Add tidy to sdmTMB_cv

[ericward-noaa]

Request to generate tables of quantities of interest: fold log likelihoods, etc

[seananderson]

I wrote a little print method for cross validation. Example:

#> fit

Cross validation of sdmTMB models with 2 folds.

Summary of the first fold model fit:

Spatial model fit by ML ['sdmTMB']
Formula: density ~ 0 + depth_scaled + depth_scaled2
Family: tweedie(link = 'log')
 
              coef.est coef.se
depth_scaled     -1.78    0.20
depth_scaled2    -1.49    0.12

Dispersion parameter: 13.20
Tweedie p: 1.62
Matérn range: 142.03
Spatial SD: 2.47
ML criterion at convergence: 3153.366

See ?tidy.sdmTMB to extract these values as a data frame.

Access the rest of the models in a list element named `models`.
E.g. `object$models[[2]]` for the 2nd fold model fit.

1 out of 2 models are consistent with convergence.
Figure out which folds these are in the `converged` list element.

Out-of-sample log likelihood for each fold: -3438.85, -3225.47.
Access these values in the `fold_loglik` list element.

Sum of out-of-sample log likelihoods: -6664.32 
More positive values imply better out-of-sample prediction.
Access this value in the `sum_loglik` list element.

I can also do a little tidy method.