Apollo

Apollo is a system for automatically assessing student's mastery of learning outcomes by looking at the code they write. This version of Apollo is built to assess Processing projects and the learning goals are adapted to the Creative Technology program at the University of Twente. Apollo integrates into Atelier to provide teachers and teaching assistants with direct feedback on uploaded programs.

This project is based on Atelier-PMD. The PMD rules for Processing used in this project are developed by Remco de Man and described in this paper. The conversion from Processing code to the Java code checked by PMD is written by Tim Blok for his Zita project, which is described in this paper.

Apollo only supports the submission hook for Atelier.

Configuration

This project exposes a single API endpoint, namely a webhook to integrate with the Atelier system. To use the system, it needs to know about your instance of Atelier and this application has to be registered with Atelier as a plugin.

Configuration of Atelier-PMD is done using a JSON configuration file, an example of which is located in the config/ folder. The location of the configuration file is set using the APOLLO_CONFIG environment variable. This configuration file contains all information the system needs to run as a plugin for Atelier:

• atelierHost: the url of your Atelier instance, without a trailing /
• atelierPluginUserID: the user ID you get in Atelier once you have registered the system as a plugin
• webhookSecret: a secret that Atelier will use to sign the webhook requests. Make sure that the value provided here is the same as the value configured in Atelier
• publicKey: the public key of this application, used for initial authentication
• privateKey: the private key corresponding with the public key
• storageLocation: file path for storing information about processed submissions

For each of these it is possible to get the values from an environment variable or a file on disk. To configure the host using an environment variable, use the following configuration:

"atelierHost": "ENV::ATELIER_HOST"

This will use an environment variable called ATELIER_HOST. Similarly, you can use a file:

"publicKey": "FILE::/run/secrets/keys/public.key"

If you specify a file path for the publicKey and privateKey variables and these files don't exist yet, Atelier-PMD will generate the keys and store them in the given locations.

Running with Docker

You can build the Docker image for Apollo using the docker build command:

docker build . -t atelier-pmd

The image exposes one volume in which you have to provide your configuration: /atelier-pmd/config. By default the configuration inside the container is called production.json, but you can override it by setting the APOLLO_CONFIG environment variable. The Tomcat server is exposed on port 8080, and the Atelier-PMD application can be reached under the /atelier-pmd path.

Debugging in IntelliJ

To debug Apollo in IntelliJ, you'll need to configure a Tomcat server in the Run/Debug Configurations. First create a new Local Tomcat Server configuration, and configure it to use your local Tomcat installation. Then you need to add the project to the server: go to the Deployment tab and add the apollo:war exploded artifact. Change the Application context to / to run the application at the root of the server.

You also need to make sure you have a valid configuration for Apollo: copy the config/example.json to config/development.json and change it to reflect your local environment. Then in the Startup/Connection tab you need to set the APOLLO_CONFIG to the full path to this development.json file. Make sure to do this for both the Run and Debug configurations.

You can now start debugging the application by starting the Tomcat configuration.