Ellipticity

[evevkovacs]

If you are submitting a PR for a validation test (either adding a new one or updating an existing one), please post a link to the DESCQA web interface for the test you are submitting.
Plots for 2 variants of this test are here:
https://portal.nersc.gov/project/lsst/descqa/v2/?run=2018-02-23_24
https://portal.nersc.gov/project/lsst/descqa/v2/?run=2018-02-23_23
Contributing code and 2 yaml files. First variant plots e distributions for varoiuos morphologies and compares to COSMOS data. Second variant plots default ellipticity distributions for catalog.
Just tested on buzzard. Throws an error that needs to be fixed, but can be done in parallel with other code improvements.

[yymao]

@evevkovacs thanks for the PR.

I've pushed some changes (clean up the code to pass pylint checks, remove the binary tarball file, and rename the tests to ellipticity_dist and ellipticity_dist_COSMOS). See my changes in detail here.

Let me know if they look good. If so then we are ready to merge.

[yymao]

@evevkovacs BTW, are the validation datasets the raw data? I think we should just check in reduced data (like histograms).

[evevkovacs]

looked good, but then I just saw your latest test run and the ellipticity distribution test was skipped for protoDC2, and that shouldn't have happened.

[evevkovacs]

Hmm.. did I forget to commit the data? The files came from BenjaminJoachimi. I don't think they are too big. Maybe we can just use them for now and decide later?

[evevkovacs]

We won't be able to change the binning if we just save the histograms