Redshift input, lens light, and basics for population transformers

[swagnercarena]

A number of changes including:

• Refactoring of xresnet code to allow for scalar inputs to fully connected layers. This also includes a new function in the dataset_generation module.
• Basic support for a population transformer. Still needs to be fleshed out more fully.
• Switch from int32 to uint32 for random seed
• Fix to the magnification calculation that also now accounts for lens light
• Fix to bug where draw_image would fail if magnification cut was not passed.
• A new distribution useful for lens light parameters. It allows the first parameter to be drawn randomly and the second parameter to be set to a noisy measurement of the first parameter.
• Extended distribution class to allow redshift sampling to include lens light.
• More informative warning when cross dict error informs you that something will be overwritten.
• COSMOS sources now take an optional parameter specifying the magnitude of the source
• GalaxyCatalog now k-corrects images
• A new source class that can be used to return a bulge+disk light parameterization for lens light
• New cosmology_utils functions that allow for k-correction calculations

[coveralls]

Pull Request Test Coverage Report for Build 2702049276

• 0 of 0 changed or added relevant lines in 0 files are covered.
• 25 unchanged lines in 4 files lost coverage.
• Overall coverage increased (+0.8%) to 95.073%

---

Files with Coverage Reduction	New Missed Lines	%
../../../../../opt/hostedtoolcache/Python/3.8.13/x64/lib/python3.8/site-packages/paltas/Sources/cosmos.py	1	99.17%
../../../../../opt/hostedtoolcache/Python/3.8.13/x64/lib/python3.8/site-packages/paltas/Sources/galaxy_catalog.py	2	96.72%
../../../../../opt/hostedtoolcache/Python/3.8.13/x64/lib/python3.8/site-packages/paltas/Analysis/train_model.py	8	91.82%
../../../../../opt/hostedtoolcache/Python/3.8.13/x64/lib/python3.8/site-packages/paltas/Configs/config_handler.py	14	95.11%
<!--	Total:	25		-->

Totals
Change from base Build 2619708538:	0.8%
Covered Lines:	2277
Relevant Lines:	2395

---

💛 - Coveralls

[swagnercarena]

Thanks for all the great comments. I made changes to the code to address all of them. As for the new fc_inputs model, it's still very much in flux. I added an activation and batch normalization as you suggested, but we can implement a more reasonable model once we know what works well.