Add goodness of fit measure to fit methods

[sosey]

email discussions with a user, Renato Martone, resulted in a request to return additional information with the line_fit that would describe the chi-squared or p-value for the fit.

specifically, he was using the non-gui interface as a replacement for psfmeasure following the discussion in this issue: https://github.com/spacetelescope/imexam/issues/76