lists in result messages

.changeset/eighty-snakes-hide.md

@@ -0,0 +1,5 @@
+---
+"rdf-validate-shacl": patch
+---
+
+Improved result messages with lists

src/dataset-utils.js

@@ -7,21 +7,52 @@ import NodeSet from './node-set.js'
  *
  * @param {DatasetCore} dataset
  * @param {Term} startNode
- * @returns Array of quads
+ * @yields {Quad}
  */
-export function extractStructure(dataset, startNode, visited = new TermSet()) {
+export function * extractStructure(dataset, startNode, visited = new TermSet()) {
   if (startNode.termType !== 'BlankNode' || visited.has(startNode)) {
-    return []
+    return
   }
 
   visited.add(startNode)
-  const quads = [...dataset.match(startNode, null, null)]
+  for (const quad of dataset.match(startNode, null, null)) {
+    yield quad
+    yield * extractStructure(dataset, quad.object, visited)
+  }
+}
+
+/**
+ * Extracts all the quads forming the structure under a blank shape node. Stops at
+ * non-blank nodes. Replaces sh:in with a comment if the list is too long.
+ *
+ * @param {Shape} shape
+ * @param {DatasetCore} dataset
+ * @param {Term} startNode
+ * @yields {Quad}
+ */
+export function * extractSourceShapeStructure(shape, dataset, startNode, visited = new TermSet()) {
+  if (startNode.termType !== 'BlankNode' || visited.has(startNode)) {
+    return
+  }
 
-  const children = quads.map((quad) => {
-    return extractStructure(dataset, quad.object, visited)
-  })
+  const { factory } = shape.context
+  const { sh, rdfs } = shape.context.ns
 
-  return quads.concat(...children)
+  const inListSize = term => {
+    const inConstraint = shape.constraints.find(x => x.paramValue.equals(term))
+    return inConstraint?.nodeSet.size
+  }
+
+  visited.add(startNode)
+  for (const quad of dataset.match(startNode, null, null)) {
+    if (quad.predicate.equals(sh.in) && inListSize(quad.object) > 3) {
+      const msg = `sh:in has ${inListSize(quad.object)} elements and has been removed from the report for brevity. Please refer the original shape`
+      yield factory.quad(quad.subject, rdfs.comment, factory.literal(msg))
+    } else {
+      yield quad
+      yield * extractSourceShapeStructure(shape, dataset, quad.object, visited)
+    }
+  }
 }
 
 /**

src/shapes-graph.js

@@ -20,7 +20,7 @@ import NodeSet from './node-set.js'
 import ValidationFunction from './validation-function.js'
 import validatorsRegistry from './validators-registry.js'
 import { extractPropertyPath, getPathObjects } from './property-path.js'
-import { getInstancesOf, isInstanceOf } from './dataset-utils.js'
+import { getInstancesOf, isInstanceOf, rdfListToArray } from './dataset-utils.js'
 
 class ShapesGraph {
   constructor(context) {
@@ -124,6 +124,14 @@ class Constraint {
   get componentMessages() {
     return this.component.getMessages(this.shape)
   }
+
+  get nodeSet() {
+    const { sh } = this.shape.context.ns
+    if (!this.inNodeSet) {
+      this.inNodeSet = new NodeSet(rdfListToArray(this.shapeNodePointer.out(sh.in)))
+    }
+    return this.inNodeSet
+  }
 }
 
 class ConstraintComponent {

src/validation-engine.js

@@ -1,7 +1,7 @@
 import clownface from 'clownface'
 import debug from 'debug'
 import ValidationReport from './validation-report.js'
-import { extractStructure } from './dataset-utils.js'
+import { extractStructure, extractSourceShapeStructure } from './dataset-utils.js'
 
 const error = debug('validation-enging::error')
 
@@ -17,10 +17,10 @@ class ValidationEngine {
     this.nestedResults = {}
   }
 
-  clone () {
+  clone() {
     return new ValidationEngine(this.context, { maxErrors: this.maxErrors })
   }
-  
+
   initReport() {
     const { rdf, sh } = this.context.ns
 
