I want to get all neighbour values of a np.array.
The array looks like:
x = np.array([ [1, 2, 3, 4 ],
[5, 6, 7, 8],
[9, 10, 11, 12],
[13, 14, 15, 16] ])
What I have is:
i = 2
j = 2
n = x[i,j-1], x[i,j], x[i,j+1], x[i-1,j], x[i+1,j], x[i-1,j-1], x[i+1,j+1], x[i+1,j-1], x[i-1,j+1]
This returns (what I want)
(10, 11, 12, 7, 15, 6, 16, 14, 8)
But also got bugs for example when i want the neightbour values of
i = 3
j = 3
That gives:
Exception has occurred: IndexError
index 4 is out of bounds for axis 1 with size 4
An other soultion is:
def find_neighbors(m, i, j, dist=1):
return [row[max(0, j-dist):j+dist+1] for row in m[max(0,-1):i+dist+1]]
and
n = find_neighbors(x, i, j)
Which gives me an array of the neightbours but also gives me not all neightbours when I set
i = 0
j = 0
because it only gives me:
[array([1, 2]), array([5, 6])]
Does anybody have a solution for this?
Thank you!
You can take advantage of python indexing wrapping around for negative indices.
def wrap_nb(x,i,j):
return x[np.ix_(*((z-1, z, z+1-S) for z,S in zip((i,j), x.shape)))].ravel()
This requires i and j to be nonnegative and less than the shape of x.
If that is not guaranteed:
def wrap_nb(x,i,j):
return x[np.ix_(*(np.r_[z-1:z+2]%S for z,S in zip((i,j), x.shape)))].ravel()
Examples:
>>> wrap_nb(x,1,-2)
array([ 2, 3, 4, 6, 7, 8, 10, 11, 12])
>>> wrap_nb(x,0,-1)
array([15, 16, 13, 3, 4, 1, 7, 8, 5])
>>> wrap_nb(x,0,0)
array([16, 13, 14, 4, 1, 2, 8, 5, 6])