runlmc.util.numpy_convenience module

Convenience functions for working with numpy arrays.

runlmc.util.numpy_convenience.begin_end_indices(lens)

runlmc.util.numpy_convenience.cartesian_product(*arrays)

runlmc.util.numpy_convenience.chunks(l, n)

Yield successive n-sized chunks from l.

runlmc.util.numpy_convenience.map_entries(f, nparr)

Map a function over a numpy array.

Parameters:

• f – single-parameter function over the same types
• nparr (np.ndarray) – arbitrary numpy array

Returns:

A numpy array with f evaluated on each element of the same shape.

runlmc.util.numpy_convenience.search_descending(x, xs, inclusive)

Parameters:

• x – threshold
• xs – descending-ordered array to search
• inclusive – whether to include values of x in xs

Returns:

the largest index index i such that xs[:i] >= x if inclusive else xs[:i] > x.

Raises:

ValueError – if array is not weakly decreasing

runlmc.util.numpy_convenience.smallest_eig(top)

Parameters:top – top row of Toeplitz matrix to get eigenvalues for Returns:the smallest eigenvalue

runlmc.util.numpy_convenience.symm_2d_list_map(f, arr, D, *args, dtype='object')

Symmetric map construction

runlmc.util.numpy_convenience.tesselate(nparr, lenit)

Create a ragged array by splitting nparr into contiguous segments of size determined by the length list lenit

Parameters:

• nparr – array to split along axis 0.
• lenit – iterator of lengths to split into.

Returns:

A list of size equal to lenit’s iteration with nparr’s segments split into corresponding size chunks.

Raises:

ValueError – if the sum of lengths doesn’t correspond to the array size.