multiple sources per tile

[jeff-regier]

This PR lets us find two sources per tile rather than just one, thus closing #925. As long as encoder.two_layers is set to false, which it is by default, it should perform like before.

I made a few other miscellaneous changes while implementing this new functionality:

• A bug in filter_tile_catalog_by_flux is fixed (the catalog needed to be sorted by flux before filtering it)
• RegionEncoder is now contained entirely within case_studies/adaptive_tiling
• the number of PSF parameters and minimum CLAHE standard deviation are no longer hard coded so that @XinyueLi1012 can apply BLISS to DC2 without changing any code (after creating a custom config file)

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[codecov]

Codecov Report

Merging #942 (b4f1bca) into master (c54a697) will decrease coverage by 0.48%.
The diff coverage is 92.78%.

@@            Coverage Diff             @@
##           master     #942      +/-   ##
==========================================
- Coverage   96.25%   95.77%   -0.48%     
==========================================
  Files          25       26       +1     
  Lines        3254     3338      +84     
==========================================
+ Hits         3132     3197      +65     
- Misses        122      141      +19     

Flag	Coverage Δ
unittests	95.77% <92.78%> (-0.48%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

Files	Coverage Δ
bliss/convnet.py	100.00% <100.00%> (ø)
bliss/data_augmentation.py	98.68% <ø> (ø)
bliss/image_normalizer.py	93.75% <100.00%> (+0.32%)	⬆️
bliss/metrics.py	99.51% <ø> (ø)
bliss/plotting.py	85.07% <100.00%> (+1.74%)	⬆️
bliss/surveys/dc2.py	100.00% <100.00%> (ø)
bliss/train.py	93.26% <ø> (ø)
bliss/variational_layer.py	100.00% <100.00%> (ø)
bliss/catalog.py	99.34% <95.65%> (-0.32%)	⬇️
bliss/encoder.py	89.10% <86.20%> (-8.97%)	⬇️

📣 Codecov offers a browser extension for seamless coverage viewing on GitHub. Try it in Chrome or Firefox today!