Feature description
In most cases, the set of summary statistics and their dimensions should coincide between all simulations and the observed data. Thus, add a check somewhere (Model class?) for consistency, allowing for an optional override if that is not desirable (non-default), e.g. if simulation dimensions can vary.
Motivation/Application
E.g. numpy is quite generous due to broadcasting, which may hide errors.
Feature description
In most cases, the set of summary statistics and their dimensions should coincide between all simulations and the observed data. Thus, add a check somewhere (Model class?) for consistency, allowing for an optional override if that is not desirable (non-default), e.g. if simulation dimensions can vary.
Motivation/Application
E.g. numpy is quite generous due to broadcasting, which may hide errors.