remove weights, rename to inv_var_weights -- and add a second type of…

[ericward-noaa]

Proposing changing the weights argument to two arguments:
inv_var_weights for weights similar to how they're being used now. And likelihood_weights for the glmmTMB/sdmTMB/brms style

Example of long format data

s <- sim_dfa(num_trends = 1, num_years = 20, num_ts = 3)
obs <- c(s$y_sim[1, ], s$y_sim[2, ], s$y_sim[3, ])
long <- data.frame("obs" = obs, "ts" = sort(rep(1:3, 20)), "time" = rep(1:20, 3))
long$likelihood_weight <- runif(nrow(long),0.01,1) # these might be CVs for example
long$inv_var_w <- 1/(long$likelihood_weight ^ 2)

# fit model with weights in likelihood
m <- fit_dfa(y = long, data_shape = "long", iter = 50, chains = 1, likelihood_weights = "likelihood_weight")
# fit model with inv var weighting 
m <- fit_dfa(y = long, data_shape = "long", iter = 50, chains = 1, inv_var_weights = "inv_var_w")