arrlp.sortloop module

arrlp.sortloop(array, *arrays, axis=None, cuda=False, **kwargs)[source]

This generator allows to loop on unique values of an array rapidly by first sorting the array.

Parameters:

  • array (ndarray) – The array to determine unique values and slices.

  • *arrays (ndarrays) – Other arrays to slice in parallel with array.

  • cuda (bool) – True to use on GPU in cupy.

  • axis (int or None) – Axis along which to find unique values. If None, flattens the array.

Returns:

  • positions (array) – Positions corresponding to the current masked elements.

  • value (float or int) – first array value which was used for defining the masked region

  • *arrays – other array elements in the masked region

Examples

>>> from arrlp import sortloop
...
>>> for pos, r, y, x in sortloop(R, Y, X) :
...     result[pos] = scalar_function(y, x) # masked for each r