use blanknodes

nbittich · Nov 23, 2024 · 420b412 · 420b412
1 parent dac8128
commit 420b412
Show file tree

Hide file tree

Showing 8 changed files with 259 additions and 149 deletions.
diff --git a/src/shared.rs b/src/shared.rs
@@ -12,6 +12,7 @@ pub const NS_TYPE: &str = "http://www.w3.org/1999/02/22-rdf-syntax-ns#type";
 pub const RDF_NIL: &str = "http://www.w3.org/1999/02/22-rdf-syntax-ns#nil";
 pub const RDF_FIRST: &str = "http://www.w3.org/1999/02/22-rdf-syntax-ns#first";
 pub const RDF_REST: &str = "http://www.w3.org/1999/02/22-rdf-syntax-ns#rest";
+#[allow(unused)]
 pub const DEFAULT_WELL_KNOWN_PREFIX: &str = "http://example.org/.well-known/genid#";
 pub static DATE_FORMATS: [&str; 9] = [
     "%Y-%m-%dT%H:%M:%S%.3f%Z",

diff --git a/src/triple_common_parser.rs b/src/triple_common_parser.rs
@@ -356,7 +356,7 @@ pub(crate) mod triple {
     pub(crate) fn labeled_bnode(s: &str) -> ParserResult<BlankNode> {
         let parse_label = delimited(
             tag(BLANK_NODE_LABEL),
-            take_while(|s: char| !s.is_whitespace()),
+            take_while(|s: char| !s.is_whitespace() && s != '.' && s != ';' && s != '<'),
             space0,
         );
 

diff --git a/src/turtle/turtle_doc.rs b/src/turtle/turtle_doc.rs
@@ -1,7 +1,8 @@
+use crate::grammar::BLANK_NODE_LABEL;
 use crate::shared::{
-    DATE_FORMATS, DEFAULT_DATE_FORMAT, DEFAULT_DATE_TIME_FORMAT, DEFAULT_TIME_FORMAT,
-    DEFAULT_WELL_KNOWN_PREFIX, RDF_FIRST, RDF_NIL, RDF_REST, TIME_FORMATS, XSD_BOOLEAN, XSD_DATE,
-    XSD_DATE_TIME, XSD_DECIMAL, XSD_DOUBLE, XSD_INTEGER, XSD_TIME,
+    DATE_FORMATS, DEFAULT_DATE_FORMAT, DEFAULT_DATE_TIME_FORMAT, DEFAULT_TIME_FORMAT, RDF_FIRST,
+    RDF_NIL, RDF_REST, TIME_FORMATS, XSD_BOOLEAN, XSD_DATE, XSD_DATE_TIME, XSD_DECIMAL, XSD_DOUBLE,
+    XSD_INTEGER, XSD_TIME,
 };
 use crate::triple_common_parser::{comments, Literal as ASTLiteral};
 use crate::triple_common_parser::{BlankNode, Iri};
@@ -40,7 +41,7 @@ fn get_uuid() -> String {
 
