Update PMD4 validations to reflect current muttnik expectations (#1249)

.gitignore

@@ -0,0 +1,2 @@
+/.cpcache
+/.nrepl-port

deps.edn

@@ -1,15 +1,29 @@
-{:aliases {:pmd4 {:extra-deps {swirrl/validations.pmd4 {:local/root "pmd4"}}}
-           :pmd-qb {:extra-deps {swirrl/validations.pmd-qb {:local/root "pmd-qb"}}}
-           :misc {:extra-deps {swirrl/validations.pmd-qb {:local/root "misc"}}}
 
-           :validate {:extra-deps { swirrl/rdf-validator {:git/url "https://github.com/Swirrl/rdf-validator.git"
-                                                          :sha "c85338c44be9f7f9726c30dca4aa47ef8bd9cfe6"
-                                                          }
-                                   ;; Import log4j2 as the logging backend
-                                   org.apache.logging.log4j/log4j-api {:mvn/version "2.11.1"}
-                                   org.apache.logging.log4j/log4j-core {:mvn/version "2.11.1"}
-                                   org.apache.logging.log4j/log4j-slf4j-impl {:mvn/version "2.11.1"}
+{ :aliases {:pmd4 {:extra-deps {swirrl/validations.pmd4 {:local/root "pmd4"}}}
+            :pmd-qb {:extra-deps {swirrl/validations.pmd-qb {:local/root "pmd-qb"}}}
+            :misc {:extra-deps {swirrl/validations.pmd-qb {:local/root "misc"}}}
+            :dev {:extra-deps { swirrl/rdf-validator {:git/url "https://github.com/Swirrl/rdf-validator.git"
+                                                      :sha "c85338c44be9f7f9726c30dca4aa47ef8bd9cfe6"}
+                               ;; Import log4j2 as the logging backend
+                               org.apache.logging.log4j/log4j-api {:mvn/version "2.11.1"}
+                               org.apache.logging.log4j/log4j-core {:mvn/version "2.11.1"}
+                               org.apache.logging.log4j/log4j-slf4j-impl {:mvn/version "2.11.1"}}
+                  :extra-paths ["." "test"]}
+            :test {:extra-deps {swirrl/rdf-validator {:git/url "https://github.com/Swirrl/rdf-validator.git"
+                                                      :sha "c85338c44be9f7f9726c30dca4aa47ef8bd9cfe6"}
+                                lambdaisland/kaocha {:mvn/version "1.0.887"}
+                               ;; Import log4j2 as the logging backend
+                               org.apache.logging.log4j/log4j-api {:mvn/version "2.11.1"}
+                               org.apache.logging.log4j/log4j-core {:mvn/version "2.11.1"}
+                               org.apache.logging.log4j/log4j-slf4j-impl {:mvn/version "2.11.1"}}
+                   :main-opts ["-m" "kaocha.runner"]
+                   :extra-paths ["." "test"]}
 
-                                   }
-                      :extra-paths ["."] ;; for log4j2.xml
-                      :main-opts ["-m" "rdf-validator.core"]}}}
+            :validate {:extra-deps {swirrl/rdf-validator {:git/url "https://github.com/Swirrl/rdf-validator.git"
+                                                          :sha "c85338c44be9f7f9726c30dca4aa47ef8bd9cfe6"}
+                                    ;; Import log4j2 as the logging backend
+                                    org.apache.logging.log4j/log4j-api {:mvn/version "2.11.1"}
+                                    org.apache.logging.log4j/log4j-core {:mvn/version "2.11.1"}
+                                    org.apache.logging.log4j/log4j-slf4j-impl {:mvn/version "2.11.1"}}
+                       :extra-paths ["."] ;; for log4j2.xml
+                       :main-opts ["-m" "rdf-validator.core"]}}}

pmd-qb/src/swirrl/validations/pmd-qb/16_SELECT_Dataset_With_MeasureDimension_Obs_Has_Exactly_1_Value.sparql

