stan deprecation warning after compilation: bracket array declaration

[japhir]

I thought I'd give you a heads up that it appears that stan has deprecated the array declaration via brackets:

I get the following warning:
Declaration of arrays by placing brackets after a variable name is deprecated and will be removed in Stan 2.32.0. Instead use the array keyword before the type. This can be changed automatically using the auto-format flag to stanc.

after I run this ulam code:

partial_pool_both_cor_ulam <- ulam(
  alist(
    D47_final ~ dnorm(mu, sigma),
    mu <- a_bins[bins] + b_bins[bins] * d18O_PDB_vit,
    c(a_bins, b_bins)[bins] ~ multi_normal(c(a, b), Rho, sigma_bins),
    a ~ normal(0.6, .1),
    b ~ normal(0, 1),
    sigma_bins ~ exponential(1),
    sigma ~ dexp(1),
    Rho ~ lkj_corr(2)
  ),
  data = dat |>
    filter(!is.na(D47_final)) |> # we don't want to impute for now
    select(d18O_PDB_vit, bins, D47_final),
  cmdstan = TRUE, log_lik = TRUE,
  chains = 4, cores = 4, iter = 3000, warmup = 1000
)

which generates this piece of stan code:

data{
    vector[1336] D47_final;
    vector[1336] d18O_PDB_vit;
    int bins[1336];
}
parameters{
    vector[8] b_bins;
    vector[8] a_bins;
    real a;
    real b;
    vector<lower=0>[2] sigma_bins;
    real<lower=0> sigma;
    corr_matrix[2] Rho;
}
model{
    vector[1336] mu;
    Rho ~ lkj_corr( 2 );
    sigma ~ exponential( 1 );
    sigma_bins ~ exponential( 1 );
    b ~ normal( 0 , 1 );
    a ~ normal( 0.6 , 0.1 );
    {
    vector[2] YY[8];
    vector[2] MU;
    MU = [ a , b ]';
    for ( j in 1:8 ) YY[j] = [ a_bins[j] , b_bins[j] ]';
    YY ~ multi_normal( MU , quad_form_diag(Rho , sigma_bins) );
    }
    for ( i in 1:1336 ) {
        mu[i] = a_bins[bins[i]] + b_bins[bins[i]] * d18O_PDB_vit[i];
    }
    D47_final ~ normal( mu , sigma );
}
generated quantities{
    vector[1336] log_lik;
    vector[1336] mu;
    for ( i in 1:1336 ) {
        mu[i] = a_bins[bins[i]] + b_bins[bins[i]] * d18O_PDB_vit[i];
    }
    for ( i in 1:1336 ) log_lik[i] = normal_lpdf( D47_final[i] | mu[i] , sigma );
}

The warning occurs for line 4, column 4, which is the bins[1336] line.

[meconk]

I get this warning as well but in addition. My stan file created by ulam starts compiling but then errors because of a <lower=0> specification before the brackets.

[augias]

No clue if this is at all related, and if the deprecation warning has come into effect. Working through the text book on chapter nine, my first encounter with ulam() is a bust. Same code as #342 for model m9.1, but this error stan notes [cid] variable identifier is not in scope. Error reads:

> data(rugged)
> d<-rugged
> d$log_gdp<-log(d$rgdppc_2000)
> dd<-d[complete.cases(d$rgdppc_2000),]
> dd$log_gdp_std<-dd$log_gdp/mean(dd$log_gdp)
> dd$rugged_std<-dd$rugged/max(dd$rugged)
> dd$cid<-ifelse(dd$cont_africa==1,1,2)

> dat_slim<- list(
+   log_gdp_std<-dd$log_gdp_std,
+   rugged_std<-dd$rugged_std,
+   cid<-as.integer(dd$cid)
+ )

> m9.1 <- ulam(
+   alist(
+     log_gdp_std ~ dnorm( mu , sigma ) ,
+     mu <- a[cid] + b[cid]*( rugged_std - 0.215 ) ,
+     a[cid] ~ dnorm( 1 , 0.1 ) ,
+     b[cid] ~ dnorm( 0 , 0.3 ) ,
+     sigma ~ dexp( 1 )
+   ) , data=dat_slim , chains=1,cmdstan = TRUE)
Compiling Stan program...
Semantic error in '/tmp/Rtmp0QkL2I/model-1e4444cbe13df.stan', line 12, column 11 to column 14:
   -------------------------------------------------
    10:      b ~ normal( 0 , 0.3 );
    11:      a ~ normal( 1 , 0.1 );
    12:      mu = a[cid] + b[cid] * (rugged_std - 0.215);
                    ^
    13:      log_gdp_std ~ normal( mu , sigma );
    14:  }
   -------------------------------------------------

Identifier 'cid' not in scope. Did you mean 'add'?

make: *** [make/program:50: /tmp/Rtmp0QkL2I/model-1e4444cbe13df.hpp] Error 1

Error: An error occured during compilation! See the message above for more information.
In addition: Warning messages:
1: In is.na(symbols[[j]]$dims) :
  is.na() applied to non-(list or vector) of type 'symbol'
2: In is.na(symbols[[j]]$dims) :
  is.na() applied to non-(list or vector) of type 'symbol'

Is this to do with stan, or have I made a gross error somewhere in my code?
System info:

> cmdstanr::cmdstan_version()
[1] "2.30.1"
> packageVersion("rethinking")
[1] ‘2.21’
> version
               _                           
platform       x86_64-pc-linux-gnu         
arch           x86_64                      
os             linux-gnu                   
system         x86_64, linux-gnu           
status                                     
major          4                           
minor          1.2                         
year           2021                        
month          11                          
day            01                          
svn rev        81115                       
language       R                           
version.string R version 4.1.2 (2021-11-01)
nickname       Bird Hippie `

[rmcelreath]

You have <- instead of = in the dat_slim list. So there are no variable names. Use =.

[japhir]

I don't think that resolves the deprecation of the id, however, so I would leave the thread open until you refactor the ulam -> stan conversion?

So the above error was indeed not related ;-)

[rmcelreath]

Right, the Experimental branch has a candidate deprecation fix. I haven't had problems, but I do want to write some more tests, and there are some advanced functions that still trigger warnings.