Skip to main content

shifty_parse/
lib.rs

1//! RDF shapes graph -> formalism IR lowering (Layer 2).
2//!
3//! [`parse_turtle`] reads a SHACL shapes graph and lowers all supported Core +
4//! AF vocabulary into the [`shifty_algebra::Schema`] IR, applying every sugar
5//! rule from `docs/01-gap-analysis.md`. Unsupported custom components, JS, and
6//! richer AF constructs produce [`Diagnostic`]s rather than silent wrong
7//! answers.
8
9pub mod diagnostics;
10pub mod graph;
11pub mod lower;
12pub mod path;
13pub mod sparql_path;
14pub mod vocab;
15
16pub use diagnostics::{DiagLevel, Diagnostic, ParseError};
17pub use graph::{Loaded, RdfFormat};
18pub use sparql_path::parse_property_path;
19
20use shifty_algebra::Schema;
21
22/// The result of lowering a shapes graph.
23pub struct ParseOutput {
24    pub schema: Schema,
25    pub diagnostics: Vec<Diagnostic>,
26}
27
28/// Load a Turtle shapes graph (for inspecting the raw RDF stage).
29pub fn load_turtle(data: &[u8], base: Option<&str>) -> Result<Loaded, ParseError> {
30    Loaded::from_turtle(data, base)
31}
32
33pub fn load_ntriples(data: &[u8]) -> Result<Loaded, ParseError> {
34    Loaded::from_ntriples(data)
35}
36
37/// Load an RDF graph using content type, source extension, lightweight content
38/// sniffing, and finally a cascading try across common RDF formats.
39pub fn load_rdf_auto(
40    data: &[u8],
41    content_type: Option<&str>,
42    source: Option<&str>,
43    base: Option<&str>,
44) -> Result<Loaded, ParseError> {
45    Loaded::from_rdf_auto(data, content_type, source, base)
46}
47
48/// Parse and lower a Turtle shapes graph into the algebra IR.
49pub fn parse_turtle(data: &[u8], base: Option<&str>) -> Result<ParseOutput, ParseError> {
50    let loaded = Loaded::from_turtle(data, base)?;
51    let lowered = lower::lower(&loaded);
52    Ok(ParseOutput {
53        schema: lowered.schema,
54        diagnostics: lowered.diagnostics,
55    })
56}
57
58/// Lower an already-loaded graph into the algebra IR.
59pub fn parse_loaded(loaded: &Loaded) -> ParseOutput {
60    let lowered = lower::lower(loaded);
61    ParseOutput {
62        schema: lowered.schema,
63        diagnostics: lowered.diagnostics,
64    }
65}
66
67#[cfg(test)]
68mod tests {
69    use super::*;
70    use shifty_algebra::render::schema_to_text;
71
72    const SHAPES: &str = r#"
73        @prefix sh: <http://www.w3.org/ns/shacl#> .
74        @prefix ex: <http://ex/> .
75        @prefix xsd: <http://www.w3.org/2001/XMLSchema#> .
76
77        ex:PersonShape a sh:NodeShape ;
78            sh:targetClass ex:Person ;
79            sh:property [
80                sh:path ex:name ;
81                sh:minCount 1 ;
82                sh:maxCount 1 ;
83                sh:datatype xsd:string ;
84            ] ;
85            sh:property [
86                sh:path [ sh:inversePath ex:child ] ;
87                sh:nodeKind sh:IRI ;
88            ] .
89    "#;
90
91    #[test]
92    fn lowers_person_shape() {
93        let out = parse_turtle(SHAPES.as_bytes(), None).unwrap();
94        assert!(out.diagnostics.is_empty(), "diags: {:?}", out.diagnostics);
95
96        let text = schema_to_text(&out.schema);
97        // a class-target statement was produced
98        assert!(text.contains("rdf:type/rdfs:subClassOf*"), "text:\n{text}");
99        // cardinality on ex:name lowered to an interval count
100        assert!(text.contains("[1..1] <http://ex/name>"), "text:\n{text}");
101        // inverse path rendered
102        assert!(text.contains("^<http://ex/child>"), "text:\n{text}");
103        // datatype facet present
104        assert!(text.contains("datatype(xsd:string)"), "text:\n{text}");
105    }
106
107    #[test]
108    fn auto_loads_rdf_xml_shapes() {
109        let rdfxml = br#"<?xml version="1.0"?>
110            <rdf:RDF
111                xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
112                xmlns:sh="http://www.w3.org/ns/shacl#"
113                xmlns:ex="http://ex/">
114              <sh:NodeShape rdf:about="http://ex/S">
115                <sh:targetClass rdf:resource="http://ex/Thing"/>
116              </sh:NodeShape>
117            </rdf:RDF>"#;
118        let loaded = load_rdf_auto(
119            rdfxml,
120            Some("application/rdf+xml"),
121            Some("https://example.test/shapes.rdf"),
122            Some("https://example.test/shapes.rdf"),
123        )
124        .unwrap();
125        let out = parse_loaded(&loaded);
126        let text = schema_to_text(&out.schema);
127        assert!(text.contains("rdf:type/rdfs:subClassOf*"), "text:\n{text}");
128    }
129
130    #[test]
131    fn lowers_triple_rule() {
132        let ttl = r#"
133            @prefix sh: <http://www.w3.org/ns/shacl#> .