@@ -26,7 +26,7 @@ WHERE {
     UNION
     {
         # Observations with more than one value of the same measure type
-        SELECT ?obs (COUNT(?val) AS ?count ) {
+        SELECT ?obs (COUNT(distinct ?val) AS ?count ) {
             ?obs qb:dataSet/qb:structure ?dsd ;
                  qb:measureType ?measure .
             ?obs ?measure ?val .

pmd-qb/src/swirrl/validations/pmd-qb/17_SELECT_All_Measures_Present_In_Measures_Dimension_Cube.sparql

@@ -12,11 +12,11 @@ PREFIX owl:     <http://www.w3.org/2002/07/owl#>
 # with the same non-measure dimension values for each of the declared measures. 
 
 # Count number of other measures found at each point 
-SELECT ?obs1 ?numMeasures (COUNT(?obs2) AS ?count) 
+SELECT ?obs1 ?numMeasures (COUNT(DISTINCT ?obs2) AS ?count)
 WHERE {
     {
         # Find the DSDs and check how many measures they have
-        SELECT ?dsd (COUNT(?m) AS ?numMeasures) WHERE {
+        SELECT ?dsd (COUNT(DISTINCT ?m) AS ?numMeasures) WHERE {
             ?dsd qb:component/qb:measure ?m.
         } GROUP BY ?dsd
     }

pmd-qb/src/swirrl/validations/pmd-qb/19_SELECT_Collection_Values_Must_Be_From_CodeList.sparql

@@ -7,7 +7,7 @@ PREFIX owl:     <http://www.w3.org/2002/07/owl#>
 
 SELECT ?obs ?dim (?v AS ?valueNotInCodeList) 
 WHERE {
-    ?obs qb:dataSet/qb:structure/qb:component/qb:componentProperty ?dim .
+    ?obs qb:dataSet/qb:structure/qb:component/qb:dimension ?dim .
     ?dim a qb:DimensionProperty ;
         qb:codeList ?list .
     ?list a skos:Collection .
@@ -22,4 +22,4 @@ WHERE {
 ####                 ORIGINAL ASK QUERY FROM SPEC  IC-19           ####
 # 
 # 
-#######################################################################
+#######################################################################

pmd-qb/src/swirrl/validations/pmd-qb/19_SELECT_ConceptScheme_Values_Must_Be_From_CodeList.sparql

@@ -7,7 +7,7 @@ PREFIX owl:     <http://www.w3.org/2002/07/owl#>
 
 SELECT ?obs ?dim (?v AS ?valueNotInCodeList) 
 WHERE {
-    ?obs qb:dataSet/qb:structure/qb:component/qb:componentProperty ?dim .
+    ?obs qb:dataSet/qb:structure/qb:component/qb:dimension ?dim .
     ?dim a qb:DimensionProperty ;
         qb:codeList ?list .
     ?list a skos:ConceptScheme .
@@ -22,4 +22,4 @@ WHERE {
 ####                 ORIGINAL ASK QUERY FROM SPEC  IC-19           ####
 # 
 # 
-#######################################################################
+#######################################################################

