select_mock_targets -- desired updated features and bug fixes

[moustakas]

Current high-priority punch list.

• Ensure all targets -- including contaminants -- have the correct / desired target density (see ensure that the QSO target densities aren't too low #192 for QSOs).

• Use simqso to generate QSO spectra with the correct underlying QSO luminosity function (see preliminary support for using simqso to generate QSO templates desisim#293)

• The MWS photometry written out in the truth catalog may have an indexing bug (see MWS colors are oddly gridded in the dc17b truth catalogs #251).

• Fix duplicate TARGETIDs (fix duplicate TARGETIDs in select_mock_spectra #178).

• Add units to target and truth files (add units to target and truth files #237).

• Validate we can use the MOCKID in the output target catalog to get back to the correct row in the original mock catalog (Encode mock provenance into TARGETID. #88).

• Use GMMs based on DR5 and deconvolved for photometric scatter (see update Gaussian mixture models? #250).

• Compare columns in mock and data target catalogs and assess (related to add additional columns to the targets catalog produced by mock.build.targets_truth #209). For example, I've already added PSFDEPTH_[W1,W2] to the mock target catalog.

[moustakas]

Most of these issues were addressed in #264, @sbailey has opened new outstanding issues, and I'll open some more. Closing this for now.