pycottas is a library for working with compressed RDF files in the COTTAS format. COTTAS stores triples as a triple table in Apache Parquet. It is built on top of DuckDB and provides an HDT-like interface.
- Compression and decompression of RDF files.
- Querying COTTAS files with triple patterns.
- RDFLib store backend for querying COTTAS files with SPARQL.
- Supports RDF datasets (quads).
- Can be used as a library or via command line.
PyPI is the fastest way to install pycottas:
pip install pycottasWe recommend to use virtual environments to install pycottas.
import pycottas
from rdflib import Graph, URIRef
pycottas.rdf2cottas('my_file.ttl', 'my_file.cottas', index='spo')
res = pycottas.search('my_file.cottas', '?s <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> ?o')
print(res)
pycottas.cottas2rdf('my_file.cottas', 'my_file.nt')
# COTTASDocument class for querying with triple patterns
cottas_doc = pycottas.COTTASDocument('my_file.cottas')
# the triple pattern can be a string (below) or a tuple of RDFLib terms
res = cottas_doc.search('?s <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> ?o')
# COTTASStore class for querying with SPARQL
graph = Graph(store=pycottas.COTTASStore('my_file.cottas'))
res = graph.query('''
PREFIX rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>
SELECT DISTINCT ?s ?o WHERE {
?s rdf:type ?o .
} LIMIT 10''')
for row in res:
print(row)To execute via command line check the docs.
pycottas is available under the Apache License 2.0.
Universidad Politécnica de Madrid.
If you used pycottas in your work, please cite the ISWC paper:
@inproceedings{arenas2025cottas,
title = {{COTTAS: Columnar Triple Table Storage for Efficient and Compressed RDF Management}},
author = {Arenas-Guerrero, Julián and Ferrada, Sebastián},
booktitle = {Proceedings of the 24th International Semantic Web Conference, ISWC},
year = {2025},
}