Morph-KGC is an engine that constructs RDF knowledge graphs from heterogeneous data sources with the R2RML and RML mapping languages. Morph-KGC is built on top of pandas and it leverages mapping partitions to significantly reduce execution times and memory consumption for large data sources.

Features ✨

• User-friendly mappings with YARRRML.
• Transformation functions with RML-FNML, including Python UDFs.
• RDF 1.2 generation with RML 1.2.
• RML views over tabular data sources and JSON files.
• Integration with RDFLib, Oxigraph.
• Stateful functions with a shared context loaded once, such as reconciliation against a SKOS vocabulary or any knowledge graph behind a SPARQL endpoint.
• Optimized to materialize large knowledge graphs.
• Remote data and mapping files.
• Input data formats:
  • Relational databases: MySQL, PostgreSQL, Oracle, Microsoft SQL Server, MariaDB, SQLite.
  • Tabular files: CSV, TSV, Excel, Parquet, Feather, ORC, Stata, SAS, SPSS, ODS.
  • Hierarchical files: JSON, XML.
  • In-memory data structures: Python Dictionaries, DataFrames.
  • Cloud data lake solutions: Databricks, Snowflake.
  • Property graph databases: Neo4j.
  • HTTP APIs: JSON over HTTP, with the requests described in the mapping with the HTTP Vocabulary in RDF.

Documentation 📑

Read the documentation.

Tutorial 👩‍🏫

Learn quickly with the tutorial in Google Colaboratory!

Getting Started 🚀

PyPI is the fastest way to install Morph-KGC:

pip install morph-kgc

We recommend to use virtual environments to install Morph-KGC.

To run the engine via command line you just need to execute the following:

morph_kgc config.ini

Check the documentation to see how to generate the configuration INI file. Here you can also see an example INI file.

It is also possible to run Morph-KGC as a library with RDFLib and Oxigraph:

import morph_kgc

# generate the triples and load them to an RDFLib graph
g_rdflib = morph_kgc.materialize('/path/to/config.ini')
# work with the RDFLib graph
q_res = g_rdflib.query('SELECT DISTINCT ?classes WHERE { ?s a ?classes }')

# generate the triples and load them to Oxigraph
g_oxigraph = morph_kgc.materialize_oxigraph('/path/to/config.ini')
# work with Oxigraph
q_res = g_oxigraph.query('SELECT DISTINCT ?classes WHERE { ?s a ?classes }')

# the methods above also accept the config as a string
config = """
            [DataSource1]
            mappings: /path/to/mapping/mapping_file.rml.ttl
            db_url: mysql+pymysql://user:password@localhost:3306/db_name
         """
g_rdflib = morph_kgc.materialize(config)

License 🔓

Morph-KGC is available under the Apache License 2.0.

Author & Contact 📬

• Julián Arenas-Guerrero - julian.arenas.guerrero@upm.es

Ontology Engineering Group, Universidad Politécnica de Madrid.

Citing 💬

If you used Morph-KGC in your work, please cite the SWJ paper:

@article{arenas2024morph,
  title     = {{Morph-KGC: Scalable knowledge graph materialization with mapping partitions}},
  author    = {Arenas-Guerrero, Julián and Chaves-Fraga, David and Toledo, Jhon and Pérez, María S. and Corcho, Oscar},
  journal   = {Semantic Web},
  year      = {2024},
  volume    = {15},
  number    = {1},
  pages     = {1-20},
  issn      = {2210-4968},
  publisher = {IOS Press},
  doi       = {10.3233/SW-223135}
}

Sponsor 🛡️