Request for option to simulate new data when NAs exist in Y

[Cole-Monnahan-NOAA]

Currently the package will not simulate new data when the original data point is NA. In some cases it could be useful to simulate new data based on the underlying estimate of x_tj so that a full time series of data could be generated. E.g., to do self-test simulation off a real data set.

If there is a way to do that currently please let me know.

Otherwise, would it be possible to add an option like ignore_NA (defaults to TRUE) in the simulate.dsem function that replaces NAs with dummy values when rebuilding the object here. This won't affect estimation but would give the user the option to get full simulated data sets. It looks like it would be needed to do rebuild the obj when resimulate_gmrf=FALSE as well.

If there is interest I could do a pull request with an example.

[James-Thorson-NOAA]

If you think it's important for our collaborations, I would definitely welcome a PR :) Hmm. You willing to it off of "add_priors" and I could merge those in at the same time?