@@ -241,7 +241,7 @@ class ValidationEngine {
       .addOut(sh.sourceShape, sourceShape)
       .addOut(sh.focusNode, focusNode)
 
-    this.copyNestedStructure(sourceShape, result)
+    this.copySourceShapeStructure(constraint.shape, result)
     this.copyNestedStructure(focusNode, result)
 
     const children = this.nestedResults[this.recordErrorsLevel + 1]
@@ -266,6 +266,13 @@ class ValidationEngine {
     }
   }
 
+  copySourceShapeStructure(shape, result) {
+    const structureQuads = extractSourceShapeStructure(shape, this.context.$shapes.dataset, shape.shapeNode)
+    for (const quad of structureQuads) {
+      result.dataset.add(quad)
+    }
+  }
+
   /**
    * Creates a result message from the validation result and the message pattern in the constraint
    */
@@ -320,7 +327,16 @@ function localName(uri) {
   return uri.substring(index + 1)
 }
 
-function nodeLabel(node) {
+function * take(n, iterable) {
+  let i = 0
+  for (const item of iterable) {
+    if (i++ === n) break
+    yield item
+  }
+}
+
+function nodeLabel(constraint, param) {
+  const node = constraint.getParameterValue(param)
   if (!node) {
     return 'NULL'
   }
@@ -331,6 +347,16 @@ function nodeLabel(node) {
   }
 
   if (node.termType === 'BlankNode') {
+    if (constraint.nodeSet) {
+      const limit = 3
+      if (constraint.nodeSet.size > limit) {
+        const prefix = Array.from(take(limit, constraint.nodeSet)).map(x => x.value)
+        return prefix.join(', ') + ` ... (and ${constraint.nodeSet.size - limit} more)`
+      } else {
+        return Array.from(constraint.nodeSet).map(x => x.value).join(', ')
+      }
+    }
+
     return 'Blank node ' + node.value
   }
 
@@ -340,7 +366,7 @@ function nodeLabel(node) {
 function withSubstitutions(messageTerm, constraint, factory) {
   const message = constraint.component.parameters.reduce((message, param) => {
     const paramName = localName(param.value)
-    const paramValue = nodeLabel(constraint.getParameterValue(param))
+    const paramValue = nodeLabel(constraint, param)
     return message
       .replace(`{$${paramName}}`, paramValue)
       .replace(`{?${paramName}}`, paramValue)

src/validators-registry.js

@@ -54,7 +54,7 @@ export default {
   [ns.sh.InConstraintComponent.value]: {
     validator: {
       func: validators.validateIn,
-      message: 'Value is not in {$in}',
+      message: 'Value is not one of the allowed values: {$in}',
     },
   },
   [ns.sh.LanguageInConstraintComponent.value]: {

src/validators.js

@@ -131,14 +131,7 @@ function validateHasValueProperty(context, focusNode, valueNode, constraint) {
 }
 
 function validateIn(context, focusNode, valueNode, constraint) {
-  const { sh } = context.ns
-  if (!constraint.nodeSet) {
-    const inNode = constraint.getParameterValue(sh.in)
-    constraint.nodeSet = new NodeSet(rdfListToArray(context.$shapes.node(inNode)))
-  }
-  const { nodeSet } = constraint
-
-  return nodeSet.has(valueNode)
+  return constraint.nodeSet.has(valueNode)
 }
 
 function validateLanguageIn(context, focusNode, valueNode, constraint) {

test/data/validation-message/message-with-list.ttl

@@ -0,0 +1,11 @@
+@prefix sh: <http://www.w3.org/ns/shacl#> .
+@prefix xsd: <http://www.w3.org/2001/XMLSchema#> .
+
+<Alice> a <Person> ; <category> "f" .
+
+<ShapeWithList> a sh:NodeShape ;
+  sh:targetClass <Person> ;
+  sh:property [
+    sh:path <category> ;
+    sh:in ("a" "b" "c" "d" "e") ;
+  ] .