pmd4/src/rdf-validator-suite.edn

@@ -21,10 +21,6 @@
 	 :source "swirrl/validations/pmd4/SELECT_ConceptSchemeExactlyOneLabel.sparql"
 	 :type :sparql
 	 :name "ConceptSchemeExactlyOneLabel"}
-        {
-	 :source "swirrl/validations/pmd4/SELECT_ConceptSchemeMustHaveCatalogEntry.sparql"
-	 :type :sparql
-	 :name "ConceptSchemeMustHaveCatalogEntry"}
         {
 	 :source "swirrl/validations/pmd4/SELECT_ConceptSchemesMaxZeroConceptsAsMembers.sparql"
 	 :type :sparql
@@ -41,26 +37,18 @@
 	 :source "swirrl/validations/pmd4/SELECT_DatasetContentsIsCorrectType.sparql"
 	 :type :sparql
 	 :name "DatasetContentsIsCorrectType"}
-        {
-	 :source "swirrl/validations/pmd4/SELECT_DatasetExactlyOneContents.sparql"
-	 :type :sparql
-	 :name "DatasetExactlyOneContents"}
-        {
-	 :source "swirrl/validations/pmd4/SELECT_DatasetExactlyOneGraph.sparql"
-	 :type :sparql
-	 :name "DatasetExactlyOneGraph"}
         {
 	 :source "swirrl/validations/pmd4/SELECT_DatasetExactlyOneIssuedDate.sparql"
 	 :type :sparql
 	 :name "DatasetExactlyOneIssuedDate"}
-        {
-	 :source "swirrl/validations/pmd4/SELECT_DatasetExactlyOneLabel.sparql"
-	 :type :sparql
-	 :name "DatasetExactlyOneLabel"}
         {
 	 :source "swirrl/validations/pmd4/SELECT_DatasetExactlyOneModifiedDate.sparql"
 	 :type :sparql
 	 :name "DatasetExactlyOneModifiedDate"}
+        {
+	 :source "swirrl/validations/pmd4/SELECT_DatasetExactlyOneTitle.sparql"
+	 :type :sparql
+	 :name "DatasetExactlyOneTitle"}
         {
 	 :source "swirrl/validations/pmd4/SELECT_DatasetGraphIsIri.sparql"
 	 :type :sparql
@@ -70,9 +58,9 @@
 	 :type :sparql
 	 :name "DatasetMaxOneContactPoint"}
         {
-	 :source "swirrl/validations/pmd4/SELECT_DatasetMaxOneContributor.sparql"
+	 :source "swirrl/validations/pmd4/SELECT_DatasetMaxOneContents.sparql"
 	 :type :sparql
-	 :name "DatasetMaxOneContributor"}
+	 :name "DatasetMaxOneContents"}
         {
 	 :source "swirrl/validations/pmd4/SELECT_DatasetMaxOneCreator.sparql"
 	 :type :sparql
@@ -89,6 +77,10 @@
 	 :source "swirrl/validations/pmd4/SELECT_DatasetMaxOnePublisher.sparql"
 	 :type :sparql
 	 :name "DatasetMaxOnePublisher"}
+        {
+	 :source "swirrl/validations/pmd4/SELECT_DatasetMinOneGraph.sparql"
+	 :type :sparql
+	 :name "DatasetMinOneGraph"}
         {
 	 :source "swirrl/validations/pmd4/SELECT_DimensionsExactlyOneLabel.sparql"
 	 :type :sparql
@@ -105,10 +97,6 @@
 	 :source "swirrl/validations/pmd4/SELECT_RecordExactlyOneIssuedDate.sparql"
 	 :type :sparql
 	 :name "RecordExactlyOneIssuedDate"}
-        {
-	 :source "swirrl/validations/pmd4/SELECT_RecordExactlyOneLabel.sparql"
-	 :type :sparql
-	 :name "RecordExactlyOneLabel"}
         {
 	 :source "swirrl/validations/pmd4/SELECT_RecordExactlyOneMetadataGraph.sparql"
 	 :type :sparql

pmd4/src/swirrl/validations/pmd4/SELECT_ConceptExactlyOneLabel.sparql

@@ -5,11 +5,11 @@ PREFIX rdfs:    <http://www.w3.org/2000/01/rdf-schema#>
 
 SELECT ?concept ?count {
     {
-        SELECT ?concept (COUNT(?lbl) AS ?count) 
+        SELECT ?concept (COUNT(DISTINCT ?lbl) AS ?count)
         WHERE {
-        { 
+        {
             ?concept a skos:Concept .
-            ?concept rdfs:label ?lbl 
+            ?concept rdfs:label ?lbl
         }
         }
         GROUP BY ?concept

pmd4/src/swirrl/validations/pmd4/SELECT_ConceptMaxOneComment.sparql

@@ -4,12 +4,12 @@ PREFIX ui:      <http://www.w3.org/ns/ui#>
 PREFIX rdfs:    <http://www.w3.org/2000/01/rdf-schema#>
 
