ecmerkle/blavaan

*blavaan* broken after latest update of *lavaan*

Closed this issue · 6 comments

strengejacke commented

This code no longer works, and lavaan was recently updated on CRAN. I think that could be a reason for the code failure.

library(blavaan)
#> Loading required package: Rcpp
#> This is blavaan 0.5-4
#> On multicore systems, we suggest use of future::plan("multicore") or
#>   future::plan("multisession") for faster post-MCMC computations.
data("PoliticalDemocracy", package = "lavaan")

model <- "
    # latent variable definitions
    dem60 =~ y1 + a*y2
    dem65 =~ y5 + a*y6

    # regressions
    dem65 ~ dem60

    # residual correlations
    y1 ~~ y5
  "

bfit <- blavaan::bsem(
  model,
  data = PoliticalDemocracy,
  n.chains = 1,
  burnin = 50,
  sample = 100
)
#> Error: lavaan->lav_options_checkvalues():  
#>    invalid value in model.type option: "bsem".

Created on 2024-06-09 with reprex v2.1.0

TDJorgensen commented

Thanks for the reprex. Have you tried the code after installing the latest development version of blavaan from GitHub?

strengejacke commented

Seems to improve the situation :-)
One error remains, probably when printing:

library(blavaan)
#> Loading required package: Rcpp
#> This is blavaan 0.5-4.1279
#> On multicore systems, we suggest use of future::plan("multicore") or
#>   future::plan("multisession") for faster post-MCMC computations.
data("PoliticalDemocracy", package = "lavaan")

model <- "
    # latent variable definitions
    dem60 =~ y1 + a*y2
    dem65 =~ y5 + a*y6

    # regressions
    dem65 ~ dem60

    # residual correlations
    y1 ~~ y5
  "

bfit <- blavaan::bsem(
  model,
  data = PoliticalDemocracy,
  n.chains = 1,
  burnin = 50,
  sample = 100
)
#> 
#> SAMPLING FOR MODEL 'stanmarg' NOW (CHAIN 1).
#> Chain 1: 
#> Chain 1: Gradient evaluation took 0.000214 seconds
#> Chain 1: 1000 transitions using 10 leapfrog steps per transition would take 2.14 seconds.
#> Chain 1: Adjust your expectations accordingly!
#> Chain 1: 
#> Chain 1: 
#> Chain 1: WARNING: There aren't enough warmup iterations to fit the
#> Chain 1:          three stages of adaptation as currently configured.
#> Chain 1:          Reducing each adaptation stage to 15%/75%/10% of
#> Chain 1:          the given number of warmup iterations:
#> Chain 1:            init_buffer = 7
#> Chain 1:            adapt_window = 38
#> Chain 1:            term_buffer = 5
#> Chain 1: 
#> Chain 1: Iteration:   1 / 150 [  0%]  (Warmup)
#> Chain 1: Iteration:  15 / 150 [ 10%]  (Warmup)
#> Chain 1: Iteration:  30 / 150 [ 20%]  (Warmup)
#> Chain 1: Iteration:  45 / 150 [ 30%]  (Warmup)
#> Chain 1: Iteration:  51 / 150 [ 34%]  (Sampling)
#> Chain 1: Iteration:  65 / 150 [ 43%]  (Sampling)
#> Chain 1: Iteration:  80 / 150 [ 53%]  (Sampling)
#> Chain 1: Iteration:  95 / 150 [ 63%]  (Sampling)
#> Chain 1: Iteration: 110 / 150 [ 73%]  (Sampling)
#> Chain 1: Iteration: 125 / 150 [ 83%]  (Sampling)
#> Chain 1: Iteration: 140 / 150 [ 93%]  (Sampling)
#> Chain 1: Iteration: 150 / 150 [100%]  (Sampling)
#> Chain 1: 
#> Chain 1:  Elapsed Time: 0.096 seconds (Warm-up)
#> Chain 1:                0.174 seconds (Sampling)
#> Chain 1:                0.27 seconds (Total)
#> Chain 1:
#> Warning: The largest R-hat is 1.12, indicating chains have not mixed.
#> Running the chains for more iterations may help. See
#> https://mc-stan.org/misc/warnings.html#r-hat
#> Warning: Bulk Effective Samples Size (ESS) is too low, indicating posterior means and medians may be unreliable.
#> Running the chains for more iterations may help. See
#> https://mc-stan.org/misc/warnings.html#bulk-ess
#> Warning: Tail Effective Samples Size (ESS) is too low, indicating posterior variances and tail quantiles may be unreliable.
#> Running the chains for more iterations may help. See
#> https://mc-stan.org/misc/warnings.html#tail-ess
#> Computing post-estimation metrics (including lvs if requested)...
#> Error in if (verbose) cat(" done.\n"): argument is of length zero

Created on 2024-06-09 with reprex v2.1.0

ecmerkle commented

Thanks, I think the remaining error is related to this recent lavaan commit. It looks like they moved verbose (and warn and debug) out of the @options slot and now treat them as environment variables.

I just pushed a simple fix to blavaan and hope to get it on CRAN in the coming day. Though I worry this is just a patch over a larger issue with the verbose option. Will try to look more soon.

yrosseel commented

Oh dear. I am sorry for the trouble. Under the hood, many changes were made in lavaan 0.6-18, in particular wrt infrastructure and housekeeping.

It is somewhat unfortunate that we did not detect this. We did detect (and fix) many issues with other packages. The revdep checks (e.g., tools::check_packages_in_dir()) rely on the examples of the package, and (mostly testthat) tests, if available. But most examples in blavaan are enclosed in a \dontrun{} statement, and that is why we did not pick this up. Of course, I understand the reason for the \dontrun{} (sampling takes a long time), but to avoid this in the future: would it be possible to remove the `\dontrun{} in a few examples, and fit the model with a (very) small number of samples? Just to make sure it 'runs'?

ecmerkle commented

No problem! I have been seeing the changes to lavaan and expected that some changes to blavaan would be needed.

Yes, blavaan is now faster for many models than it was originally, so I will take a couple out of \dontrun. I also wonder what is the optimal thing for blavaan to do, for the new treatment of the verbose and warn and debug options. Right now, I just added verbose back to the Options slot to restore existing functionality.

ecmerkle commented

A new version is now on CRAN, which should fix the problems. (the submission process was unusually fast today)