 struct Context<'a> {
     base: Option<&'a str>,
-    well_known_prefix: String,
+    well_known_prefix: Option<String>,
     prefixes: BTreeMap<&'a str, &'a str>,
 }
 #[derive(Serialize, PartialEq, Deserialize, Clone, Debug)]
@@ -88,6 +89,7 @@ pub enum Node<'a> {
     Literal(Literal<'a>),
     Ref(Arc<Node<'a>>),
     List(Vec<Node<'a>>),
+    LabeledBlankNode(String),
 }
 #[derive(PartialEq, PartialOrd, Debug, Clone)]
 pub struct Statement<'a> {
@@ -97,7 +99,7 @@ pub struct Statement<'a> {
 }
 #[derive(PartialEq, PartialOrd, Debug, Default)]
 pub struct TurtleDoc<'a> {
-    well_known_prefix: String,
+    well_known_prefix: Option<String>,
     base: Option<&'a str>,
     prefixes: BTreeMap<Cow<'a, str>, Cow<'a, str>>,
     statements: Vec<Statement<'a>>,
@@ -139,20 +141,35 @@ impl<'a> PartialEq for Node<'a> {
         match (self, other) {
             (Node::Iri(n1), Node::Iri(n2)) => n1.eq(n2),
             (Node::Literal(n1), Node::Literal(n2)) => n1 == n2,
-            (l @ Node::Literal(_) | l @ Node::Iri(_) | l @ Node::List(_), Node::Ref(n2)) => {
-                n2.as_ref().eq(l)
-            }
-
-            (Node::Ref(n1), r @ Node::Iri(_) | r @ Node::Literal(_) | r @ Node::List(_)) => {
-                n1.as_ref().eq(r)
-            }
+            (
+                l @ Node::Literal(_)
+                | l @ Node::Iri(_)
+                | l @ Node::List(_)
+                | l @ Node::LabeledBlankNode(_),
+                Node::Ref(n2),
+            ) => n2.as_ref().eq(l),
+
+            (
+                Node::Ref(n1),
+                r @ Node::Iri(_)
+                | r @ Node::Literal(_)
+                | r @ Node::List(_)
+                | r @ Node::LabeledBlankNode(_),
+            ) => n1.as_ref().eq(r),
             (Node::Ref(n1), Node::Ref(n2)) => n1 == n2,
             (Node::List(n1), Node::List(n2)) => n1 == n2,
+            (Node::LabeledBlankNode(n1), Node::LabeledBlankNode(n2)) => n1 == n2,
             (Node::Iri(_), Node::Literal(_))
+            | (Node::Iri(_), Node::LabeledBlankNode(_))
             | (Node::Iri(_), Node::List(_))
+            | (Node::LabeledBlankNode(_), Node::Iri(_))
+            | (Node::LabeledBlankNode(_), Node::Literal(_))
+            | (Node::LabeledBlankNode(_), Node::List(_))
             | (Node::Literal(_), Node::Iri(_))
+            | (Node::Literal(_), Node::LabeledBlankNode(_))
             | (Node::Literal(_), Node::List(_))
             | (Node::List(_), Node::Iri(_))
+            | (Node::List(_), Node::LabeledBlankNode(_))
             | (Node::List(_), Node::Literal(_)) => false,
         }
     }
@@ -407,8 +424,8 @@ impl<'a> TurtleDoc<'a> {
         turtle_values: Vec<TurtleValue<'a>>,
         well_known_prefix: Option<String>,
     ) -> Result<Self, TurtleDocError> {
-        let well_known_prefix =
-            well_known_prefix.unwrap_or_else(|| DEFAULT_WELL_KNOWN_PREFIX.to_string());
+        // let well_known_prefix =
+        //     well_known_prefix.unwrap_or_else(|| DEFAULT_WELL_KNOWN_PREFIX.to_string());
         let mut context = Context {
             base: None,
             well_known_prefix,
@@ -606,7 +623,6 @@ impl<'a> TurtleDoc<'a> {
         ctx: &'x Context,
         statements: &'x mut Vec<Statement<'a>>,
     ) -> Result<Node<'a>, TurtleDocError> {
-        let well_known_prefix = ctx.well_known_prefix.as_str();
         match value {
             v @ TurtleValue::Iri(_) | v @ TurtleValue::Literal(_) => {
                 let prefixes: BTreeMap<Cow<str>, Cow<str>> = ctx
@@ -618,11 +634,19 @@ impl<'a> TurtleDoc<'a> {
                 Self::simple_turtle_value_to_node(v, base, prefixes, true)
             }
             TurtleValue::BNode(BlankNode::Labeled(label)) => {
-                Ok(Node::Iri(Cow::Owned(well_known_prefix.to_owned() + label)))
+                if let Some(well_known_prefix) = ctx.well_known_prefix.as_ref() {
+                    Ok(Node::Iri(Cow::Owned(well_known_prefix.to_owned() + label)))
+                } else {
+                    Ok(Node::LabeledBlankNode(label.into()))
+                }
             }
             TurtleValue::BNode(BlankNode::Unlabeled) => {
                 let uuid = get_uuid();
-                Ok(Node::Iri(Cow::Owned(format!("{well_known_prefix}{uuid}"))))
+                if let Some(well_known_prefix) = ctx.well_known_prefix.as_ref() {
+                    Ok(Node::Iri(Cow::Owned(format!("{well_known_prefix}{uuid}"))))
+                } else {
+                    Ok(Node::LabeledBlankNode(uuid.replace("-", "")))
+                }
             }
             TurtleValue::Statement {
                 subject,
@@ -786,7 +810,14 @@ impl From<&Node<'_>> for RdfJsonNodeResult {
     fn from(value: &Node<'_>) -> Self {
         let typ_uri = "uri".into();
         let typ_literal = "literal".into();
+        let typ_bnode = "bnode".into();
         match value {
+            Node::LabeledBlankNode(bnode) => RdfJsonNodeResult::SingleNode(RdfJsonNode {
+                typ: typ_bnode,
+                datatype: None,
+                lang: None,
+                value: bnode.into(),
+            }),
             Node::Iri(iri) => RdfJsonNodeResult::SingleNode(RdfJsonNode {
                 typ: typ_uri,
                 datatype: None,
@@ -916,6 +947,9 @@ impl Display for Node<'_> {
             Node::Literal(Literal::Time(d)) => {
                 write!(f, r#""{}"^^<{}>"#, d.format(DEFAULT_TIME_FORMAT), XSD_TIME)
             }
+            Node::LabeledBlankNode(bnode) => {
+                write!(f, "{BLANK_NODE_LABEL}{bnode}")
+            }
             Node::List(list) => {
                 panic!("encountered node list where we shouldn't {list:?}");
             }
@@ -1214,6 +1248,15 @@ mod test {
         assert_eq!(9, triples.len());
     }
 
