Return type

Optional[ForwardRef]

docarray.math.ndarray.ravel(value, docs, field)[source]#

Ravel value into doc.field of each documents

Parameters
  • docs (DocumentArray) – the docs to set

  • field (str) – the field of the doc to set

  • value (ArrayType) – the value to be set on doc.field

Return type

None

docarray.math.ndarray.get_array_type(array, raise_error_if_not_array=True)[source]#

Get the type of ndarray without importing the framework

Parameters

array (ArrayType) – any array, scipy, numpy, tf, torch, etc.

Return type

Tuple[str, bool]

Returns

a tuple where the first element represents the framework, the second represents if it is sparse array

docarray.math.ndarray.to_numpy_array(value)[source]#

Return the value always in numpy.ndarray regardless the framework type.

Return type

ndarray

Returns

the value in numpy.ndarray.

docarray.math.ndarray.to_list(value)[source]#
Return type

List[float]

docarray.math.ndarray.get_array_rows(array)[source]#

Get the number of rows of the ndarray without importing all frameworks

Parameters

array (ArrayType) – input array

Return type

Tuple[int, int]

Returns

(num_rows, ndim)

Examples

>>> get_array_rows([1,2,3])
1, 1
>>> get_array_rows([[1,2,3], [4,5,6]])
2, 2
>>> get_array_rows([[1,2,3], [4,5,6], [7,8,9]])
3, 2
>>> get_array_rows(np.array([[1,2,3], [4,5,6], [7,8,9]]))
3, 2
docarray.math.ndarray.check_arraylike_equality(x, y)[source]#

Check if two array type objects are the same with the supported frameworks.

Examples

>>> import numpy as np
    x = np.array([[1,2,0,0,3],[1,2,0,0,3]])
    check_arraylike_equality(x,x)
True
>>> from scipy import sparse as sp
    x = sp.csr_matrix([[1,2,0,0,3],[1,2,0,0,3]])
    check_arraylike_equality(x,x)
True
>>> import torch
    x = torch.tensor([1,2,3])
    check_arraylike_equality(x,x)
True
docarray.math.ndarray.detach_tensor_if_present(x)[source]#

Check if input is a dense torch array and detaches the tensor from the current graph. :param array: input array :rtype: Any :return: (num_rows, ndim)