Fitting a linear or power-law continuum

[keflavich]

In the low frequency end of the ALMA bands, there can apparently be measurable spectral slope, which invalidates the 'flat continuum' assumption implicit in all of statcont. We should include a way to fit a linear or power-law model to the spectrum while simultaneously rejecting the emission/absorption lines. We can probably do this fitting with http://docs.astropy.org/en/stable/api/astropy.modeling.fitting.FittingWithOutlierRemoval.html#astropy.modeling.fitting.FittingWithOutlierRemoval.

[betacygni]

I have prepared the files that can be used as a test case, I will try to upload them here