GSoC 2025 Project: Exploring and Abstracting Triplestore Alternatives

This repository hosts the work carried out as part of Google Summer of Code 2025 under the mentorship of GFOSS – Open Technologies Alliance. The goal of the project is to explore, test, benchmark, and abstract multiple RDF triplestore implementations into a unified Python interface.

Project Overview

Organization: GFOSS – Open Technologies Alliance

Project: Exploring and Abstracting Triplestore Alternatives

Contributor: Maria-Malamati Papadopoulou(goes by Maira Papadopoulou)

Mentor: Alexios Zavras

Citation

The released version of this software has been archived on Zenodo and is available under the following DOI:

DOI: 10.5281/zenodo.20759436

The Problem

RDF triplestores are essential for managing linked data and semantic web applications. However, the current ecosystem suffers from:

• Fragmentation: many triplestore implementations, each with its own API and quirks
• Steep learning curve: developers must repeatedly adapt to different query/load interfaces
• Limited abstraction: no unified layer to seamlessly switch between backends

This makes experimentation and adoption harder for developers, researchers, and organizations working with RDF data.

The Solution

Triplestore Abstraction Library: a Python package providing a unified API across multiple triplestores.

Core Components Delivered

Unified Python API: Consistent interface for loading RDF, querying with SPARQL, and modifying triples

Backend Implementations: Connectors for Jena, GraphDB, Blazegraph, AllegroGraph,and Oxigraph

Comprehensive Documentation: HOWTO.md, REFERENCE.md, and backend-specific configuration guides

Testing Framework: Automated pytest suite for backend validation

Project Characteristics

Implemented Features

• Unified Python API: Provides a single interface to interact with multiple triplestore backends.
• RDF Data Loading:
  • Supports loading data exclusively in Turtle (.ttl) format.
  • Large files can be ingested, with some backends offering streaming upload.
• SPARQL Query Execution:
  • Full support for all query forms: SELECT, ASK, CONSTRUCT, DESCRIBE.
  • Query results returned in Python-native structures (lists/dicts).
• Data Modification: add(), delete(), and clear() supported in all backends.
• Multiple Backends: Ready-to-use connectors for Jena, GraphDB, Blazegraph, AllegroGraph, Oxigraph.
• Error Handling: Consistent exception system for backend discovery, configuration errors, and runtime issues.
• Testing & Benchmarking:
  • Pytest-based test suite for all backends.
  • Benchmarking tools for comparing performance.

Optional / Backend-Specific Features

• Reasoning: Available only in backends that support inference (e.g., GraphDB, AllegroGraph).

📚 Documentation

• REFERENCE.md: Detailed API reference
• HOWTO.md: Usage and configuration guide
• alternatives.md: Candidate triplestores and their characteristics
• GSoC.md: Project Report
• BENCHMARKING.md: Benchmarking report

Candidates

The set of triplestore implementations that might be handled is listed in a separate file.

License

All code in this repository is licensed under the Apache-2.0 license.

Notice

Some of the contents may have been developed with support from one or more generative Artificial Intelligence solutions.