 
-SELECT ?concept (COUNT(?comment) AS ?count) 
+SELECT ?concept (COUNT(DISTINCT ?comment) AS ?count)
         WHERE {
-        { 
+        {
             ?concept a skos:Concept .
             ?concept rdfs:comment ?comment
         }
         }
         GROUP BY ?concept
-        HAVING (?count > 1)
+        HAVING (?count > 1)

pmd4/src/swirrl/validations/pmd4/SELECT_ConceptMaxOneNotation.sparql

@@ -2,13 +2,13 @@ PREFIX qb:      <http://purl.org/linked-data/cube#>
 PREFIX skos:    <http://www.w3.org/2004/02/skos/core#>
 PREFIX ui:      <http://www.w3.org/ns/ui#>
 
-SELECT ?concept (COUNT(?n) AS ?count) 
+SELECT ?concept (COUNT(DISTINCT ?n) AS ?count)
         WHERE {
-        { 
+        {
             ?concept a skos:Concept .
             ?concept skos:notation ?n .
         }
         }
         GROUP BY ?concept
         HAVING (?count > 1)
-        LIMIT 10
+        LIMIT 10

pmd4/src/swirrl/validations/pmd4/SELECT_ConceptSchemeExactlyOneLabel.sparql

@@ -5,15 +5,15 @@ PREFIX rdfs:    <http://www.w3.org/2000/01/rdf-schema#>
 
 SELECT ?scheme ?count ?lbl  {
     {
-        SELECT ?scheme ?lbl (COUNT(?lbl) AS ?count) 
+        SELECT ?scheme ?lbl (COUNT(DISTINCT ?lbl) AS ?count)
         WHERE {
-        { 
+        {
             ?scheme a skos:ConceptScheme .
             ?scheme rdfs:label ?lbl, ?lbl2
             FILTER(?lbl != ?lbl2)
         }
         }
-        GROUP BY ?scheme ?lbl 
+        GROUP BY ?scheme ?lbl
         HAVING (?count > 1)
     }
     UNION {

pmd4/src/swirrl/validations/pmd4/SELECT_ConceptsMaxOfOneSortPriority.sparql

@@ -2,9 +2,9 @@ PREFIX qb:      <http://purl.org/linked-data/cube#>
 PREFIX skos:    <http://www.w3.org/2004/02/skos/core#>
 PREFIX ui:      <http://www.w3.org/ns/ui#>
 
-SELECT ?concept  (COUNT(?sp) AS ?count) ?scheme
+SELECT ?concept  (COUNT(DISTINCT ?sp) AS ?count) ?scheme
 WHERE {
-   { 
+   {
     ?concept a skos:Concept ;
              ui:sortPriority ?sp .
     ?concept skos:inScheme ?scheme .
@@ -13,4 +13,4 @@ WHERE {
 GROUP BY ?concept ?scheme
 HAVING (?count > 1)
 ORDER BY ?scheme ?concept
-LIMIT 10
+LIMIT 10

pmd4/src/swirrl/validations/pmd4/SELECT_DatasetContentsIsCorrectType.sparql

@@ -3,29 +3,30 @@ PREFIX skos:    <http://www.w3.org/2004/02/skos/core#>
 PREFIX ui:      <http://www.w3.org/ns/ui#>
 PREFIX dcat:    <http://www.w3.org/ns/dcat#>
 PREFIX foaf:    <http://xmlns.com/foaf/0.1/>
-PREFIX dct:     <http://purl.org/dc/terms/> 
+PREFIX dct:     <http://purl.org/dc/terms/>
 PREFIX pmdcat:  <http://publishmydata.com/pmdcat#>
 
