Add tests demonstrating forward-slash behaviors in Turtle, JSON-LD, a…

…nd SPARQL (#1872) This patch adds test to verify that forward-slashes, and escaped forward slashes, work correctly in Turtle, JSON-LD and SPARQL. Some of the added tests check that SPARQL queries that contain seemingly escaped forward slashes as part of `PN_LOCAL` fail to execute as the SPARQL spec does not allow for escaped forward slashes in `PN_LOCAL`. Currently the RDFLib SPARQL processor does however allow escaped forward slashes as part of `PN_LOCAL`, so these tests have been marked with xfail as RDFLib should provide some way to parse SPARQL in strict mode, and in this mode leaniancies should be disabled and any non-standard SPARQL should be treated as invalid. Co-authored-by: Iwan Aucamp <[email protected]>
RDFLib · May 17, 2022 · ccb9c4a · ccb9c4a
1 parent 1e16caf
commit ccb9c4a
Show file tree

Hide file tree

Showing 8 changed files with 253 additions and 0 deletions.
diff --git a/test/data/variants/forward_slash-asserts.json b/test/data/variants/forward_slash-asserts.json
@@ -0,0  1,10 @@
 {
   "quad_count": 4,
   "exact_match": true,
   "has_subject_iris": [
     "http://example.org/kb/individual-a",
     "http://example.org/kb/individual-b",
     "http://example.org/ontology/core/MyClassA",
     "http://example.org/ontology/core/MyClassB"
   ]
 }
diff --git a/test/data/variants/forward_slash-variant-prefixed.jsonld b/test/data/variants/forward_slash-variant-prefixed.jsonld
@@ -0,0  1,33 @@
 {
   "@context": {
     "ex": "http://example.org/ontology/",
     "kb": "http://example.org/kb/",
     "owl": "http://www.w3.org/2002/07/owl#"
   },
   "_comment": [
     "The JSON-LD spec does not provide a grammar production rule set in,",
     "EBNF.  However, the section on compact IRIs indicates that an IRI can",
     "be prefixed at any point that would not result in a suffix starting",
     "with \"//\".  Hence, an unpaired forward slash, as a legal character of",
     "an IRI, can appear in the suffix component of a compact IRI.",
     "https://json-ld.org/spec/latest/json-ld/#compact-iris"
   ],
   "@graph": [
     {
       "@id": "kb:individual-a",
       "@type": "ex:core/MyClassA"
     },
     {
       "@id": "ex:core/MyClassA",
       "@type": "owl:Class"
     },
     {
       "@id": "kb:individual-b",
       "@type": "ex:core/MyClassB"
     },
     {
       "@id": "ex:core/MyClassB",
       "@type": "owl:Class"
     }
   ]
 }
diff --git a/test/data/variants/forward_slash-variant-prefixed.ttl b/test/data/variants/forward_slash-variant-prefixed.ttl
@@ -0,0  1,18 @@
 @prefix ex: <http://example.org/ontology/> .
 @prefix kb: <http://example.org/kb/> .
 @prefix owl: <http://www.w3.org/2002/07/owl#> .
 
 # Spell a class name with prefixing, but have the prefixing NOT include
 # one of the forward-slashed path components.
 # The forward slash must be escaped, according to Turtle grammar
 # production rules grammar rules including and between PN_LOCAL and
 # PN_LOCAL_ESC.
 # https://www.w3.org/TR/turtle/#sec-grammar-grammar
 
 ex:core\/MyClassA a owl:Class .
 
 kb:individual-a a ex:core\/MyClassA .
 
 ex:core\/MyClassB a owl:Class .
 
