add a function to extract regions from a 2D image

[awalter-bnl]

This commit adds a function <extract_regions> which extracts regions from a 2D array and returns a dictionary containing the regions as well as the arrays related to those regions.

[codecov-io]

Codecov Report

❗ No coverage uploaded for pull request base (master@bd0b95f). Click here to learn what that means.
The diff coverage is 100%.

@@            Coverage Diff            @@
##             master       #1   +/-   ##
=========================================
  Coverage          ?   49.04%           
=========================================
  Files             ?        5           
  Lines             ?      314           
  Branches          ?        0           
=========================================
  Hits              ?      154           
  Misses            ?      160           
  Partials          ?        0

Impacted Files	Coverage Δ
rixs/tests/test_preliminary_processing.py	100% <100%> (ø)
rixs/preliminary_processing.py	100% <100%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update bd0b95f...2d86efe. Read the comment docs.

[awalter-bnl]

FYI, I am waiting on the feedback, via basecamp, on weather this should be here or in NSLS-II-SIX/sixtools before removing the "Do not merge" tag.

[danielballan]

The return value of this function -- a dictionary with labels mapped to subarrays -- is not the most useful representation. It is more useful to pass around slices:

regions = {1: [slice(1, 2), slice(3, 4)], 2: [slice(3, 6), slice(2, 4)]}

or so-called label arrays, which are a "paint by number" representation of this array:

labels = np.zeros_like(arr)  # an array of zeros with the same dimensions as arr
for i, region.items()
    # Paint the subarray region with the number i.
    labels[region] = i

Then our label array labels looks like:

array([[0., 0., 0., 0., 0., 0., 0., 0., 0., 0.],
       [0., 0., 0., 1., 0., 0., 0., 0., 0., 0.],
       [0., 0., 0., 0., 0., 0., 0., 0., 0., 0.],
       [0., 0., 2., 2., 0., 0., 0., 0., 0., 0.],
       [0., 0., 2., 2., 0., 0., 0., 0., 0., 0.],
       [0., 0., 2., 2., 0., 0., 0., 0., 0., 0.],
       [0., 0., 0., 0., 0., 0., 0., 0., 0., 0.],
       [0., 0., 0., 0., 0., 0., 0., 0., 0., 0.],
       [0., 0., 0., 0., 0., 0., 0., 0., 0., 0.],
       [0., 0., 0., 0., 0., 0., 0., 0., 0., 0.]])

Which of these is better depends on that context. Scikit-image provides tools for switching between them.

[danielballan]

Recommended reading: http://scikit-image.org/docs/dev/user_guide/numpy_images.html

[awalter-bnl]

This is a good point, and much better form a memory usage point of view. I am about to move this stuff to NSLS-II-SIX/sixtools but will make this change at the same time.

[awalter-bnl]

closing this as it is being moved to NSLS-II-SIX/sixtools