fix: blank node conflicts

.changeset/poor-cherries-reply.md

@@ -0,0 +1,5 @@
+---
+"@hydrofoil/sparql-processor": patch
+---
+
+Added overridable method `processSubquery`

.changeset/three-donkeys-join.md

@@ -0,0 +1,5 @@
+---
+"@hydrofoil/shape-to-query": patch
+---
+
+When SPARQL constraints generated blank nodes, the exact same blank nodes could be used in the query multiple times, causing invalid SPARQL queries. This change ensures that blank nodes are unique within the query, preventing such issues.

packages/processor/index.ts

@@ -194,13 +194,15 @@ export default abstract class ProcessorImpl<F extends DataFactory = DataFactory>
       .with({ type: 'minus' }, minus => this.processMinus(minus))
       .with({ type: 'filter' }, filter => this.processFilter(filter))
       .with({ type: 'bind' }, bind => this.processBind(bind))
-      .with({ type: 'query', queryType: 'SELECT' }, query => {
-        // clones this instance to provide a separate context for the subquery
-        return this.clone().processQuery(query)
-      })
+      .with({ type: 'query', queryType: 'SELECT' }, query => this.processSubquery(query))
       .otherwise(p => p)
   }
 
+  processSubquery(query: sparqljs.SelectQuery): sparqljs.SelectQuery {
+    // clones this instance to provide a separate context for the subquery
+    return this.clone().processQuery(query)
+  }
+
   processBind(bind: sparqljs.BindPattern): sparqljs.Pattern {
     return {
       ...bind,

packages/shape-to-query/lib/optimizer/BlankNodeScopeFixer.ts

@@ -0,0 +1,55 @@
+import Processor from '@hydrofoil/sparql-processor'
+import type { BlankNode, DataFactory } from '@rdfjs/types'
+import type { GroupPattern, Pattern, UnionPattern } from 'sparqljs'
+import type { Environment } from '@rdfjs/environment/Environment.js'
+import type { TermMapFactory } from '@rdfjs/term-map/Factory.js'
+
+/**
+ * Ensures that blank nodes are renamed when necessary
+ * to avoid using the same labels in different scopes
+ */
+export class BlankNodeScopeFixer extends Processor<Environment<TermMapFactory | DataFactory>> {
+  constructor(
+    factory: Environment<TermMapFactory | DataFactory>,
+    private scopeCounter = 0,
+    private blankNodes = factory.termMap()) {
+    super(factory)
+  }
+
+  clone() {
+    return new BlankNodeScopeFixer(this.factory, this.scopeCounter, this.blankNodes)
+  }
+
+  processGroup(group: GroupPattern): Pattern {
+    this.incrementScope()
+    return super.processGroup(group)
+  }
+
+  processUnion(union: UnionPattern): Pattern {
+    const patterns = union.patterns.map(pattern => {
+      this.incrementScope()
+      return this.processPattern(pattern)
+    })
+
+    return {
+      ...union,
+      patterns,
+    }
+  }
+
+  processBlankNode(blankNode: BlankNode): BlankNode {
+    if (!this.blankNodes.has(blankNode)) {
+      this.blankNodes.set(blankNode, this.scopeCounter)
+    }
+
+    if (this.blankNodes.get(blankNode) === this.scopeCounter) {
+      return blankNode
+    }
+
+    return this.factory.blankNode(`s${this.scopeCounter}_${blankNode.value}`)
+  }
+
+  private incrementScope() {
+    this.scopeCounter++
+  }
+}

packages/shape-to-query/lib/shapeToQuery.ts

@@ -8,12 +8,14 @@ import type { Processor } from '@hydrofoil/sparql-processor'
 import type { Options } from './shapeToPatterns.js'
 import { shapeToPatterns } from './shapeToPatterns.js'
 import { UnionRepeatedPatternsRemover } from './optimizer/UnionRepeatedPatternsRemover.js'
+import { BlankNodeScopeFixer } from './optimizer/BlankNodeScopeFixer.js'
 
