Installing

This wiki gives information on how to get and install the Rosie project at the University of Michigan. These instructions are targeted toward Ubuntu Linux, although the base Rosie installation can work on any platform.

Soar

To use Rosie, you first need to have Soar installed. Note that currently the only way to get python3 SML support is building from source (only required if you want to use Rosie's python interface).

Option 1: Direct Download
You can download a copy from here. Extract it to the desired location, then set up the following environment variables:

# (Make sure that SOAR_HOME is set to the bin directory)
export SOAR_HOME=$HOME/path/to/soar/bin
export PYTHONPATH=$PYTHONPATH:$SOAR_HOME
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:$SOAR_HOME
export CLASSPATH=$CLASSPATH:$SOAR_HOME/java/sml.jar:$SOAR_HOME/SoarJavaDebugger.jar

Option 2: Build from Source
You need java and swig installed

git clone https://github.com/soargroup/soar
cd soar
python3 scons/scons.py all

Set the same environment variables as above, except that SOAR_HOME will be set to soar/out

Pysoarlib

If you want to use the Rosie python interface, you will need the pysoarlib library.

Simply clone https://github.com/amininger/pysoarlib and add the parent directory to the $PYTHONPATH. Note that it requires Soar to be built with the sml_python targeting python 3 and SOAR_HOME on the PYTHONPATH.

Rosie

git clone https://github.com/soargroup/rosie

Building Rosie:
Simply run ant in the rosie/tools/java directory, then ant in the rosie/java directory. (Make sure the CLASSPATH variables are already set)

Alternatively, in the rosie directory there is a file called source-rosie-env.sh. If you set the $ROSIE_HOME environment variable to the rosie path, and source that file, it will set up the environment variables automatically, as well as adding useful terminal commands you can run from anywhere.

export ROSIE_HOME=$HOME/path/to/rosie
source $ROSIE_HOME/source-rosie-env.sh

Once this is sourced, you can execute the terminal commands build_rosie_tools and build_rosie from anywhere.

Environment Variables:
As described in the previous section, if you set $ROSIE_HOME and source the source-rosie-env.sh file in your bash profile, these environment variables will be added automatically.

export CLASSPATH=$CLASSPATH:$ROSIE_HOME/java/rosie.jar
export CLASSPATH=$CLASSPATH:$ROSIE_HOME/tools/java/rosie-tools.jar
export CLASSPATH=$CLASSPATH:$ROSIE_HOME/tools/antlr-4.5-complete.jar

export PYTHONPATH=$PYTHONPATH:$ROSIE_HOME/python