Download the stanford core NLP models from https://github.com/hippo-digital/tna-prototype/releases/tag/0.1-nlp and unzip the multipart zip into the FeatureExtractor project directory.

The solution assumes NEO4J is running in a docker container on the local machine. To create a neo4j docker instance run

docker pull neo4j 

docker run --publish=7474:7474 --publish=7687:7687 --volume=$HOME/neo4j/data:/data neo4j

Once the container is up and running you should be able to navigate to localhost:7474 and see the Neo4J admin console.

Open the solution file in Visual Studio / JetBrains Rider and run the GraphBuilder project. This will build a graph based on the C14242 series.

Once the load has finished you can now navigate to the Neo4j console and run a query to see a graph for example,

MATCH p=(n:location)-[:title|:nationality|:city|:person|:documenttype|:date|:location*0..2]-(a) WHERE n.name = "Crimea" or n.name="Hospital" RETURN DISTINCT n, collect(a)[..25]

should return something like

Optional

You can also run the FeatureExtractor project. This will output a set of JSON files to the processed folder in the root directory of this project. The GraphBuilder project will then read all of the files into the processed folder load all of the results to the Neo4J instance.