Installation

The following commands should install ToMCAT and its dependencies for users that use either MacPorts, Homebrew, or apt as their package manager.

git clone https://github.com/ml4ai/tomcat
cd tomcat && ./tools/install.sh

For people using other package managers (such as yum), note that ToMCAT depends on the following: CMake 3.15+, Boost 1.69+, a C++17 compatible compiler (tested with GCC 9 and AppleClang 11.0 so far), libfmt, doxygen, ffmpeg, OpenCV 4, dlib, Java 8, and Gradle.

Please make sure your internet connection is active while the install.sh script is running. macOS users may want to turn off their firewalls.

We don’t officially support Windows right now, but pull requests that add Windows support are welcome.

Running experiments

To run an experiment, run the following command from the tomcat directory.

./tools/run_session.sh

The data from the experiment will be saved in a folder whose name contains a timestamp corresponding to the start of the experiment, of the form

tomcat/data/participant_data/session_<Year>_<Month>_<Day>_<Hour>_<Minute>_<Second>

This folder will contain four files:

  • webcam_video.mpg : A video recording of the player’s face taken using the built-in webcam.

  • screen_video.mpg : A video recording of the Minecraft playthrough (i.e. what the player sees on their screen)

  • player_audio.wav : A WAV file containing the audio recording of the player.

  • malmo_data.tgz : A gzipped tarball containing Malmo data about player’s position, nearby entities, etc.

For developers

To speed up builds, create a file called gradle.properties and add org.gradle.daemon=true to it:

echo "org.gradle.daemon=true" > ~/.gradle/gradle.properties