Create a .env file in the root (app) folder. Then edit the following key-value pairs and paste them into the paste them into the previously created env file.

SECRET_KEY = "4307fff64d15d8366384e22cb48f8d4f6fc3bd74c89eb19a1c93bdfc5f94434d"
ALGORITHM = "HS256"
ACCESS_TOKEN_EXPIRE_MINUTES = 120
API_ADMIN_KEY = "73d088b3ce965980dcb1ccf3b5b63b54cafb93232ab0889151c52207fe5a104b"
INITIAL_ADMIN_PW = "AlignStudiesPassword"
INITIAL_ADMIN_EMAIL = "norply@studyalign.com"

CORS_ORIGINS = "http://localhost:*, http://localhost, http://localhost:8080, http://localhost:63343, http://localhost:3333, http://localhost:3000, http://localhost:3366"

These key-value pairs are mainly used for authentication (issuing and validating of JWT tokens). In a production environment they would be replaced accordingly. Configuration of env variables might differ between hosting services.

Create a database.env file in the database directory and paste the content from below. Docker-compose reads values from these files. For a prod or dev environment you would create a database.prod.env or databse.dev.env.

Replace YOUR-USER and YOUR-PASSWORD with unique credentials (also in the SQLALCHEMY_DB_URL value).

POSTGRES_USER=YOUR-USER
POSTGRES_PASSWORD=YOUR-PASSWORD
POSTGRES_DB=studyalign
SQLALCHEMY_DB_URL=postgresql+psycopg2://YOUR-USER:YOUR-PASSWORD@studyalign-database/studyalign

To initialize the database cd backend and execute: docker exec -it studyalign-backend python /app/init_db.py On another instance, e.g. dev you would run: docker exec -it studyalign-backend-dev python /app/init_db.py

You can connect to the database with a tool like DataGrip. You will find the development credentials within database/database.env, e.g. using the url: jdbc:postgresql://localhost:5433/postgres. The port 5433 gets mapped to port 5432 in the database container.

You can run the StudyAlign backend but also LimeSurvey with the next steps. If you don't intent to use LimeSurvey, you can skip steps 2 and 3

1. Create a docker network by making the shell script create-network.sh executable sudo chmod +x create-network.sh and running it ./create-network.sh.
2. Start nginx: Open a new shell window, change the directory to loadbalancer, execute: docker-compose up
3. Start LimeSurvey: Edit credentials in limesurvey/docker-compose.yml. Open a new shell window, change the directory to limesurvey, execute: docker-compose up
4. Start StudyAlign: Open a new shell window, execute docker-compose up from the project's root folder
   1. On production, run: docker-compose -f docker-compose.prod.yml up
   2. On dev, run: docker-compose -f docker-compose.dev.yml up

Access limesurvey locally through http://localhost:8082. Access the StudyAlign backend locally through http://localhost:8080

Since we are using SQLalchemy with alembic for migrations, check out the official tutorial: https://alembic.sqlalchemy.org/en/latest/tutorial.html

docker exec -it studyalign-backend alembic revision -m "short description of migration"

When the docker containers are running:

docker exec -it studyalign-backend alembic upgrade head

This is executing all migrations up to the latest.

{
  "access_token": "eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJzdWIiOiJmNTc3NjVmNy0yYTIxLTRkMzMtYTNmMC1hYmNkZjdkZmM0MTIiLCJleHAiOjE1OTY3MjQzMzB9.tQ6clupKwY5YfmD7p7xQBWjx22O0dNAF5q6iCnVWZos",
  "refresh_token": "eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJzdWIiOiJmNTc3NjVmNy0yYTIxLTRkMzMtYTNmMC1hYmNkZjdkZmM0MTIiLCJyZWZyZXNoIjp0cnVlLCJleHAiOjkxMDE2MjQwMTZ9.wPaN3NMUU2n4ST3tjpuz0ZfPALK3dO3TZ6wzW9zPY7g",
  "token_type": "bearer"
}

We run all containers on one host.

The services limesurvey and studyalign-backend will be attached to the studyalign-network. As well, will be the nginx service studyalign-nginx. The nginx serves as a reverse proxy to balance network requests.

By executing docker network inspect studyalign-network from the shell you can see the details about the studyalign-network. The studyalign-network is accessible from outside, e.g. via port 443 (SSL). Whereas we isolate LimeSurvey's mariadb and StudyAlign's postgres and make it only available within a container.

The LimeSurvey instance, and the StudyAlign instance run in separate containers.

If you decide to host your machine learning model on a separate VM, use the study-align-prediction example project.

You can run the VM locally with docker-compose up as well. When adding a model to the database enter the URL http://172.17.0.1:8081/api/v1. This way the services are able to communicate with each other.

• relatedTarget property is optional and must be created manually => see createMouseInteraction() in frontend lib.