* `impute()` gains an `extra` argument.
Use it for observations not in the model that you still want to add in the
follow-up analysis.
For example: exclude rare observations from the model but you want them in the
aggregations.
* `impute()` on INLA models now also handles the binomial, the zero-inflated
Poison (type 0 and 1) and the zero-inflated negative binomial (type 0 and 1)
distributions.
* Add `hurdle_impute()` to fit a hurdle model based on a model of the presences
and a model of the counts.
* Added validation rules for `rawImputed` and `aggregatedImputed` objects.
* Update [`checklist`](https://inbo.github.io/checklist/) infrastructure.
