add_factor_prior

gammapy.astro.darkmatter.add_factor_prior(model, sigma, mu=1.0)

Attach a Log Normal nuisance prior on scale for J/D-factor uncertainty.

The J/D-factor is kept fixed at its nominal value; the associated uncertainty is instead expressed as an equivalent prior on scale, since the predicted flux depends only on the product scale * jfactor. Placing the prior directly on a second parameter (e.g. log10_jfactor) would make it perfectly degenerate with scale. This reparametrisation is a pure shift, so the prior retains the same shape and sigma, centered at scale = 1 instead of at the nominal log10(J).

Parameters:

modelDarkMatterSpectralModel

Model whose scale parameter will get the prior attached. scale is unfrozen as part of this call.

sigmafloat

Uncertainty on log10(J) (or log10(D)), in dex.

mufloat, optional

Center of the prior, in units of scale. Default is 1.0, i.e. the nominal J/D-factor value.

Returns:

modelDarkMatterSpectralModel

The same model instance, with the prior attached, for chaining.