Glossary - Technical Terms and Jargon explained

The Basic Formal Ontology is a so called upper ontology consisting of very general terms. BFO is adopted as a foundational ontology principally in the areas of biomedical ontology, security and defense (intelligence) ontology, and industry ontologies.

In the context of mapping, dictionaries are a common pattern for deciding the class of an entity based on the values in a database column. Ontopic Studio's built-in support for this pattern makes it easier to use this pattern, reducing typos and mismatches between values and classes.

A node in a knowledge graph. An entity is usually identified by an IRI.

A column (or collection of columns) that refers to the primary key (non-nullable uniquely identified values) in another table. It acts as a cross-reference between tables. A foreign key must always refer to a primary key somewhere.

The Internationalized Resource Identifier identifies a logical or physical resource used by web technologies. Example: http://data.mycompany.com/employee/1234.

In Ontopic Studio, a lens is a logical view of the original database table that allows you to visualize and transform the content without altering the original data stored in the data source.

In the case of a relational data source, a (schema) mapping connects tables and columns from the source to the components of the semantic layer, which for a knowledge graph are classes and properties. A common standard language for mapping relational databases to RDF graphs is R2RML, but other languages exist. A mapping consists of individual mapping entries.

A mapping entry is the detailed rule describing how to create a particular part of the knowledge graph from a well-defined part of the data source.

The creation of RDF triples in the knowledge graph. In this context, it is typically data from relational databases, from CSV files or spreadsheet files being transformed to a file of to RDF triples or ingested directly into specialized graph databases, also known as triple stores.

Ontop can build Virtual knowledge graphs as described by R2RML mappings or in the custom OBDA file format. Both file formats are equivalent in terms of expressiveness. The OBDA file format is preferred by many practitioners as it isperceived as easier to read.

The OBO ontologies are related to the life sciences and created and maintained by the Open Biological and Biomedical Ontologies (OBO) Foundry. OBO use BFO as an upper ontology.

Ontop translates queries from SPARQL, the query language used for RDF graphs, into SQL, the query language for traditional databases. Data from relational (traditional) databases therefore appears as a Knowlegde Graph, but without the need to copy and transform data from a relational database into a specialized graph database. This is then called a virtual knowledge graph. Ontop also allows you to transform batch-wise entire database tables, as well as entire databases, into RDF graphs. This modality is then called materialization. Ontop is Open Source software and is been developed by several universities and companies. Ontopic provides commercial support and services for Ontop.

In the context of knowledge graphs, an ontology defines classes, properties and the relationships between them and is used to provide a struture for data. Ontologies can also be used for automated reasoning.

Mappings from relational databases to RDF datasets can be expressed using the R2RML mapping language. R2RML is a W3C standard.

SPARQL is a language for querying RDF datasets. It is deliberately similar to SQL, but adapted to query graphs, whereas SQL is designed for querzing relational databases. SPARQL is a W3C standard.

A SPARQL endpoint is characterized by a process, typically associated with a database, that is capable of understanding SPARQL queries and either returning the requested data or modyfying the data, as specified in the query. SPARQL endpoints use the HTTP protocol for communication and each endpoint has a defined URL.

A template is used to generate subject IRIs for entities. Example: http://data.mycompany.com/employee/{empno}, where {empno} is replaced with a unique key in order to generate unique IRIs.

Answering SPARQL queries does not involve any materialization in RDF. Instead SPARQL queries are translated into SQL queries, which are executed over the data sources. One of the most widely used and advanced Virtual knowledge graph engines is Ontop.