 const generator = new sparqljs.Generator()
 
 const defaultOptimizers = (): Processor[] => [
   new DuplicatePatternRemover(rdf),
   new UnionRepeatedPatternsRemover(rdf),
+  new BlankNodeScopeFixer(rdf),
   new PrefixExtractor(rdf),
 ]
 

packages/shape-to-query/package.json

@@ -54,6 +54,7 @@
     "chai": "^5.1.2",
     "chai-string": "^1.5.0",
     "debug": "^4.4.0",
+    "docker-compose": "^1.1.0",
     "glob": "^11.0.0",
     "mocha-chai-jest-snapshot": "^1.1.6",
     "n3": "^1.16.3",

packages/shape-to-query/test/example/full-text-search.ttl

@@ -0,0 +1,68 @@
+@prefix sparql: <http://datashapes.org/sparql#> .
+@prefix xsd: <http://www.w3.org/2001/XMLSchema#> .
+@prefix hydra: <http://www.w3.org/ns/hydra/core#> .
+@prefix sh: <http://www.w3.org/ns/shacl#> .
+@prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .
+@prefix schema: <http://schema.org/> .
+@prefix meta: <https://cube.link/meta/> .
+@prefix s2q: <https://hypermedia.app/shape-to-query#> .
+
+<>
+  a sh:NodeShape ;
+  sh:property
+    [
+      sh:path schema:hasDefinedTerm ;
+      sh:node
+        [
+          sh:property
+            [
+              sh:path rdf:type ;
+            ] ;
+          sh:property
+            [
+              sh:path schema:name ;
+              hydra:freetextQuery "foo"
+            ]
+        ] ;
+      sh:values
+        [
+          sh:distinct
+            [
+              sh:limit 10 ;
+              sh:nodes
+                [
+                  sh:offset 0 ;
+                  sh:nodes
+                    [
+                      sh:orderBy
+                        [
+                          sh:path schema:name ;
+                        ] ;
+                      sh:nodes
+                        [
+                          sh:orderBy
+                            [
+                              sh:path [ sh:inversePath schema:inDefinedTermSet ] ;
+                            ] ;
+                          sh:nodes
+                            [
+                              sh:nodes
+                                [
+                                  sh:path [ sh:inversePath schema:inDefinedTermSet ] ;
+                                ] ;
+                              sh:filterShape
+                                [
+                                  sh:property
+                                    [
+                                      sh:path rdf:type ;
+                                      sh:hasValue schema:DefinedTerm ;
+                                    ] ;
+                                ] ;
+                            ] ;
+                        ] ;
+                    ]
+                ]
+            ]
+        ] ;
+    ] ;
+.