+    #[test]
+    #[serial]
+    fn test_labeled_bnode_err() {
+        FAKE_UUID_GEN.store(0, std::sync::atomic::Ordering::SeqCst);
+        let mut buf_c = String::new();
+        let turtle_c =
+            TurtleDoc::from_file("tests/labeled_bnode_err.ttl", None, &mut buf_c).unwrap();
+        assert_eq!(turtle_c.len(), 40)
+    }
     #[test]
     #[serial]
     fn other_test() {
@@ -1228,9 +1271,10 @@ mod test {
         println!("{}", turtle_expected.difference(&turtle_c).unwrap());
         assert_eq!(turtle_c.difference(&turtle_expected).unwrap().len(), 0);
     }
+
     #[test]
     #[serial]
-    fn complex_test() {
+    fn complex_test_with_bnode() {
         FAKE_UUID_GEN.store(0, std::sync::atomic::Ordering::SeqCst);
         let mut buf_c = String::new();
         let mut buf_e = String::new();

diff --git a/src/turtle/turtle_parser.rs b/src/turtle/turtle_parser.rs
@@ -261,7 +261,32 @@ mod test {
         );
     }
 
-    #[test] // NORDINE
+    #[test]
+    fn predicate_labeled_bnode_test() {
+        let s = r#"
+        <http://example.org/ns#ComplexResource> <http://example.org/ns#hasNestedObject> _:1.
+
+        "#;
+        let (rest, res) = statements(s).unwrap();
+        assert!(rest.trim().is_empty());
+        assert_eq!(
+            res,
+            vec![TurtleValue::Statement {
+                subject: Box::new(TurtleValue::Iri(Iri::Enclosed(
+                    "http://example.org/ns#ComplexResource",
+                ),)),
+                predicate_objects: [TurtleValue::PredicateObject {
+                    predicate: Box::new(TurtleValue::Iri(Iri::Enclosed(
+                        "http://example.org/ns#hasNestedObject",
+                    ),)),
+                    object: Box::new(TurtleValue::BNode(BlankNode::Labeled("1",),)),
+                },]
+                .into(),
+            },]
+        );
+    }
+
+    #[test]
     fn collection_test_from_other() {
         let s = r#"
         ex:techProducts ex:hasProducts ( ex:product1 ex:product2 )."#;

