• Fix broken tests on CRAN
  • Extensive refactoring of internal code
  • Don’t drop output dimensions when only one parameter saved
  • Summary method now returns the summary matrix
  • Add tests
  • Add vignette
  • Remove View function due to problematic interactions with RStudio
  • Improvements to traceplot and densityplot
  • Make sure output is reproducible when parallel=FALSE
  • Minor bugfixes
  • Fix issue when chains adapted for different numbers of iterations
  • Fix crash when DIC was requested but JAGS couldn’t calculate it
  • Stop auto-converting 1 row/1 col matrices into vectors
  • Improve pp.check plotting function
  • Add some warnings about functions and arguments that will be deprecated soon
  • Fix issues with NAs in parameters.
  • Fix missing arguments in autojags().
  • Handle errors in output processing so that samples are not lost completely.
  • Make sure a specified random seed is used for initial value functions.
  • Fix error where summary stats were printed in the wrong order for some parameters.
  • Don’t collapse 1 row/1col matrices into vectors automatically.
  • Fix bug with DIC output introduced in 1.4.7.
  • Minor changes to prep for CRAN submission.
  • Fix more issues with random seeds and problems with DIC.
  • Fix issue with setting seed affecting other random number generation.
  • Allow turning deviance/DIC monitoring on/off when updating an existing model.
  • Overhaul handling of adaptation; default to running a sequence of adapt() calls until JAGS reports adapation is adequate. Suggest minimum of 1000 adaptation iterations otherwise in man pages.
  • Change output of summary function. Add ability to control factories.
  • Change default approach to generating random seeds.
  • Add S3 method for View()
  • Convert traceplot() to an S3 method from S4.
  • Fix to allow running in parallel when dependencies are in non-standard libraries.
  • Fix bug when updating jagsbasic class.
  • Changes to NAMESPACE to get clean build check on R-devel.
  • Add ability to specify number of CPU cores to use when running in parallel.
  • Fix a problem where saving output for a single scalar parameter broke output processing.
  • Fix a problem where nonscalar estimated parameters with missing (i.e., non-estimated) values broke the display of summary stats.
  • Explicitly import functions from default packages to meet new standards for building with R-devel.
  • Change minimum required version of rjags to 3-13.
  • Add verbose argument to functions to allow suppression of all text output in the console as the function runs.
  • Change method for closing connection to clusters when running in parallel to avoid closing unrelated connections.
  • Adjust output for jags.basic.
  • Fix problem with max.iter argument in autojags(). Clarify documentation and output for autojags().
  • Add option save.all.iter in autojags() function to combine MCMC samples from all iterative updates into final posterior.
  • Fix issue with calculating stats that sometimes occurred with a constant posterior distribution.
  • Fix error where autojags did not handle NA rhat values properly.
  • Added more helpful warning when at least one rhat value = NA.
  • Initial CRAN upload.