New physics mean model extension: veiling from circumstellar disk

I am adding a new module to support inferences of an emission component from circumstellar disks. My assumptions for the purpose of this implementation are:

1. The emission component is spectrally smooth.
2. The emission can be characterized by two additional labels: a black body temperature, T_BB, and a solid angle, Omega_BB.

Veiling is often characterized by an r value, which represents the ratio of the flux in the smoothly varying excess component to the continuum flux in the stellar photosphere at a given wavelength or spectral band, e.g. K-band.

We wish to compute this r_K value for historical comparisons to literature and ease translation between theoretical parameters (T_BB and Omega_BB) and observational parameters (r_K). Such a computation should be calculated at the time of the likelihood call and then tabluated with the Arbitrary metadata blobs pattern in emcee.

This extension applies to phenomena beyond circumstellar disks. Any model that can be coarsely approximated by a black body + stellar photosphere will work. Circumstellar disks will have the additional limitation that the black body temperature cannot exceed the sublimation temperature of dust. This information can be included in a prior probability distribution function for T_BB.

The constraint on T_BB is likely to be very weak--and strongly degenerate with Omega_BB--- for narrow bandwidth spectra. But for whole-spectrum fitting (vis-a-vis Starfish Classic, Czekala et al. 2015) with high-resolution, high-bandwidth near-IR spectra (vis-a-vis IGRINS, Park et al. 2014), this approach could yield exceptional constraints on T_BB, assuming a black body model is sufficiently accurate to capture most of the departures in the line depths from spectral model predictions.

I'm calling this new mode star_BB.py modeled after the existing prototype star_veil.py, which employed a constant value for r_K over the input spectral bandwidth. The implementation with black bodies described here has the benefit of both providing r_K and delivering physically interpretable quantities, as noted above.

Review the existing prototype star_veil.py and the proposed star_BB.py mode, then trace where likelihood calls are made. Use the emcee arbitrary metadata blobs pattern as the reference for exposing r_K and the black-body quantities, and verify that the new mode supports the stated T_BB and Omega_BB parameters.