NEWS

Version 0.5.0

Update check for catch and lambda factors in predicted stock composition
Update selectivity names, using ‘total’ and ‘mature’ instead of ‘B’ and ‘SB’, respectively. For good practice, units are not necessarily in biomass.
Add option to fix index catchability in Dsurvey@qest_i
Update plot_index() to make facetted plots and plot_CAL() to plot aggregate length composition

Update calc_eqdist() where movement is indexed within the same season as stock distribution (previously applied to previous seasonal time step)

Use the potentially faster default pipe |> instead of magittr’s pipe %>%
Fix movement indexing in calc_population() (model previously had a 1 season lag that was erroneous)
Initial rec devs are length na if advance age after season 1 (obviously in seasonal models), otherwise remains length na-1
Remove some loops with apply() to speed up calc_F()
Length-age matrices, when modified by selectivity have a tiny number added to denominator to avoid division by zero
Update how selectivity arrays are filled in, fixes issue when time blocks are used
Use parallel package functions instead of snowfall for parallel computation with profiling and retrospectives
Fix indexing with predictions of stock composition by coercing a vector of 1’s and 0’s to boolean
Add do_jitter() function
Clean up various internal functions: like_comp(), optimize_RTMB(), get_sdreport()
Export get_likelihood_components()
Reporet fitting time in fit_MSA()

New selectivity options: constant over size and age range, mapping a subset of length or age from fleet to index
Add calc_init_population() for spool-up in spatial or seasonal models
Report most state variables and fits to data invisibly in plotting function
Add multivariate logitnormal likelihood to comp data
Update residual calculation for composition data
Fix predictions of tag transitions. Model previously had a 1 season lag that was erroneous.

More robust max F check when calc_F() to prevent numerical overflow (check in log space rather than normal space)
Various checks for NA’s in plotting functions
Fix internal function collapse_yearseason that converts year and season dimensions of arrays into single time dimension
Various fixes for cleaner console reporting
Profile function exports model object and parameter values to individual cores in parallel mode
Properly dispatch S4 generics for MSAassess (previously used S3 methods)
Remove stats::uniroot import for compatibility with RTMB 1.9