Before you start, don't forget to enable your conda environment using
source activate scientific
on Unix-based systems, or
activate scientific
on Windows. If you named your environment differently, you can figure out its name by running the command
conda info --envs
After you activated your environment, make sure to install all requirements of this current week by changing to your project-folder and running
pip install -r requirements.txt
For this week's bonus task you will do a exercise which will require you to understand shape and slice-access of numpy arrays.
Your task is to create a function subarray that will return a
sub-array of a given array. You can specify the shape of the new sub-array (how
many rows and columns) and the point it is centered on. If the sub-array
overlaps with the boundary of the original array, a fill-value can be provided,
otherwise the overlap should be omitted. The function signature should be as
follows:
import numpy as np
def sub_array(array, shape, center, fill=None):
# ... your code here
You can assume, that the input array is a numpy-array with arbitrary
values. The shape parameter will be a tuple with (rows, cols),
the center will be a tuple with the coordinates (row, col)
and fill is defaulted to None, but can be assigned to
any value that should be used for filling overlap. The function should return
the appropriate sub-array as a numpy-array. Here you can see a couple of
illustrations of the function's behavior:
Example 1: Input-array has shape (5, 6). Sub-array-center (orange) is (2, 2), sub-array (red) has shape (3, 3). The result is a 3x3 sub-patch.
Example 2: Input-array has shape (5, 6). Sub-array-center (orange) is (2, 2), sub-array (red) has shape (4, 3). The result is a 4x3 sub-patch. Note the location of the center for even number of rows or columns.
Example 3: Input-array has shape (5, 6). Sub-array-center (orange)
is (3, 2), sub-array (red) has shape (4, 5). Here fill=None. The result is a 3x5 sub-patch.
Example 4: Input-array has shape (5, 6). Sub-array-center (orange)
is (3, 4), sub-array (red) has shape (5, 5). Here fill=1. The result is a 5x5 sub-patch.
Good luck!