diff --git a/tests/expected_complex.ttl b/tests/expected_complex.ttl
@@ -4,18 +4,18 @@
 <http://example.org/ns#ComplexResource> <http://purl.org/dc/elements/1.1/created> "2024-11-22T12:34:56Z"^^<http://www.w3.org/2001/XMLSchema#dateTime>.
 <http://example.org/ns#ComplexResource> <http://example.org/ns#hasValue> "42"^^<http://www.w3.org/2001/XMLSchema#integer>.
 <http://example.org/ns#ComplexResource> <http://example.org/ns#hasValue> "forty-two"@en.
-<http://example.org/.well-known/genid#1> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://example.org/ns#NestedType>.
-<http://example.org/.well-known/genid#1> <http://example.org/ns#nestedProperty> "Nested Value"^^<http://www.w3.org/2001/XMLSchema#string>.
-<http://example.org/.well-known/genid#1> <http://example.org/ns#linksTo> <http://example.org/another-resource>.
-<http://example.org/ns#ComplexResource> <http://example.org/ns#hasNestedObject> <http://example.org/.well-known/genid#1>.
-<http://example.org/.well-known/genid#2> <http://www.w3.org/1999/02/22-rdf-syntax-ns#first> "Item1"^^<http://www.w3.org/2001/XMLSchema#string>.
-<http://example.org/.well-known/genid#3> <http://www.w3.org/1999/02/22-rdf-syntax-ns#first> "Item2"^^<http://www.w3.org/2001/XMLSchema#string>.
-<http://example.org/.well-known/genid#5> <http://example.org/ns#innerProperty> "Inner Value"^^<http://www.w3.org/2001/XMLSchema#string>.
-<http://example.org/.well-known/genid#4> <http://www.w3.org/1999/02/22-rdf-syntax-ns#first> <http://example.org/.well-known/genid#5>.
-<http://example.org/.well-known/genid#4> <http://www.w3.org/1999/02/22-rdf-syntax-ns#rest> <http://www.w3.org/1999/02/22-rdf-syntax-ns#nil>.
-<http://example.org/.well-known/genid#3> <http://www.w3.org/1999/02/22-rdf-syntax-ns#rest> <http://example.org/.well-known/genid#4>.
-<http://example.org/.well-known/genid#2> <http://www.w3.org/1999/02/22-rdf-syntax-ns#rest> <http://example.org/.well-known/genid#3>.
-<http://example.org/ns#ComplexResource> <http://example.org/ns#hasCollection> <http://example.org/.well-known/genid#2>.
+_:1 <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://example.org/ns#NestedType>.
+_:1 <http://example.org/ns#nestedProperty> "Nested Value"^^<http://www.w3.org/2001/XMLSchema#string>.
+_:1 <http://example.org/ns#linksTo> <http://example.org/another-resource>.
+<http://example.org/ns#ComplexResource> <http://example.org/ns#hasNestedObject> _:1.
+_:2 <http://www.w3.org/1999/02/22-rdf-syntax-ns#first> "Item1"^^<http://www.w3.org/2001/XMLSchema#string>.
+_:3 <http://www.w3.org/1999/02/22-rdf-syntax-ns#first> "Item2"^^<http://www.w3.org/2001/XMLSchema#string>.
+_:5 <http://example.org/ns#innerProperty> "Inner Value"^^<http://www.w3.org/2001/XMLSchema#string>.
+_:4 <http://www.w3.org/1999/02/22-rdf-syntax-ns#first> _:5.
+_:4 <http://www.w3.org/1999/02/22-rdf-syntax-ns#rest> <http://www.w3.org/1999/02/22-rdf-syntax-ns#nil>.
+_:3 <http://www.w3.org/1999/02/22-rdf-syntax-ns#rest> _:4.
+_:2 <http://www.w3.org/1999/02/22-rdf-syntax-ns#rest> _:3.
+<http://example.org/ns#ComplexResource> <http://example.org/ns#hasCollection> _:2.
 <http://example.org/ns#ComplexResource> <http://www.w3.org/2004/02/skos/core#note> "This resource demonstrates a variety of RDF features."@en.
 <http://example.org/ns#ComplexResource> <http://www.w3.org/2002/07/owl#sameAs> <http://example.org/ns#AliasResource>.
 <http://example.org/ns#Type1> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://www.w3.org/2002/07/owl#Class>.
@@ -24,27 +24,27 @@
 <http://example.org/ns#SuperType> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://www.w3.org/2002/07/owl#Class>.
 <http://example.org/ns#SuperType> <http://www.w3.org/2000/01/rdf-schema#label> "Super Type"@en.
 <http://example.org/ns#SuperType> <http://www.w3.org/2000/01/rdf-schema#comment> "A superclass for demonstration purposes."@en.
-<http://example.org/.well-known/genid#6> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://example.org/ns#AnonymousType>.
-<http://example.org/.well-known/genid#6> <http://example.org/ns#anonymousProperty> "I am a blank node."^^<http://www.w3.org/2001/XMLSchema#string>.
+_:6 <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://example.org/ns#AnonymousType>.
+_:6 <http://example.org/ns#anonymousProperty> "I am a blank node."^^<http://www.w3.org/2001/XMLSchema#string>.
