[Feature #25] Schema loading from JOPA model.

kbss-cvut · May 17, 2022 · 3fbb0c9 · 3fbb0c9
1 parent 887bbbf
commit 3fbb0c9
Show file tree

Hide file tree

Showing 5 changed files with 83 additions and 24 deletions.
diff --git a/s-pipes-modules/module-tabular/src/main/java/cz/cvut/spipes/modules/TabularModule.java b/s-pipes-modules/module-tabular/src/main/java/cz/cvut/spipes/modules/TabularModule.java
@@ -1,16 +1,23 @@
 package cz.cvut.spipes.modules;
 
+import cz.cvut.kbss.jopa.model.EntityManager;
 import cz.cvut.spipes.constants.CSVW;
 import cz.cvut.spipes.constants.KBSS_MODULE;
 import cz.cvut.spipes.constants.SML;
 import cz.cvut.spipes.engine.ExecutionContext;
 import cz.cvut.spipes.engine.ExecutionContextFactory;
 import cz.cvut.spipes.exception.ResourceNotFoundException;
 import cz.cvut.spipes.exception.ResourceNotUniqueException;
+import cz.cvut.spipes.modules.model.Column;
+import cz.cvut.spipes.modules.model.TableSchema;
+import cz.cvut.spipes.modules.template.InvalidTemplateException;
+import cz.cvut.spipes.modules.template.UriTemplate;
+import cz.cvut.spipes.modules.util.JopaPersistenceUtils;
 import cz.cvut.spipes.registry.StreamResource;
 import cz.cvut.spipes.registry.StreamResourceRegistry;
 import cz.cvut.spipes.util.JenaUtils;
 import org.apache.jena.datatypes.xsd.XSDDatatype;
+import org.apache.jena.iri.IRI;
 import org.apache.jena.rdf.model.*;
 import org.apache.jena.vocabulary.RDF;
 import org.apache.jena.shared.PropertyNotFoundException;
@@ -110,6 +117,7 @@ public class TabularModule extends AbstractModule {
     @Override
     ExecutionContext executeSelf() {
         Model inputModel = executionContext.getDefaultModel();
+        EntityManager em = JopaPersistenceUtils.getEntityManager("model", inputModel);
 
         outputModel = ModelFactory.createDefaultModel();
 
@@ -131,6 +139,15 @@ ExecutionContext executeSelf() {
             Set<String> columnNames = new HashSet<>();
             List<RDFNode> columns = new LinkedList<>();
 
+            TableSchema tableSchema = em.createNativeQuery(
+                    "PREFIX csvw: <http://www.w3.org/ns/csvw#>" +
+                            "SELECT ?t WHERE {" +
+                            "?t rdf:type csvw:TableSchema" +
+                            "}",
+                    TableSchema.class).getSingleResult();
+            List<Column> schemaColumns = tableSchema.getColumns();
+
+            int j = 0;
             for (String columnTitle : header) {
                 Resource columnResource = ResourceFactory.createResource();
                 String columnName = normalize(columnTitle);
@@ -166,22 +183,15 @@ ExecutionContext executeSelf() {
                     ResourceFactory.createStringLiteral(columnTitle)
                 );
 
-                String columnAboutUrl = null; //TODO get from inputModel
-                if (columnAboutUrl != null && !columnAboutUrl.isEmpty()) {
-                    outputModel.add(
-                            columnResource,
-                            CSVW.aboutUrl,
-                            outputModel.createTypedLiteral(columnAboutUrl, CSVW.uriTemplate)
-                    );
-                } else {
-                    outputModel.add(
-                            T_Schema,
-                            CSVW.aboutUrl,
-                            outputModel.createTypedLiteral(sourceResource.getUri() + "#row-{_row}", CSVW.uriTemplate)
-                    );
-                }
+                Column column = schemaColumns.get(j);
+                String columnAboutUrlStr = column.getAboutUrl();
+                outputModel.add(
+                        columnResource,
+                        CSVW.aboutUrl,
+                        outputModel.createTypedLiteral(columnAboutUrlStr, CSVW.uriTemplate)
+                );
 
-                String columnPropertyUrl = null; //TODO get from inputModel
+                String columnPropertyUrl = column.getPropertyUrl();
                 if (columnPropertyUrl != null && !columnPropertyUrl.isEmpty()) {
                     outputModel.add(
                             columnResource,
@@ -201,6 +211,8 @@ ExecutionContext executeSelf() {
                             ResourceFactory.createPlainLiteral(sourceResource.getUri() + "#" + URLEncoder.encode(columnName, "UTF-8"))
                     );
                 }
+
+                j++;
             }
 
             RDFList columnList = outputModel.createList(columns.iterator());
@@ -260,11 +272,12 @@ ExecutionContext executeSelf() {
                     Resource schemaColumnResource = columns.get(i).asResource();
 
                     // 4.6.8.1
-                    String columnAboutUrl = getAboutUrlFromSchema(schemaColumnResource);
-                    Resource S = ResourceFactory.createResource(columnAboutUrl.replace(
-                            "{_row}",
-                            Integer.toString(listReader.getRowNumber())
-                    ));
+                    String columnAboutUrlStr = getAboutUrlFromSchema(schemaColumnResource);
+                    UriTemplate aboutUrlTemplate = new UriTemplate(columnAboutUrlStr);
+                    aboutUrlTemplate.initialize(null, Arrays.asList(header));
+                    IRI columnAboutUrl = aboutUrlTemplate.getUri(row);
+
+                    Resource S = ResourceFactory.createResource(columnAboutUrl.toString());
 
