Motivation

Continuous Deployment: Building an Integrated DevOps Environment

This project aims to achieve a fast and easy deployment for oTree experiments. oTree is a Python library that helps you to create experiments in the field of economics. The Karlsruhe Institute of Technology uses oTree in their decision & design lab. Currently, the provisioning of new experiments is outdated and mainly manual. We wanted the new approach to be low effort for experimenters and largely automated. For this, we laid out the following user story:

1. The experimenter clones a base project from the universities GitLab.
2. The user adds their experiment code to the repository and makes some minor configurations. They then push the code to a new project. This triggers a CI/CD pipeline.
3. The pipeline builds the new code and makes it ready for deployment. Deploying is done with a single click.
4. If deployed, the application is accessible via HTTPS under an assigned URL (e.g. experiment-name.k2lab.kit.edu).
5. If the experimenter updates their code, these changes trigger a new pipeline. After it succeeds, the user can deploy new changes to production.
6. Besides deploying new code, the user should be able to
   • delete their experiment production
   • reset their oTree database on the production server

This is a classic application of a DevOps approach. Changes happen iteratively and can be deployed with no overhead. Our goal is to build an automated deployment process for multiple oTree experiments, naturally we refer to our project as oForest.

How to Navigate this Documentation

We have tried to allow for non-linear progression through this documentation. We want to take different levels of knowledge into consideration. Therefore, we have separated concepts from tutorials and code examples and provide frequent cross-links. If you are not exactly sure what a concept means, we encourage you to click them and come back later.

Still, you will find a recommended next read at the end of most sections. This will provide you with some guidance on which topic to proceed with.

Read Next

Understand the DevOps approach and get to know some terminology.

Related

You want to check out a sample experiment repository? Check out this.