This repository contains sources for the recommendation system ejob-o-mat. This work is part of the project "ELLI – Excellent Teaching and Learning in Engineering Sciences" and was funded by the federal ministry of education and research, Germany. This work was carried out at the IMA of RWTH Aachen University.
To set up the project, a clone of the repository is necessary. Afterwards, two options for running the recommendation system are available: Run as Docker container or run as Docker Compose service. In order to run the system as Docker container, a MongoDB database instance must be ready. For the Docker Compose solution, this precondition is not necessary: The necessary database gets started automatically.
First of all, the Docker container must be build. In order to do so, perform the following command:
docker build -t ejob-o-mat .
In order to run the system as Docker container, perform the following command:
docker run --name ejob-o-mat --link MY_MONGODB --restart=always -d -p 80:40000 -p 127.0.0.1:50000:50000 -e "dbHost=MY_MONGODB" -e "dbPassword=MY_PASSWORD" -e "dbUser=ejob-o-mat" ejob-o-mat
Here, the MongoDB database can be also run as Docker container or an external database can be used. Please complement the above command with the database’s properties. After starting the recommendation system, the public facing site runs at port 80
. Further, the internal administration site runs on port 50000
, but local to the Docker machine.
In order to run the system with Docker Compose, perform the following command from the base directory of this repository:
docker-compose up -d --build
Please change ROOT_PASS
and USER_PASS
in docker-compose.yml
locally to provide a safe password. After starting the recommendation system, the public facing site runs at port 80
. Further, the internal administration site runs on port 50000
, but local to the Docker machine.
After pulling the latest version the update process differs on whether Docker or Docker Compose is used to run the system. The database does not need to be touched during update.
Stop and remove the current container. All data is kept in the database and thus safe.
docker rm -f ejob-o-mat
Afterwards build the new container and start it as described above under Run with Docker Compose.
Updating with Docker Compose is very straightforward and works the same way as installing:
docker-compose up -d --build
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License. Commercial use of this project is not permitted. Please contact the IMA of RWTH Aachen University.
In order to get in contact with the author, please write a mail to thorsten.sommer@ima-zlw-ifu.rwth-aachen.de Please notice, that the author cannot provide support.