processmaker4/executor-r
Script Task Executor Engine with R Runtime
This docker image provides a sandboxed protected environment to run custom R scripts that are written in ProcessMaker 4. User created script tasks should be isolated however have utilities available to them in order to get most common tasks done. This R environment has R packages available and autoloaded so Script Tasks can take advantage of the following libraries:
The execution requires a data.json, config.json and an output.json file be present on the host system. The data.json represents the Request instance data. The config.json represents configuration specific for this Script Task. And the output.json should be a blank file that will be populated by the successful output of the script task. The script task is represented by a script.r file. It is the responsibility of the caller to have these files prepared before executing the engine via command line (or docker API).
When writing a Script Task, three variables are available. They are:
Your script should execute quickly. Once the script is complete, the output variable will converted to JSON which will be stored in the output.json file. Once the docker execution is complete, you should use the return code of the docker execution. If the code is 0, then the script task executed successfully and you can read output.json for the valid output. If it is non-zero, then you should review STDERR to see the error that was displayed during execution.
{
"firstname": "Taylor"
}
output <- data
output$first_name <- toupper(output$first_name)
{"firstname":"TAYLOR"}
$ docker run -v <path to local data.json>:/opt/executor/data.json \
-v <path to local config.json>:/opt/executor/config.json \
-v <path to local script.r>:/opt/executor/script.r \
-v <path to local output.json>:/opt/executor/output.json \
processmaker/executor:r \
Rscript /opt/executor/bootstrap.r
docker pull processmaker4/executor-r