The interface of the fit() calibration method currently requires that users pass in a partially filled fitinfo dictionary, x, y, and emu. The method adds results to the dictionary. Typically the calibration methods also set x, y, and emu into fitinfo as part of the fit.
Conceptually, however, (emu, x, y, yvar, prior) are part of the specification of the calibration problem. Therefore, it makes more sense for users to set those into fitinfo and only pass that to fit().
The expectation is that this refactoring should not change results in any way if the same problem is run before and after using the same random seed.
The interface of the
fit()calibration method currently requires that users pass in a partially filledfitinfodictionary,x,y, andemu. The method adds results to the dictionary. Typically the calibration methods also setx,y, andemuintofitinfoas part of the fit.Conceptually, however,
(emu, x, y, yvar, prior)are part of the specification of the calibration problem. Therefore, it makes more sense for users to set those intofitinfoand only pass that tofit().The expectation is that this refactoring should not change results in any way if the same problem is run before and after using the same random seed.