Experiment

How to run

1. Install the requirements with

   pip install -r lsl-server/requirements.txt

2. Set the environment variables for the oTree image in docker-compose.yml.

   • PID: The participant ID number. This defines the order of the tasks and AI-assistance.
   • DEBUG: Toggles the debug mode. This defines whether continue buttons are available in the user interface.

3. Run the lsl-server and docker containers

   python lsl-server/server.py
   docker compose up --build

Components

• code-server

  This is the server for self-hosting VS Code inside a Docker container. By doing this, we are able to remove the headers that disallows us from embedding it in an iframe. It also allows us to pre-install the redhat.java extension and a JDK.

• exp-server

  This is a Python flask server, which acts as an endpoint, responsible for the behind-the-scenes actions for the experiment, such as running tests, setting the experiements up, and sending requests to the OpenAI API.

• java-src

  This directory contains the Java tasks for the experiment, including the UML diagrams and tests for the tasks.

• lsl-server

  This is a Python flash server, which handles the collection and streaming of Lab-Streaming-Layer (LSL) events. This must be run locally on the host machine for LabRecorder to capture the stream.

• otree

  This is the client, which manages the user interface that the experiment.

• nginx

  We use nginx as a reverse proxy to remove the headers for the code-server and to redirect requests within the docker network to the lsl-server, which runs on the host machine.

Prerequisites

Python and Docker

This experiment requires Python>=v3 and Docker.

Requestly

The browser extension Requestly should be installed, and two Header-modification rules should be added.

• Remove response header "content-security-policy"
• Remove response header "x-frame-options"

This is necessary for the iframes to work.

OpenAI API Key

The system running this should have a .env file in the root folder with an environment variable like this: OPENAI_API_KEY=<key here>.