Overview

• The Need

• High level Architecture

• Interoperability & Reuse

• Demo

• Conclusion

Need for parser modularity

• Rapid experimentation of new features
• Reproducibility in a SPARQL environment that grow in heterogeneity
• Maintainability of heterogeneous query tooling

Lexer and Parser using Chevrotain

1. Lexical Analysis (scanning): transform character stream into token stream.
2. Syntax Analysis (parsing): transform token stream into abstract syntax tree.
3. Semantic Analysis: validate constraints not enforced by grammar alone.

Constructing a simple lexer builder

© refactoring.guru

Define each grammar rule as an object

© refactoring.guru

Construct parser builder

© refactoring.guru

Interoperability & Reuse

We avoid big monolith using modular packaging

Demo

Adjust

Builtin Adjust Query

SELECT ?s ?p (ADJUST(
    ?o,
    "-PT10H"^^<http://www.w3.org/2001/XMLSchema#dayTimeDuration>
) as ?adjusted) WHERE {
  ?s ?p ?o
}

Construct Quad

Construct Quad Query

CONSTRUCT WHERE { GRAPH <wow> { ?s ?p ?o } }

Text Search

Text Search Query

SELECT * WHERE {
  ?s ?p ?o OPTION (score ?sc)
}

SELECT * WHERE {
  ?s ?p ?o OPTION (score ?sc)
}

No Optional

Additional

Complete SPARQL flow

Future Work

• SPARQL 1.2
• Modular Algebra Transformer
• Modular Round Tripping Generator
• Release V1