This beginner's kit runs entirely in a Docker container so as to minimise the effort required to get started and the amount of software packages and frameworks to be installed. Docker Engine is the only piece of software required.
Install Docker Engine from https://www.docker.com.
Please, double-check that Docker has been given an adequate space for containers by checking under Settings -> Resources -> Virtual Disk limit. In our tests, we set this parameter to 64 GB. For the sake of completeness, our default configuration was set to 8 CPUs, 8 GB memory, and 1 GB swap.
First, git clone this repository, then open a Terminal window and get inside the cloned folder.
Once you are in position, run the following command
docker build --rm -t openaire-beginners-kit .
This will take care of everything is needed to create a Docker image; do not worry, building the image will take some time as Docker needs to fetch plenty of dependencies and try to virtualise an entire Apache Hadoop/Spark cluster on your local workstation. At the end, you will have a sandboxed environment with Python, Jupyter notebooks, Pandas, Hadoop, Spark and another bunch of modules installed. Easy peasy.
In order to start the container, open Docker Engine, locate the Docker image you just built and run it by clicking the Play icon (►), having care of specifying the port 8889 in the proper field.
You can also provide Docker with a name for the container, e.g., kit-container.
You can run the container from the Terminal too with the following command
docker run --name kit-container -p 8889:8889 openaire-beginners-kit
In order to stop the container, head back to Docker Engine, locate the running container and click on the Stop icon (■) in order to stop it.
You can stop the container from the Terminal too with the following command
docker stop kit-container
Later, you can go again to Docker Engine to restart the container by clicking the Play icon (►). This will resume the container as you left it (data included).
You can resume the container from the Terminal with the command
docker restart kit-container
To start playing with OpenAIRE data, you need to head to JupiterLab and locate the Jupyter notebook we prepared.
Upon starting the container you should see a bunch of logs.
Locate a line containing something like http://127.0.0.1:8889/lab?token=...; that's our guy, click on it.
The browser will open a page with JupyterLab.
Open the notebook beginners_kit.ipynb; this is where you can start playing with the OpenAIRE Graph.
Just follow the instructions provided in there.