Add local caching of slicing for data reduction

[mraspaud]

This PR avoids computing the slices for data reduction repeatedly when a bunch of datasets are resampled from the same area to the same area at the same time

• Tests added
• Tests passed
• Passes git diff origin/master -- "*py" | flake8 --diff
• Fully documented

[djhoese]

Could this (the use of a reductions dictionary) be cleaned up by using iter_by_area or something similar? Could the reduction functionality shown here then be shared with the other similar methods like crop and slice? Maybe something like an iterated slicing/reducing method?

[coveralls]

Coverage increased (+0.01%) to 78.303% when pulling 49604aa on mraspaud:feature-data-reduction-cache into ab4956a on pytroll:master.

[codecov]

Codecov Report

Merging #635 into master will increase coverage by 0.01%.
The diff coverage is 100%.

@@            Coverage Diff            @@
##           master    #635      +/-   ##
=========================================
+ Coverage   78.28%   78.3%   +0.01%     
=========================================
  Files         136     136              
  Lines       19723   19735      +12     
=========================================
+ Hits        15441   15453      +12     
  Misses       4282    4282

Impacted Files	Coverage Δ
satpy/scene.py	88.13% <100%> (+0.13%)	⬆️
satpy/tests/test_scene.py	99.45% <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 ab4956a...49604aa. Read the comment docs.

[mraspaud]

@djhoese 's comments are recorded in issue #637