                     // 4.6.8.2
                     if (R != null) {
@@ -278,7 +291,8 @@ ExecutionContext executeSelf() {
                     String columnPropertyUrl = getPropertyUrlFromSchema(schemaColumnResource);
                     Property P = ResourceFactory.createProperty(columnPropertyUrl);
 
-                    String valueUrl = null; //TODO get from inputModel
+                    Column column = schemaColumns.get(i);
+                    String valueUrl = column.getValueUrl();
 
                     if (valueUrl != null && !valueUrl.isEmpty()) {
                         // 4.6.8.4
@@ -309,7 +323,7 @@ ExecutionContext executeSelf() {
                 }
             }
 
-        } catch (IOException e) {
+        } catch (IOException | InvalidTemplateException e) {
             LOG.error("Error while reading file from resource uri {}", sourceResource, e);
         } finally {
             if( listReader != null ) {

diff --git a/s-pipes-modules/module-tabular/src/main/java/cz/cvut/spipes/modules/model/Column.java b/s-pipes-modules/module-tabular/src/main/java/cz/cvut/spipes/modules/model/Column.java
@@ -1,5 +1,6 @@
 package cz.cvut.spipes.modules.model;
 
+import cz.cvut.kbss.jopa.model.annotations.Id;
 import cz.cvut.kbss.jopa.model.annotations.OWLAnnotationProperty;
 import cz.cvut.kbss.jopa.model.annotations.OWLClass;
 import cz.cvut.spipes.constants.CSVW;
@@ -10,6 +11,9 @@
 @OWLClass(iri = CSVW.ColumnUri)
 public class Column {
 
+    @Id(generated = true)
+    private String id;
+
     @OWLAnnotationProperty(iri = CSVW.nameUri)
     private String name;
 

diff --git a/s-pipes-modules/module-tabular/src/main/java/cz/cvut/spipes/modules/model/TableSchema.java b/s-pipes-modules/module-tabular/src/main/java/cz/cvut/spipes/modules/model/TableSchema.java
@@ -17,6 +17,12 @@ public class TableSchema {
     @OWLAnnotationProperty(iri = CSVW.aboutUrlUri)
     private String aboutUrl;
 
+    @OWLAnnotationProperty(iri = CSVW.propertyUrlUri)
+    private String propertyUrl;
+
+    @OWLAnnotationProperty(iri = CSVW.valueUrlUri)
+    private String valueUrl;
+
     @OWLObjectProperty(iri = CSVW.columnsUri)
     private List<Column> columns;
 
@@ -28,6 +34,22 @@ public void setAboutUrl(String aboutUrl) {
         this.aboutUrl = aboutUrl;
     }
 
+    public String getPropertyUrl() {
+        return propertyUrl;
+    }
+
+    public void setPropertyUrl(String propertyUrl) {
+        this.propertyUrl = propertyUrl;
+    }
+
+    public String getValueUrl() {
+        return valueUrl;
+    }
+
+    public void setValueUrl(String valueUrl) {
+        this.valueUrl = valueUrl;
+    }
+
     public List<Column> getColumns() {
         return columns;
     }

diff --git a/...-modules/module-tabular/src/main/java/cz/cvut/spipes/modules/template/StringTemplate.java b/...-modules/module-tabular/src/main/java/cz/cvut/spipes/modules/template/StringTemplate.java
@@ -6,12 +6,18 @@
 import java.util.List;
 
 /**
- * Implementation of the first two levels of RFC6570
+ * Implementation of the first two levels of RFC6570.
+ * <p>
+ * This implementation of the String Template has been heavily inspired by the one that is part of the
+ * project Linkedpipes ETL developed by the Linkedpipes ETL Team
+ * (<a href="https://etl.linkedpipes.com/team/">ETL Team</a>).
+ *
  * @see <a href="https://tools.ietf.org/html/rfc6570#section-2">RFC6570 section 2</a>
+ * @see <a href="https://etl.linkedpipes.com/">Linkedpipes ETL</a>
  */
 public class StringTemplate {
 
-    public static final String TABLE_RESOURCE_REF = "__A83N48X1_TABLE_URI__"; //TODO what is this?
+    public static final String TABLE_RESOURCE_REF = "__A83N48X1_TABLE_URI__";
 
     private interface Token {
 

diff --git a/...pes-modules/module-tabular/src/main/java/cz/cvut/spipes/modules/template/UriTemplate.java b/...pes-modules/module-tabular/src/main/java/cz/cvut/spipes/modules/template/UriTemplate.java
@@ -1,5 +1,8 @@
 package cz.cvut.spipes.modules.template;
 
+import org.apache.jena.iri.IRI;
+import org.apache.jena.iri.IRIFactory;
+
 import java.util.List;
 
 public class UriTemplate {
@@ -13,4 +16,14 @@ public UriTemplate(String templateAsString) {
     public void initialize(String tableUri, List<String> header) throws InvalidTemplateException {
         template.initialize(tableUri, header);
     }
+
+    public IRI getUri(List<String> row) {
+        String val = template.process(row);
+
+        if (val != null) {
+            return IRIFactory.iriImplementation().create(val);
+        }
+
+        return null;
+    }
 }