uniform_partition_fromgrid

odl.discr.partition.uniform_partition_fromgrid(grid, min_pt=None, max_pt=None)

Return a partition of an interval product based on a given grid.

This method is complementary to uniform_partition_fromintv in that it infers the set to be partitioned from a given grid and optional parameters for min_pt and max_pt of the set.

Parameters

gridRectGrid

Grid on which the partition is based

min_pt, max_ptfloat, sequence of floats, or dict, optional

Spatial points defining the lower/upper limits of the intervals to be partitioned. The points can be specified in two ways:

float or sequence: The values are used directly as min_pt and/or max_pt.

dict: Index-value pairs specifying an axis and a spatial coordinate to be used in that axis. In axes which are not a key in the dictionary, the coordinate for the vector is calculated as:

min_pt = x[0] - (x[1] - x[0]) / 2
max_pt = x[-1] + (x[-1] - x[-2]) / 2

See Examples below.

In general, min_pt may not be larger than grid.min_pt, and max_pt not smaller than grid.max_pt in any component. None is equivalent to an empty dictionary, i.e. the values are calculated in each dimension.

See also

uniform_partition_fromintv

Examples

Have min_pt and max_pt of the bounding box automatically calculated:

>>> grid = odl.uniform_grid(0, 1, 3)
>>> grid.coord_vectors
(array([ 0. ,  0.5,  1. ]),)
>>> part = odl.uniform_partition_fromgrid(grid)
>>> part.cell_boundary_vecs
(array([-0.25,  0.25,  0.75,  1.25]),)

min_pt and max_pt can be given explicitly:

>>> part = odl.uniform_partition_fromgrid(grid, min_pt=0, max_pt=1)
>>> part.cell_boundary_vecs
(array([ 0.  ,  0.25,  0.75,  1.  ]),)

Using dictionaries, selective axes can be explicitly set. The keys refer to axes, the values to the coordinates to use:

>>> grid = odl.uniform_grid([0, 0], [1, 1], (3, 3))
>>> part = odl.uniform_partition_fromgrid(grid,
...                                       min_pt={0: -1}, max_pt={-1: 3})
>>> part.cell_boundary_vecs[0]
array([-1.  ,  0.25,  0.75,  1.25])
>>> part.cell_boundary_vecs[1]
array([-0.25,  0.25,  0.75,  3.  ])