 kb:individual-b a ex:core\/MyClassB .
diff --git a/test/data/variants/forward_slash.jsonld b/test/data/variants/forward_slash.jsonld
@@ -0,0  1,24 @@
 {
   "@context": {
     "kb": "http://example.org/kb/",
     "owl": "http://www.w3.org/2002/07/owl#"
   },
   "@graph": [
     {
       "@id": "kb:individual-a",
       "@type": "http://example.org/ontology/core/MyClassA"
     },
     {
       "@id": "http://example.org/ontology/core/MyClassA",
       "@type": "owl:Class"
     },
     {
       "@id": "kb:individual-b",
       "@type": "http://example.org/ontology/core/MyClassB"
     },
     {
       "@id": "http://example.org/ontology/core/MyClassB",
       "@type": "owl:Class"
     }
   ]
 }
diff --git a/test/data/variants/forward_slash.nt b/test/data/variants/forward_slash.nt
@@ -0,0  1,4 @@
 <http://example.org/ontology/core/MyClassA> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://www.w3.org/2002/07/owl#Class> .
 <http://example.org/kb/individual-a> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://example.org/ontology/core/MyClassA> .
 <http://example.org/ontology/core/MyClassB> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://www.w3.org/2002/07/owl#Class> .
 <http://example.org/kb/individual-b> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://example.org/ontology/core/MyClassB> .
diff --git a/test/data/variants/forward_slash.ttl b/test/data/variants/forward_slash.ttl
@@ -0,0  1,10 @@
 @prefix kb: <http://example.org/kb/> .
 @prefix owl: <http://www.w3.org/2002/07/owl#> .
 
 <http://example.org/ontology/core/MyClassA> a owl:Class .
 
 kb:individual-a a <http://example.org/ontology/core/MyClassA> .
 
 <http://example.org/ontology/core/MyClassB> a owl:Class .
 
 kb:individual-b a <http://example.org/ontology/core/MyClassB> .
diff --git a/test/test_graph/test_variants.py b/test/test_graph/test_variants.py
@@ -49,14  49,26 @@ class GraphAsserts:
 
     quad_count: Optional[int] = None
     exact_match: bool = False
     has_subject_iris: Optional[List[str]] = None
 
     def check(
         self, first_graph: Optional[ConjunctiveGraph], graph: ConjunctiveGraph
     ) -> None:
         """
         if `first_graph` is `None` then this is the first check before any
         other graphs have been processed.
         """
         if self.quad_count is not None:
             assert self.quad_count == len(list(graph.quads()))
         if first_graph is not None and self.exact_match:
             GraphHelper.assert_quad_sets_equals(first_graph, graph)
         if first_graph is None and self.has_subject_iris is not None:
             subjects_iris = {
                 f"{subject}"
                 for subject in graph.subjects()
                 if isinstance(subject, URIRef)
             }
             assert set(self.has_subject_iris) == subjects_iris
 
 
 @dataclass(order=True)
@@ -219,6  231,8 @@ def test_variants(graph_variant: GraphVariants) -> None:
     assert len(graph_variant.variants) > 0
     first_graph: Optional[ConjunctiveGraph] = None
     first_path: Optional[Path] = None
     logging.debug("graph_variant.asserts = %s", graph_variant.asserts)
 
     for variant_key, variant_path in graph_variant.variants.items():
         logging.debug("variant_path = %s", variant_path)
         format = guess_format(variant_path.name, fmap=SUFFIX_FORMAT_MAP)

diff --git a/test/test_sparql/test_forward_slash_escapes.py b/test/test_sparql/test_forward_slash_escapes.py
@@ -0,0  1,140 @@
 #!/usr/bin/env python3
 
 # This software was developed at the National Institute of Standards
 # and Technology by employees of the Federal Government in the course
 # of their official duties. Pursuant to title 17 Section 105 of the
 # United States Code this software is not subject to copyright
 # protection and is in the public domain. NIST assumes no
 # responsibility whatsoever for its use by other parties, and makes
 # no guarantees, expressed or implied, about its quality,
 # reliability, or any other characteristic.
 #
 # We would appreciate acknowledgement if the software is used.
 
 """
 This test-set demonstrates usage of identifier prefixing and the
 forward-slash character in Turtle, JSON-LD, and SPARQL.  The motivating
 use case originated with attempts to interact with IANA Media Types as
 prefixed concepts, e.g. "application/json" somehow being
 "mime:application/json".
 """
 
 from test.data import TEST_DATA_DIR
 from test.utils.graph import cached_graph
 from typing import Set
 
 import pytest
 
 from rdflib import Graph
 from rdflib.plugins.sparql.processor import prepareQuery
 from rdflib.plugins.sparql.sparql import Query
 
 query_string_expanded = r"""
 SELECT ?nIndividual
 WHERE {
   ?nIndividual a <http://example.org/ontology/core/MyClassB> .
 }"""
 
