sparqlkit 0.3.0

This package provides parsers and serializers for the SPARQL 1.1 Query and Update Language.

SPARQLKit

This package provides parsers and serializers for the SPARQL 1.1 Query Language.

Note: This is not a SPARQL processing engine.

Install

pip install sparqlkit

For CLI support:

pip install sparqlkit[cli]

CLI

SPARQLKit provides a command-line interface for formatting SPARQL files.

sparql --help
sparql -h

Commands

format

Format SPARQL files in-place:

sparql format path/to/query.rq
sparql format path/to/directory/

Check if files are formatted without making changes:

sparql format --check path/to/query.rq

Usage

import sparqlkit

query = r'''
PREFIX : <http://www.example.org/>
SELECT * WHERE { ?s ?p ?o }
'''

# Use the convenience function
formatted = sparqlkit.format_string(query)
print(formatted)

Preserving comments

Comments are preserved end-to-end through format_string and parse/serialize by default. To disable comment preservation, use preserve_comments=False:

import sparqlkit

query = "SELECT * WHERE { # comment\n  ?s ?p ?o }\n"
formatted = sparqlkit.format_string(query)
print(formatted)

tree = sparqlkit.parse(query)
print(sparqlkit.serialize(tree))

no_comments = sparqlkit.format_string(query, preserve_comments=False)
print(no_comments)

Notes:

• Comments are preserved using stable anchoring (nearby-token association), not exact original spacing.
• Comments are emitted as standalone lines by default for safety, but common inline forms are preserved:
  • SELECT ?x # comment (inline after a token)
  • WHERE { # comment (inline after {)
  • FILTER(... ) # comment (inline after ))

For advanced usage with the AST:

from sparqlkit.parser import sparql_query_parser
from sparqlkit.serializer import SparqlSerializer

tree = sparql_query_parser.parse(query)
serializer = SparqlSerializer()
result = serializer.visit_topdown(tree)
print(result)

Features

Iterative Stack-Based Serializer

The SPARQL serializer uses an iterative stack-based approach, allowing serialization of queries with arbitrary complexity and nesting depth (e.g., 1500+ nested OPTIONALs) without triggering Python's RecursionError.

Deep Nesting Example

from sparqlkit.parser import sparql_query_parser
from sparqlkit.serializer import SparqlSerializer

# Create a deeply nested query string
depth = 2000
query = "SELECT * WHERE { " + ("OPTIONAL { " * depth) + "?s ?p ?o" + (" }" * depth) + " }"

# Parse and serialize (no RecursionError)
tree = sparql_query_parser.parse(query)
serializer = SparqlSerializer()
result = serializer.visit_topdown(tree)
print(f"Successfully serialized query with nesting depth {depth}")

Extensibility

The serializer can be extended through subclassing to customize output:

from sparqlkit.serializer import SparqlSerializer
from lark import Tree

class CustomSerializer(SparqlSerializer):
    def _build_handler_map(self):
        handlers = super()._build_handler_map()
        handlers["var"] = {"enter": CustomSerializer._custom_var_enter, "exit": None}
        return handlers

    def _custom_var_enter(self, tree: Tree, context: dict) -> bool:
        self._parts.append(tree.children[0].value.upper())
        self._parts.append(" ")
        return True

Conformance

The parser and serializer passes all 1,070+ tests including those from the https://github.com/w3c/rdf-tests repository.