packages/shape-to-query/test/lib/optimizer/BlankNodeScopeFixer/fulltext-lists.rq

@@ -0,0 +1,73 @@
+PREFIX schema: <http://schema.org/>
+PREFIX rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>
+CONSTRUCT {
+  ?resource1 schema:hasDefinedTerm ?resource2.
+  ?resource2 rdf:type ?resource4.
+  ?resource2 schema:name ?resource7.
+}
+WHERE {
+  {
+    SELECT ?resource1 ?resource2 WHERE {
+      {
+        SELECT DISTINCT ?resource1 ?resource2 WHERE {
+          ?resource1 ^schema:inDefinedTermSet ?resource2.
+          ?resource2 rdf:type ?resource4.
+          VALUES ?resource4 {
+            schema:DefinedTerm
+          }
+          OPTIONAL { ?resource2 ^schema:inDefinedTermSet ?resource5. }
+          OPTIONAL { ?resource2 schema:name ?resource6. }
+        }
+        ORDER BY (?resource5) (?resource6)
+        LIMIT 10
+      }
+      ?resource1 schema:hasDefinedTerm ?resource10.
+      {
+        ?resource10 schema:name ?resource12.
+        {
+          ?resource10 <http://jena.apache.org/text#query> _:b261;
+            schema:name ?resource12.
+          _:b261 rdf:first schema:name;
+            rdf:rest _:b262.
+          _:b262 rdf:first "foo*";
+            rdf:rest rdf:nil.
+          FILTER(REGEX(?resource12, "^foo", "i"))
+        }
+      }
+    }
+  }
+  UNION
+  {
+    SELECT ?resource2 ?resource4 ?resource7 WHERE {
+      {
+        SELECT DISTINCT ?resource1 ?resource2 WHERE {
+          ?resource1 ^schema:inDefinedTermSet ?resource2.
+          ?resource2 rdf:type ?resource4.
+          VALUES ?resource4 {
+            schema:DefinedTerm
+          }
+          OPTIONAL { ?resource2 ^schema:inDefinedTermSet ?resource5. }
+          OPTIONAL { ?resource2 schema:name ?resource6. }
+        }
+        ORDER BY (?resource5) (?resource6)
+        LIMIT 10
+      }
+      { ?resource2 rdf:type ?resource4. }
+      UNION
+      { ?resource2 schema:name ?resource7. }
+      ?resource1 schema:hasDefinedTerm ?resource10.
+      {
+        ?resource10 schema:name ?resource12.
+        {
+          ?resource10 <http://jena.apache.org/text#query> _:b261;
+            schema:name ?resource12.
+          _:b261 rdf:first schema:name;
+            rdf:rest _:b262.
+          _:b262 rdf:first "foo*";
+            rdf:rest rdf:nil.
+          FILTER(REGEX(?resource12, "^foo", "i"))
+        }
+      }
+    }
+  }
+}

packages/shape-to-query/test/lib/optimizer/BlankNodeScopeFixer/multiple-levels.rq

@@ -0,0 +1,14 @@
+PREFIX ex: <http://example.org/>
+
+ASK {
+    _:foo a ex:Foo .
+    {
+        _:foo a ex:Bar .
+        {
+            _:foo a ex:Baz .
+            {
+                _:foo a ex:Qux .
+            }
+        }
+    }
+}

packages/shape-to-query/test/lib/optimizer/BlankNodeScopeFixer/sub-query.rq

@@ -0,0 +1,15 @@
+PREFIX ex: <http://example.org/>
+
+ASK {
+    {
+        SELECT * {
+          _:foo a ex:Foo .
+        }
+    }
+    UNION
+    {
+      SELECT * {
+        _:foo a ex:Bar .
+      }
+    }
+}

packages/shape-to-query/test/lib/optimizer/BlankNodeScopeFixer/union.rq

@@ -0,0 +1,15 @@
+PREFIX ex: <http://example.org/>
+
+ASK {
+    {
+        _:foo a ex:Bar .
+    }
+    UNION
+    {
+        _:foo a ex:Foo .
+    }
+    UNION
+    {
+        _:foo a ex:Baz .
+    }
+}

packages/shape-to-query/test/lib/optimizer/BlankNodeScopeFixer/used-twice-in-scope.rq

@@ -0,0 +1,9 @@
+PREFIX ex: <http://example.org/>
+
+ASK {
+    _:foo a ex:Foo .
+    {
+        _:foo a ex:Bar .
+        _:foo a ex:Baz .
+    }
+}

packages/shape-to-query/test/lib/optimizer/BlankNodeScopeFixer/used-twice-top-scope.rq

@@ -0,0 +1,6 @@
+PREFIX ex: <http://example.org/>
+
+ASK {
+  _:foo a ex:Foo .
+  _:foo a ex:Bar .
+}

packages/shape-to-query/test/lib/optimizer/BlankNodeScopeFixer/used-twice.rq

@@ -0,0 +1,8 @@
+PREFIX ex: <http://example.org/>
+
+ASK {
+    _:foo a ex:Foo .
+    {
+        _:foo a ex:Bar .
+    }
+}