Skip to content

Why Morph-KGC

Before starting with the development of Morph-KGC, we analyzed the performance and features of several knowledge graph construction engines, as described in this paper. Most of these engines presented issues when processing large volumes of data, as well as limited functionality or poor compliance with R2RML and RML. Morph-KGC has been designed with performance in mind, while remaining robust and feature-rich. In addition, it is currently the only engine that supports RML-star, enabling the generation of the emerging RDF-star data model.

Our Reasons to Recommend You to Use Morph-KGC

Performance

Morph-KGC relies on the usage of mapping partitioning to achieve efficient knowledge graph materialization. Morph-KGC can run mapping rules in parallel using the full power of the CPU. For scenarios that require to maintain the memory usage low, it is possible to use sequential processing, preventing the entire knowledge graph to be loaded in memory.

Additional optimizations are also implemented to increase efficiency: redundant self-join elimination, vectorized operations, hash joins and more.

W3C Compliance

Morph-KGC adopts the W3C Recommendation R2RML mapping language to map relational databases to RDF. In addition, it supports RML and RML-star, which are being further developed by the Knowledge Graph Construction W3C Community Group.

Reliability

Morph-KGC is being now used for all our knowledge graph construction projects at the Ontology Engineering Group, and other organizations are starting to adopt it as well for their RDF and RDF-star data materialization pipelines. This is why we put strong emphasis in keeping it stable, with solid releases. The engine is under continuous integration using R2RML test cases, RML test cases and RML-star test cases, in addition to more complex ones.

We also frequently test Morph-KGC in scenarios involving large volumes of data with the LUBM4OBDA, NPD and GTFS-Madrid benchmarks.

RDF-star

RDF-star extends the RDF data model with a compact alternative to the standard RDF reification that has been implemented by several triplestore vendors. Morph-KGC is currently the only knowledge graph construction engine implementing RML-star, allowing to generate RDF-star knowledge graphs in a systematic and declarative manner.

Free & Open Source

Morph-KGC is available under the permissive Apache License 2.0, which allows commercial use, modification, distribution, patent use and private use.

Integrated In

OEG UPM