area2 - Return array of indices defining a 2d subarea of a 2d array.
source_anchor_y source_anchor_x
area_height area_width area_anchor_y area_anchor_x
area2 -> [2d-indices]
(Examples are illustrated):
Ex. 1: source array: (height=5), (width=10), anchor=(0,0)
subarea : height=3, width= 3, anchor=(0,0)
xoo.......
ooo.......
ooo.......
..........
..........
0 0 3 3 0 0 area2 -> [0 0 0 1 0 2
1 0 1 1 1 2
2 0 2 1 2 2]
Ex. 1b:source array: (height=5), (width=10), anchor=(2,2)
subarea : height=3, width= 3, anchor=(2,2)
ooo.......
ooo.......
oox.......
..........
..........
2 2 3 3 2 2 area2 -> [0 0 0 1 0 2
1 0 1 1 1 2
2 0 2 1 2 2]
Ex. 1c:Note that anchor point may lie outside both
arrays' bounds:
source array: (height=5), (width=10), anchor=(1,12)
subarea : height=3, width= 3, anchor=(1,12)
ooo.......
ooo....... x
ooo.......
..........
..........
1 12 3 3 1 12 area2 -> [0 0 0 1 0 2
1 0 1 1 1 2
2 0 2 1 2 2]
Ex. 2: source array: (height=6), (width=15), anchor=(2,5)
subarea : height=4, width= 5, anchor=(1,3)
...............
..ooooo........
..oooxo........
..ooooo........
..ooooo........
...............
2 5 4 5 1 3 area2 -> [1 2 1 3 1 4 1 5 1 6
2 2 2 3 2 4 2 5 2 6
3 2 3 3 3 4 3 5 3 6
4 2 4 3 4 4 4 5 4 6]
Ex. 3: Note that subarea doesn't need to lie
inside bounds of source array:
source array: (height=4), (width= 8), anchor=(4,-1)
subarea : height=2, width= 3, anchor=(1, 0)
........
........
........
ooo......
xoo
4 -1 2 3 1 0 area2 -> [3 -1 3 0 3 1
4 -1 4 0 4 1]
Given a -- hypothetical -- twodimensional array,
"area" tells you, what indices you need to
subscript a contiguous, twodimensional subarea.
The subarea is defined by specifying it's size
(width and height), as well as its location in the
source array. The location is defined by specifying
an anchor point in the source array as well as in
the subarea. Anchor points are matched, see
illustration, and examples below:
source array: height=6, width=15, anchor=(2,5)
subarea : height=4, width= 5, anchor=(1,3)
...............
..ooooo........
..oooxo........
..ooooo........
..ooooo........
...............
"area2" returns an array of TWODIMENSIONAL indices.
There is a SLI function called "area" returning
onedimensional indices, as well as the conversion
functions "cv1d" and "cv2d".
(For information on the order of subscription in NEST
arrays, see references below.)
In: "area2" takes six integer arguments (three pairs).
These arguments describe (1) the height and width of the
subarea to be indexed in the (hypothetical) source array, as
well as (2&3) an anchor point in each of the two arrays (see
illustration above):
source_anchor_y,
source_anchor_x: position of the anchor point relative
to ORIGIN OF THE SOURCE ARRAY
area_heigh t : height of the subarea to be subscribed
area_width : width of the subarea to be subscribed
area_anchor_y,
area_anchor_x : position of the anchor point relative
to ORIGIN OF THE SUBAREA
Out: "area" returns an array of ONEDIMENSIONAL indices:
[2d-indices] : flat integer array containing the indices
that can be used to subscript the
(hypothetical) source array in order to
access the desired subarea.
Indices are twodimensional. The returned
array is flat and has the following order:
[1y 1x 2y 2x 3y 3x ... ny nx]
That is, each pair of numbers indicates the
y- and the x-component of a respective
index.
The indices 1..n are returned in standard
NEST counting order. (For information on the
order of subscription in NEST arrays, see
references below.)
"area2" may raise the following SLI interpreter errors:
StackUnderflowError
ArgumentTypeError
No argument range check is performed. The returned
indices may indicate regions outside the source
array (see Example 3). This is not a bug, it's a
feature :-). Note that restrictions apply to the
related function "area".
However, anchor points may lie outside the array
bounds.
Note that arguments source_width and source_height are not used in
computation, and do not appear in the argument list.
(TO BE DONE: NEST layer indexing conventions)
Ruediger Kupper
/var/www/debian/nest/nest-simulator-2.20.0/sli/sliarray.cc