 <http://example.org/ns#MultilingualResource> <http://purl.org/dc/elements/1.1/title> "Título en Español"@es.
 <http://example.org/ns#MultilingualResource> <http://purl.org/dc/elements/1.1/title> "Title in English"@en.
 <http://example.org/ns#MultilingualResource> <http://purl.org/dc/elements/1.1/title> "Titre en Français"@fr.
 <http://example.org/ns#DataTypedLiterals> <http://example.org/ns#integerValue> "123"^^<http://www.w3.org/2001/XMLSchema#integer>.
 <http://example.org/ns#DataTypedLiterals> <http://example.org/ns#decimalValue> "123.45"^^<http://www.w3.org/2001/XMLSchema#decimal>.
 <http://example.org/ns#DataTypedLiterals> <http://example.org/ns#booleanValue> "true"^^<http://www.w3.org/2001/XMLSchema#boolean>.
 <http://example.org/ns#DataTypedLiterals> <http://example.org/ns#customValue> "custom-datatype"^^<http://example.org/ns#CustomDatatype>.
-<http://example.org/.well-known/genid#7> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://www.w3.org/1999/02/22-rdf-syntax-ns#Statement>.
-<http://example.org/.well-known/genid#7> <http://www.w3.org/1999/02/22-rdf-syntax-ns#subject> <http://example.org/ns#ComplexResource>.
-<http://example.org/.well-known/genid#7> <http://www.w3.org/1999/02/22-rdf-syntax-ns#predicate> <http://purl.org/dc/elements/1.1/creator>.
-<http://example.org/.well-known/genid#7> <http://www.w3.org/1999/02/22-rdf-syntax-ns#object> "Author Name"@en.
-<http://example.org/.well-known/genid#8> <http://www.w3.org/1999/02/22-rdf-syntax-ns#first> "First Item"^^<http://www.w3.org/2001/XMLSchema#string>.
-<http://example.org/.well-known/genid#10> <http://www.w3.org/1999/02/22-rdf-syntax-ns#first> "Nested Item 1"^^<http://www.w3.org/2001/XMLSchema#string>.
-<http://example.org/.well-known/genid#11> <http://www.w3.org/1999/02/22-rdf-syntax-ns#first> "Nested Item 2"^^<http://www.w3.org/2001/XMLSchema#string>.
-<http://example.org/.well-known/genid#11> <http://www.w3.org/1999/02/22-rdf-syntax-ns#rest> <http://www.w3.org/1999/02/22-rdf-syntax-ns#nil>.
-<http://example.org/.well-known/genid#10> <http://www.w3.org/1999/02/22-rdf-syntax-ns#rest> <http://example.org/.well-known/genid#11>.
-<http://example.org/.well-known/genid#9> <http://www.w3.org/1999/02/22-rdf-syntax-ns#first> <http://example.org/.well-known/genid#10>.
-<http://example.org/.well-known/genid#12> <http://www.w3.org/1999/02/22-rdf-syntax-ns#first> "Second Item"^^<http://www.w3.org/2001/XMLSchema#string>.
-<http://example.org/.well-known/genid#12> <http://www.w3.org/1999/02/22-rdf-syntax-ns#rest> <http://www.w3.org/1999/02/22-rdf-syntax-ns#nil>.
-<http://example.org/.well-known/genid#9> <http://www.w3.org/1999/02/22-rdf-syntax-ns#rest> <http://example.org/.well-known/genid#12>.
-<http://example.org/.well-known/genid#8> <http://www.w3.org/1999/02/22-rdf-syntax-ns#rest> <http://example.org/.well-known/genid#9>.
-<http://example.org/ns#ComplexList> <http://example.org/ns#hasList> <http://example.org/.well-known/genid#8>.
+_:7 <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://www.w3.org/1999/02/22-rdf-syntax-ns#Statement>.
+_:7 <http://www.w3.org/1999/02/22-rdf-syntax-ns#subject> <http://example.org/ns#ComplexResource>.
+_:7 <http://www.w3.org/1999/02/22-rdf-syntax-ns#predicate> <http://purl.org/dc/elements/1.1/creator>.
+_:7 <http://www.w3.org/1999/02/22-rdf-syntax-ns#object> "Author Name"@en.
+_:8 <http://www.w3.org/1999/02/22-rdf-syntax-ns#first> "First Item"^^<http://www.w3.org/2001/XMLSchema#string>.
+_:10 <http://www.w3.org/1999/02/22-rdf-syntax-ns#first> "Nested Item 1"^^<http://www.w3.org/2001/XMLSchema#string>.
+_:11 <http://www.w3.org/1999/02/22-rdf-syntax-ns#first> "Nested Item 2"^^<http://www.w3.org/2001/XMLSchema#string>.
+_:11 <http://www.w3.org/1999/02/22-rdf-syntax-ns#rest> <http://www.w3.org/1999/02/22-rdf-syntax-ns#nil>.
+_:10 <http://www.w3.org/1999/02/22-rdf-syntax-ns#rest> _:11.
+_:9 <http://www.w3.org/1999/02/22-rdf-syntax-ns#first> _:10.
+_:12 <http://www.w3.org/1999/02/22-rdf-syntax-ns#first> "Second Item"^^<http://www.w3.org/2001/XMLSchema#string>.
+_:12 <http://www.w3.org/1999/02/22-rdf-syntax-ns#rest> <http://www.w3.org/1999/02/22-rdf-syntax-ns#nil>.
+_:9 <http://www.w3.org/1999/02/22-rdf-syntax-ns#rest> _:12.
+_:8 <http://www.w3.org/1999/02/22-rdf-syntax-ns#rest> _:9.
+<http://example.org/ns#ComplexList> <http://example.org/ns#hasList> _:8.