handling of double censoring in fit_disc_gamma and related functions

[seabbs]

This issue is related to reconhub/distcrete#21

As mentioned there it looks to me like the handling of discretisation in this package (both in the gamma fitting functionality and elsewhere) is single censored versus double censored and so will induce some bias (see the linked issues for some related articles).

I think there are similar options here as to over in distcrete with the additional options of wrapping some of the fitting tools from primarycensored (as discussed here epiverse-trace/epiparameter#8)

[seabbs]

Coming back to this it also looks like this doesn't have support for right truncation adjustment but mentions use in outbreaks in the docs. Perhaps that should have a warning label or ideally be updated?