134            @prefix ex: <http://ex/> .
135            ex:S a sh:NodeShape ;
136                sh:targetClass ex:Rectangle ;
137                sh:rule [
138                    a sh:TripleRule ;
139                    sh:subject sh:this ;
140                    sh:predicate ex:area ;
141                    sh:object [ sh:path ex:width ] ;
142                    sh:condition ex:S ;
143                    sh:order 1 ;
144                ] .
145        "#;
146        let out = parse_turtle(ttl.as_bytes(), None).unwrap();
147        assert!(out.diagnostics.is_empty(), "diags: {:?}", out.diagnostics);
148        assert_eq!(out.schema.rules.len(), 1);
149        let r = &out.schema.rules[0];
150        assert_eq!(r.order, Some(1));
151        assert_eq!(r.conditions.len(), 1);
152        use shifty_algebra::{NodeExpr, RuleHead};
153        match &r.head {
154            RuleHead::Triple {
155                subject,
156                predicate,
157                object,
158            } => {
159                assert!(matches!(subject, NodeExpr::This));
160                assert!(matches!(predicate, NodeExpr::Constant(_)));
161                assert!(matches!(object, NodeExpr::Path(_)));
162            }
163            other => panic!("expected TripleRule, got {other:?}"),
164        }
165    }
166
167    #[test]
168    fn lowers_sparql_rule_opaque() {
169        let ttl = r#"
170            @prefix sh: <http://www.w3.org/ns/shacl#> .
171            @prefix ex: <http://ex/> .
172            ex:S a sh:NodeShape ;
173                sh:targetNode ex:x ;
174                sh:rule [ a sh:SPARQLRule ; sh:construct "CONSTRUCT { ?this ex:p ?this } WHERE {}" ] .
175        "#;
176        let out = parse_turtle(ttl.as_bytes(), None).unwrap();
177        assert_eq!(out.schema.rules.len(), 1);
178        assert!(matches!(
179            out.schema.rules[0].head,
180            shifty_algebra::RuleHead::Sparql(_)
181        ));
182    }
183
184    #[test]
185    fn lowers_sparql_constraint() {
186        let ttl = r#"
187            @prefix sh: <http://www.w3.org/ns/shacl#> .
188            @prefix ex: <http://ex/> .
189            ex:S a sh:NodeShape ;
190                sh:targetNode ex:x ;
191                sh:sparql [ sh:select "SELECT $this WHERE {}" ] .
192        "#;
193        let out = parse_turtle(ttl.as_bytes(), None).unwrap();
194        assert!(out.diagnostics.is_empty(), "diags: {:?}", out.diagnostics);
195        let root = out.schema.statements[0].shape;
196        let shifty_algebra::Shape::Annotated {
197            severity, shape, ..
198        } = out.schema.arena.get(root)
199        else {
200            panic!("expected severity annotation");
201        };
202        assert_eq!(severity, &shifty_algebra::Severity::Violation);
203        assert!(matches!(
204            out.schema.arena.get(*shape),
205            shifty_algebra::Shape::Sparql(_)
206        ));
207    }
208
209    #[test]
210    fn lowers_sparql_target() {
211        let ttl = r#"
212            @prefix sh: <http://www.w3.org/ns/shacl#> .
213            @prefix ex: <http://ex/> .
214            ex:S a sh:NodeShape ;
215                sh:target [ sh:select "SELECT ?this WHERE { ?this a ex:Person }" ] .
216        "#;
217        let out = parse_turtle(ttl.as_bytes(), None).unwrap();
218        assert!(out.diagnostics.is_empty(), "diags: {:?}", out.diagnostics);
219        assert!(matches!(
220            out.schema.statements[0].selector,
221            shifty_algebra::Selector::Sparql(_)
222        ));
223    }
224}