output conditional log likelihood and number of resampled particles in filter

[lwhittles]

Would be nice to have an idea of pinch points in fitting model over a data time series, in particular outputting the conditional log likelihood contribution of each data point as well as the number of particles being resampled to see where degeneracy is occuring.

If possible an option to output these in a similar way to saved trajectories in the fitting methods (e.g. pmcmc)