 SELECT ?dataset {
-    ?dataset a pmdcat:Dataset .
+    ?dataset a pmdcat:Dataset ;
+             pmdcat:datasetContents ?contents .
     FILTER(
         NOT EXISTS {
             SELECT ?dataset {
-                ?dataset a pmdcat:Dataset .
-                ?dataset pmdcat:datasetContents ?contents .
+                ?dataset a pmdcat:Dataset ;
+                         pmdcat:datasetContents ?contents .
                 ?contents a ?type .
                 FILTER(?type IN (
-                    pmdcat:Dataset, 
-                    pmdcat:FileDataset, 
-                    pmdcat:URLDataset, 
-                    pmdcat:LinkedDataset, 
-                    pmdcat:DataCube, 
-                    pmdcat:GraphDatasetContents, 
-                    pmdcat:Ontology, 
+                    pmdcat:Dataset,
+                    pmdcat:FileDataset,
+                    pmdcat:URLDataset,
+                    pmdcat:LinkedDataset,
+                    pmdcat:DataCube,
+                    pmdcat:GraphDatasetContents,
+                    pmdcat:Ontology,
                     pmdcat:ConceptScheme
                 ))
             }
         }
     )
 
-}
+}

pmd4/src/swirrl/validations/pmd4/SELECT_DatasetExactlyOneLabel.sparql → pmd4/src/swirrl/validations/pmd4/SELECT_DatasetExactlyOneTitle.sparql

@@ -3,17 +3,17 @@ PREFIX skos:    <http://www.w3.org/2004/02/skos/core#>
 PREFIX ui:      <http://www.w3.org/ns/ui#>
 PREFIX dcat:    <http://www.w3.org/ns/dcat#>
 PREFIX foaf:    <http://xmlns.com/foaf/0.1/>
-PREFIX dct:     <http://purl.org/dc/terms/> 
+PREFIX dct:     <http://purl.org/dc/terms/>
 PREFIX pmdcat:  <http://publishmydata.com/pmdcat#>
 PREFIX rdfs:    <http://www.w3.org/2000/01/rdf-schema#>
 
 SELECT ?dataset ?count {
     {
-        SELECT ?dataset (COUNT(?lbl1) AS ?count) 
+        SELECT ?dataset (COUNT(DISTINCT ?title1) AS ?count)
         WHERE {
             ?dataset a pmdcat:Dataset .
-            ?dataset rdfs:label ?lbl1, ?lbl2 .
-            FILTER(?lbl1 != ?lbl2)
+            ?dataset dct:title ?title1, ?title2 .
+            FILTER(?title1 != ?title2)
         }
         GROUP BY ?dataset
         HAVING (?count > 1)
@@ -22,7 +22,7 @@ SELECT ?dataset ?count {
         SELECT ?dataset ?count {
             ?dataset a pmdcat:Dataset .
             FILTER NOT EXISTS {
-                ?dataset rdfs:label ?o 
+                ?dataset dct:title ?o
             }
             BIND(0 AS ?count)
         }

pmd4/src/swirrl/validations/pmd4/SELECT_DatasetMaxOneContactPoint.sparql

@@ -3,13 +3,13 @@ PREFIX skos:    <http://www.w3.org/2004/02/skos/core#>
 PREFIX ui:      <http://www.w3.org/ns/ui#>
 PREFIX dcat:    <http://www.w3.org/ns/dcat#>
 PREFIX foaf:    <http://xmlns.com/foaf/0.1/>
-PREFIX dct:     <http://purl.org/dc/terms/> 
+PREFIX dct:     <http://purl.org/dc/terms/>
 PREFIX pmdcat:  <http://publishmydata.com/pmdcat#>
 
-SELECT ?dataset (COUNT(?o) AS ?count) 
+SELECT ?dataset (COUNT(DISTINCT ?o) AS ?count)
 WHERE {
     ?dataset a pmdcat:Dataset .
     ?dataset dcat:contactPoint ?o
 }
 GROUP BY ?dataset
-HAVING (?count > 1)
+HAVING (?count > 1)