Comunica: a Modular SPARQL Query Engine
for the Web

Querying Linked Data on the Web
is a multifaceted story

• Query languages and extensions
• Publication interfaces
• Querying algorithms
• ...

This variety of facets leads to comparability problems

Alternative algorithms are implemented in different engines → unfair!

Need for a flexible query engine
where various querying techniques
can be implemented and compared

Need for a flexible engine to compare alternative algorithms

Need for a flexible query engine
where various querying techniques
can be implemented and compared

Modularity using the actor model

Logic is separated into different actors.

Each actor independently performs a specific task.

• JSON-LD parsing actor
• SPARQL union actor A
• SPARQL union actor B

Actors can be combined for solving more complex tasks.

Actors subscribe to task-specific buses.

Mediators are responsible for selecting an actor to solve a certain task.

Example: Parse as fast as possible

Modules are wired together through semantic configuration files

Components.js: a semantic dependency injection framework.

Configuration files declare and parameterize actors, mediators and buses.

Implementation details

• Open-source under MIT license
• Implemented in TypeScript/JavaScript
• 140 separate modules (actors, buses, mediators)
• 2071 unit tests with a coverage of 100%
• 16 integration tests across different actors
• Continuous execution of the 268 SPARQL 1.1 Query spec tests
• Compatible with JavaScript API specification of RDFJS (W3C community group)

Current features

• SPARQL 1.1 support Passes 89% of spec tests (WIP)
• Federated querying of heterogeneous sources: Any combination of sources
• • Linked Data documents (Turtle, N-Triples, TriG, N-Quads, JSON-LD, RDFa)
  • Triple Pattern Fragments
  • SPARQL endpoints
  • HDT files
  • RDFJS Sources
  • SmartKG (WIP)

Using Comunica

• in your browser:

  http://query.linkeddatafragments.org/

• in JavaScript applications:

  Node.js: npm install @comunica/actor-init-sparql
  Browser: https://github.com/rdfjs/comunica-browser

• for development:

  https://github.com/comunica/comunica/

• → Documentation & tutorials:

  https://github.com/comunica https://comunica.readthedocs.io/en/latest/

Current usage of Comunica

• Research

  New querying algorithms and publishing interfaces

• Solid

  Authenticated querying across multiple data pods

• Improved developer experience

  Foundation for tools such as LDflex and GraphQL-LD

Comunica enables research

• Different trade-offs between server and client effort

  Triple Pattern Fragments and variants (versioning, Bloom filters, ...)

• New query language features

  SPARQL 1.2 community group

• Link-Traversal-Based Query Processing

  Live querying through Linked Data on the Web

Comunica enables Solid apps

• Lightweight client-side Web apps

  JavaScript runs directly in the browser
  Small bundle sizes by cherry-picking modules

• Querying private data

  Solid-specific HTTP actor for enabling authenticated requests

• Querying across multiple data pods

  Enabled by ability to federate over heterogeneous sources

Comunica enables high-level abstractions

Developers don't like writing SPARQL queries → developer-friendly abstractions

• LDflex

  DSL for querying Linked Data through JavaScript expressions
  Live playground

• GraphQL-LD

  Linked Data Querying with the highly popular GraphQL query language
  Live playground

• A modular framework for querying Linked Data on the Web.

https://comunica.linkeddatafragments.org/