[FIXED] How to crop zero edges of a numpy array?

Issue

I have this ugly, un-pythonic beast:

def crop(dat, clp=True):
    '''Crops zero-edges of an array and (optionally) clips it to [0,1].

    Example:
    >>> crop( np.array(
    ...       [[0,0,0,0,0,0],
    ...        [0,0,0,0,0,0],
    ...        [0,1,0,2,9,0],
    ...        [0,0,0,0,0,0],
    ...        [0,7,4,1,0,0],
    ...        [0,0,0,0,0,0]]
    ...     ))
    array([[1, 0, 1, 1],
           [0, 0, 0, 0],
           [1, 1, 1, 0]])
    '''
    if clp: np.clip( dat, 0, 1, out=dat )
    while np.all( dat[0,:]==0 ):
        dat = dat[1:,:]
    while np.all( dat[:,0]==0 ):
        dat = dat[:,1:]
    while np.all( dat[-1,:]==0 ):
        dat = dat[:-1,:]
    while np.all( dat[:,-1]==0 ):
        dat = dat[:,:-1]
    return dat
    # Below gets rid of zero-lines/columns in the middle
    #+so not usable.
    #dat = dat[~np.all(dat==0, axis=1)]      
    #dat = dat[:, ~np.all(dat == 0, axis=0)]

How do I tame it, and make it beautiful?

Solution

Try incorporating something like this:

# argwhere will give you the coordinates of every non-zero point
true_points = np.argwhere(dat)
# take the smallest points and use them as the top left of your crop
top_left = true_points.min(axis=0)
# take the largest points and use them as the bottom right of your crop
bottom_right = true_points.max(axis=0)
out = dat[top_left[0]:bottom_right[0]+1,  # plus 1 because slice isn't
          top_left[1]:bottom_right[1]+1]  # inclusive

This could be expanded without reasonable difficulty for the general n-d case.

Answered By - SCB

Answer Checked By - Willingham (PHPFixing Volunteer)