This repository creates the tooling project for the Triplifier.

• Soest J van, Choudhury A, Gaikwad N, Sloep M, Dumontier M, Dekker A (2019) Annotation of Existing Databases using Semantic Web Technologies: Making Data more FAIR. CEUR-WS, Edinburgh, pp 94–101

This tool can be executed in two modes:

• Stand-alone java runnable jar
• Docker container (and service) mode

For the runnable jar, it needs a computer with Java 8 runtime installed. For the docker container, it needs Docker Community Edition (native on Ubuntu, "for Windows" or "for Mac").

The basic configuration can run with the following command:

java -jar triplifier.jar -p triplifier.properties

The properties file mentioned here contains the JDBC connection information, several examples are displayed below, for different database systems.

PostgreSQL

jdbc.url = jdbc:postgresql://localhost/my_database
jdbc.user = postgres
jdbc.password = postgres
jdbc.driver = org.postgresql.Driver

MySQL

jdbc.url = jdbc:mysql://localhost/my_database
jdbc.user = user
jdbc.password = pass
jdbc.driver = com.mysql.cj.jdbc.Driver

Microsoft SQL Server

jdbc.url = jdbc:sqlserver://localhost;databaseName=my_database
jdbc.user = my_username
jdbc.password = my_password
jdbc.driver = com.microsoft.sqlserver.jdbc.SQLServerDriver

H2 database file

jdbc.url = jdbc:h2:file:/db/database_file_name
jdbc.user = sa
jdbc.password = sa
jdbc.driver = org.h2.Driver

SQLite database file

jdbc.url = jdbc:sqlite:/my.db
jdbc.user = user
jdbc.password = pass
jdbc.driver = org.sqlite.JDBC

Folder with CSV files

jdbc.url = jdbc:relique:csv:C:\\Users\\johan\\test?fileExtension=.csv
jdbc.user = user
jdbc.password = pass
jdbc.driver = org.relique.jdbc.csv.CsvDriver

To upload data directly into an RDF endpoint (and hence save the memory footprint), please add the following variables in the properties file (default: triplifier.properties):

• repo.type: Indicating the type of RDF endpoint. The following values are allowed for this variable:
  • "memory": this is the default, in-memory (and file-based) RDF/OWL output
  • "sparql": upload the triples directly in a SPARQL endpoint using the SPARQL protocol
  • "rdf4j": upload the triples directly in an RDF4j repository (e.g. a RDF4j or GraphDB database)
• repo.url: the URL of the RDF endpoint (not applicable or needed for repo.type = memory). In case of a SPARQL endpoint, include the URL of the actual SPARQL endpoint. In case of RDF4j, indicate the URL of the repository server
• repo.id: for repo.type = rdf4j indicate the repository ID
• repo.user: optional, specify the username for password-protected rdf4j repositories
• repo.pass: optional, specify the password for password-protected rdf4j repositories

An example of the full specification is given below:

repo.type = rdf4j
repo.url = http://localhost:7200
repo.id = data
repo.user = userNameTest
repo.pass = test

By default, the tool will generate an ontology file (ontology.owl) and a turtle file containing the materialized triples (output.ttl) relative to the execution folder. To change this, the following additional arguments can be used:

• -o <output_path_for_materialized_triples_file>
• -t <output_path_for_ontology_file>
• -c
• -b
• --ontologyAndOrData [ontology|data]

To run the triplifier as Docker container, you can run the following command:

On Linux/Unix/macOS systems:

docker run --rm \
   -v $(pwd)/output.ttl:/output.ttl \
   -v $(pwd)/ontology.owl:/ontology.owl \
   -v $(pwd)/triplifier.properties:/triplifier.properties \
   ghcr.io/maastrichtu-cds/triplifier:latest

On windows systems:

docker run --rm ^
   -v %cd%/output.ttl:/output.ttl ^
   -v %cd%/ontology.owl:/ontology.owl ^
   -v %cd%/triplifier.properties:/triplifier.properties ^
   ghcr.io/maastrichtu-cds/triplifier:latest

The example below shows how to run the container as a service, where the materialization process is called every interval time (defined by SLEEPTIME in seconds).

docker run --rm \
   -e SLEEPTIME=10 \
   --link graphdb:graphdb \
   -v $(pwd)/triplifier.properties:/triplifier.properties \
   ghcr.io/maastrichtu-cds/triplifier:latest

In this example, there is already a GraphDB docker container running, hence we can connect the docker containers. Therefore, the repo.url in the properties file should containthe hostname "graphdb", as inserted by the --link option. If the endpoint is running at a different location, you can specify the full URL of that location in the properties file, an omit the --link option.

An example of annotations (and insertions) can be found in the following repository:

• https://gitlab.com/UM-CDS/fair/data/maastro-rectum