VisionEval 3.1.2 Public Release

This release supports (and defaults to) R 4.4.0 and earlier R 4.x versions. Installers for many of the most recent R versions are in the "Assets" section. You will need to install RTools44 in order to build VisionEval. RStudio should do (or offer to do) that for you if you set it to use R 4.4.0. See further instructions below.

Principal changes in this version include

• RTools44 required if you want to build an R 4.4.0 version (which is now the default, so you need RTools44 to "just build it"). If you want to build for an earlier R version, you can install (and may already have) an earlier RTools version, but you'll need to set RStudio to your chosen R Version or explicitly set the R version (for example) ve.build(r.version="4.3.3.") or set an operating system environment variable such as VE_R_VERSION=4.3.3 which will work for both ve.build and for the command line Makefile version.
• Better diagnostic messages for model input files (thanks to Minnesota DOT for helping us identify those);
• Better support for different versions of PowertrainsAndFuels using package aliases in visioneval.cnf
• Added an installModel version of VE-State that does year staging (one stage for base year, one for default future)
• Improved startup scripts
• Added the ability to have additional (non-specified) columns in .csv input files - previously, any field that was not in the module I/O specification would cause a fatal error. Now it just causes a warning and the input field is ignored. This change will help have "check" columns in the file (e.g. the total of separate proportions) so it is easy to check during development and you don't have to remove the field in order to run the model.
• VisionEval now strips trailing spaces from Azone, Bzone, Marea and Geo columns in the .csv input files. Some common data preparation approaches attach spaces to Azone or Bzone names, which prevents them from matching input rows (and which is also very hard to see or visualize in Excel). One less way to frustrate new users.
• Made the BaselineValue output field configurable in query exports using the longScenarios flag (That field is used with the Tableau sample format). Rather than automatically using the first version of the BaseYear query results, it can use any arbitrary model stage or year for the BaselineValues. The percentage comparisons in Tableau can then refer to the default future year scenario rather than the base year. Add the baseScenario parameter to the query export function with longScenarios=TRUE. baseScenario can be a one or two element character vector. First element is the name of a model stage, and the second element is the year to use (default is overall model's BaseYear). Do like this: openModel("myModel")$run()$query('fullQuery")$export(longScenarios=TRUE,baseScenario=c("future-year","2045") (you'll need to change the model name and decide what query file to use). Then look for a .csv file with the query results in results/outputs.

Contact Jeremy Raw (jeremy.raw@dot.gov) with questions or problems.