easy conversion to draws from rstantools format

I am struggling with converting posteriors I get from rstantools things like posterior_linpred to a draws object correctly. The problem is that the chain information gets dropped. Here is an example illustrating what I'd like to have:

library(posterior)
#> Warning: package 'posterior' was built under R version 4.1.2
#> This is posterior version 1.2.2
#> 
#> Attaching package: 'posterior'
#> The following objects are masked from 'package:stats':
#> 
#>     mad, sd, var


samp <- as_draws_matrix(example_draws())

## posterior_* functions from stan tools return matrices like

rstantools_samp <- matrix(as.matrix(samp), niterations(samp)*nchains(samp), nvariables(samp))
colnames(rstantools_samp) <- variables(samp)
head(rstantools_samp)
#>            mu      tau   theta[1]    theta[2]   theta[3] theta[4]    theta[5]
#> [1,] 2.005831 2.767367  3.9617520  0.27123540 -0.7431706 2.104805  0.92348879
#> [2,] 1.458316 6.979976  0.1237101 -0.06901539  0.9518270 7.281225 -0.06195211
#> [3,] 5.814947 9.677075 21.2510465 14.93055775  1.8290945 1.381443  0.53106337
#> [4,] 6.849586 4.788366 14.6996540  8.58618604  2.6749150 4.393232  4.75807198
#> [5,] 1.805168 2.848165  5.9600546  1.15573721  3.1088628 1.994890  0.76885094
#> [6,] 3.841243 4.083357  5.7601096  9.90920447 -0.9956266 5.328625  5.88894271
#>       theta[6]  theta[7]  theta[8]
#> [1,]  1.650237  3.320019  4.848542
#> [2,] 11.257502  9.621128 -8.640446
#> [3,]  7.155371 14.802013 -1.736363
#> [4,]  8.101547  9.491277  5.281551
#> [5,]  4.656270  1.208251 -4.540236
#> [6,] -1.701463  2.780403  7.075855

dim(rstantools_samp)
#> [1] 400  10

## things are order by chain, so we have

all(rstantools_samp[1:100,1] == subset_draws(samp, variable="mu", chain=1))
#> [1] TRUE
all(rstantools_samp[101:200,1] == subset_draws(samp, variable="mu", chain=2))
#> [1] TRUE

## now we should have a posterior function which lets me create from
## rstantools_samp a posterior draws thing which knows the number of
## chains. This does not work:

as_draws_matrix(rstantools_samp, .nchains=4)
#> # A draws_matrix: 400 iterations, 1 chains, and 10 variables
#>     variable
#> draw   mu tau theta[1] theta[2] theta[3] theta[4] theta[5] theta[6]
#>   1  2.01 2.8     3.96    0.271    -0.74      2.1    0.923      1.7
#>   2  1.46 7.0     0.12   -0.069     0.95      7.3   -0.062     11.3
#>   3  5.81 9.7    21.25   14.931     1.83      1.4    0.531      7.2
#>   4  6.85 4.8    14.70    8.586     2.67      4.4    4.758      8.1
#>   5  1.81 2.8     5.96    1.156     3.11      2.0    0.769      4.7
#>   6  3.84 4.1     5.76    9.909    -1.00      5.3    5.889     -1.7
#>   7  5.47 4.0     4.03    4.151    10.15      6.6    3.741     -2.2
#>   8  1.20 1.5    -0.28    1.846     0.47      4.3    1.467      3.3
#>   9  0.15 3.9     1.81    0.661     0.86      4.5   -1.025      1.1
#>   10 7.17 1.8     6.08    8.102     7.68      5.6    7.106      8.5
#> # ... with 390 more draws, and 2 more variables

^{Created on 2022-08-04 by the reprex package (v2.0.1)}

What I can do is to crudely set the nchains attribute to the number of chains. So I think the above should just work and give me a draws thing with 4 chains... this obviously requires documented formatting of the input samples to be column major sorted...

Start with the as_draws_matrix conversion shown in the issue and review how matrix dimensions and chain metadata are handled. Use the rstantools_samp example as the reproduction case; done means converting the chain-ordered matrix with .nchains=4 produces draws with four chains and preserves the expected values.