 # NOTE: This is expected to fail.  The SPARQL grammar production rules
 # for prefixed IRIs, especially at production rule PN_LOCAL, have no way
 # to reach the forward-slash or backslash characters.
 # https://www.w3.org/TR/rdf-sparql-query/#grammar
 query_string_prefixed = r"""
 PREFIX ex: <http://example.org/ontology/>
 SELECT ?nIndividual
 WHERE {
   # NOTE: Syntax is incorrect - forward slash cannot be included in
   # local component of name.
   ?nIndividual a ex:core\/MyClassB .
 }"""
 
 PN_LOCAL_BACKSLASH_XFAIL_REASON = """
     Contrary to the ratified SPARQL 1.1 grammar, the RDFlib SPARQL propcessor
     accepts backslashes as part of PN_LOCAL which it treats as escape
     characters.
 
     There should be a way to instruct the SPARQL parser to operate in strict
     mode, and in strict mode backslashes should not be permitted in PN_LOCAL.
 
     See https://github.com/RDFLib/rdflib/issues/1871
 """
 
 
 def _test_query_prepares(query_string: str) -> None:
     """
     Confirm parse behavior of SPARQL engine when a concept would be
     prefixed at a point that leaves a forward-slash character in the
     suffix.
     """
     nsdict = {
         "ex": "http://example.org/ontology/",
         "kb": "http://example.org/kb/",
         "owl": "http://www.w3.org/2002/07/owl#",
     }
     # TODO: A 'strict' flag for prepareQuery is under consideration to
     # adjust parse behavior around backslash characters.
     query_object = prepareQuery(query_string, initNs=nsdict)
     assert isinstance(query_object, Query)
 
 
 def test_query_prepares_expanded() -> None:
     _test_query_prepares(query_string_expanded)
 
 
 @pytest.mark.xfail(reason=PN_LOCAL_BACKSLASH_XFAIL_REASON)
 def test_query_prepares_prefixed() -> None:
     with pytest.raises(ValueError):
         _test_query_prepares(query_string_prefixed)
 
 
 def _test_escapes_and_query(
     graph: Graph, query_string: str, expected_query_compiled: bool
 ) -> None:
     """
     Confirm search-results behavior of SPARQL engine when a concept
     would be prefixed at a point that leaves a forward-slash character
     in the suffix.
 
     Note that _test_query_prepares also exercises the expected parse
     failure.  This parameterized test is more for demonstrating that
     searching can work without prefixes.
     """
     expected: Set[str] = {
         "http://example.org/kb/individual-b",
     }
     computed: Set[str] = set()
 
     query_compiled: bool = False
     try:
         query_object = prepareQuery(query_string)
         query_compiled = True
     except Exception:
         pass
     assert expected_query_compiled == query_compiled
 
     for result in graph.query(query_object):
         computed.add(str(result[0]))
 
     assert expected == computed
 
 
 def test_escapes_and_query_turtle_expanded() -> None:
     graph = cached_graph((TEST_DATA_DIR / "variants/forward_slash.ttl",))
     _test_escapes_and_query(graph, query_string_expanded, True)
 
 
 @pytest.mark.xfail(reason=PN_LOCAL_BACKSLASH_XFAIL_REASON, raises=AssertionError)
 def test_escapes_and_query_turtle_prefixed() -> None:
     graph = cached_graph((TEST_DATA_DIR / "variants/forward_slash.ttl",))
     _test_escapes_and_query(graph, query_string_prefixed, False)
 
 
 def test_escapes_and_query_jsonld_expanded() -> None:
     graph = cached_graph((TEST_DATA_DIR / "variants/forward_slash.jsonld",))
     _test_escapes_and_query(graph, query_string_expanded, True)
 
 
 @pytest.mark.xfail(reason=PN_LOCAL_BACKSLASH_XFAIL_REASON, raises=AssertionError)
 def test_escapes_and_query_jsonld_prefixed() -> None:
     graph = cached_graph((TEST_DATA_DIR / "variants/forward_slash.jsonld",))
     _test_escapes_and_query(graph, query_string_prefixed, False)