test/data/validation-sourceShape/long-list.ttl

@@ -0,0 +1,12 @@
+@prefix sh: <http://www.w3.org/ns/shacl#> .
+@prefix ex: <http://example.com/> .
+
+ex:p1 a ex:Person ; ex:category 3 .
+ex:p2 a ex:Person ; ex:category 30, 40 .
+
+ex:shape sh:targetClass ex:Person ;
+  sh:property [
+    sh:path ex:category ;
+    sh:minCount 2 ;
+    sh:in (1 2 3 4 5) ;
+  ] .

test/data/validation-sourceShape/mandatory-list.ttl

@@ -0,0 +1,11 @@
+@prefix sh: <http://www.w3.org/ns/shacl#> .
+@prefix ex: <http://example.com/> .
+
+ex:p1 a ex:Person .
+
+ex:shape sh:targetClass ex:Person ;
+  sh:property [
+    sh:path ex:category ;
+    sh:minCount 1 ;
+    sh:in (1 2 3) ;
+  ] .

test/validation_message_test.js

@@ -89,4 +89,17 @@ describe('validation messages', () => {
       { value: 'Mon message de validation', language: 'fr' },
     ]))
   })
+
+  it('Lists first items in message', async () => {
+    const dataPath = path.join(rootPath, 'message-with-list.ttl')
+    const data = await loadDataset(dataPath)
+    const shapes = data
+
+    const validator = new SHACLValidator(shapes)
+    const report = validator.validate(data)
+
+    assert.strictEqual(report.results.length, 1)
+    assert.strictEqual(report.results[0].message.length, 1)
+    assert.strictEqual(report.results[0].message[0].value, 'Value is not one of the allowed values: a, b, c ... (and 2 more)')
+  })
 })

test/validation_sourceShape_test.js

@@ -0,0 +1,52 @@
+/* eslint-env mocha */
+import path from 'path'
+import assert from 'assert'
+import * as url from 'url'
+import clownface from 'clownface'
+import SHACLValidator from '../index.js'
+import ns from '../src/namespaces.js'
+import { rdfListToArray } from '../src/dataset-utils.js'
+import { loadDataset } from './utils.js'
+
+const { rdfs, sh } = ns
+
+const __dirname = url.fileURLToPath(new URL('.', import.meta.url))
+const rootPath = path.join(__dirname, '/data/validation-sourceShape')
+
+describe('validation source shapes', () => {
+  it('Includes source shape without long list', async () => {
+    const dataPath = path.join(rootPath, 'long-list.ttl')
+    const data = await loadDataset(dataPath)
+    const shapes = data
+
+    const validator = new SHACLValidator(shapes)
+    const report = validator.validate(data)
+
+    // three results for the same source shape
+    assert.strictEqual(report.results.length, 3)
+    const sourceShapes = new Set(report.results.map(result => result.sourceShape))
+    assert.strictEqual(sourceShapes.size, 1)
+    const [shape] = sourceShapes
+    // the shape has a comment instead of sh:in
+    assert.deepEqual([], [...report.dataset.match(shape, sh.in, null)])
+    const [comment] = report.dataset.match(shape, rdfs.comment, null)
+    assert.strictEqual(comment.object.value, 'sh:in has 5 elements and has been removed from the report for brevity. Please refer the original shape')
+  })
+  it('Includes source shape with list', async () => {
+    const dataPath = path.join(rootPath, 'mandatory-list.ttl')
+    const data = await loadDataset(dataPath)
+    const shapes = data
+
+    const validator = new SHACLValidator(shapes)
+    const report = validator.validate(data)
+
+    assert.strictEqual(report.results.length, 1)
+    const sourceShapes = new Set(report.results.map(result => result.sourceShape))
+    assert.strictEqual(sourceShapes.size, 1)
+    const [shape] = sourceShapes
+    // the shape has the lst of allowed values even though the violated constraint is sh:minCount and not sh:in
+    const list = clownface({ dataset: report.dataset, term: shape }).out(sh.in)
+    const actual = rdfListToArray(list).map(term => term.value)
+    assert.deepEqual(actual, ['1', '2', '3'])
+  })
+})