[Feature #15] Fix normalize propertyUris and column names

kbss-cvut · Sep 17, 2021 · 5f42e64 · 5f42e64
1 parent e01d2f0
commit 5f42e64
Show file tree

Hide file tree

Showing 2 changed files with 18 additions and 4 deletions.
diff --git a/s-pipes-core/src/main/java/cz/cvut/spipes/constants/CSVW.java b/s-pipes-core/src/main/java/cz/cvut/spipes/constants/CSVW.java
@@ -25,6 +25,7 @@ protected static Property property(String local )
     public static final Property aboutUrl = property("aboutUrl");
     public static final Property propertyUrl = property("propertyUrl");
     public static final Property name = property("name");
+    public static final Property title = property("title");
     public static final Property valueUrl = property("valueUrl");
     public static final Property tableSchema = property("tableSchema");
     public static final Property column = property("column");

diff --git a/s-pipes-core/src/main/java/cz/cvut/spipes/modules/TabularModule.java b/s-pipes-core/src/main/java/cz/cvut/spipes/modules/TabularModule.java
@@ -103,8 +103,11 @@ ExecutionContext executeSelf() {
 
             String[] header = listReader.getHeader(true); // skip the header (can't be used with CsvListReader)
 
+
             for (String columnName : header) {
                 Resource columnResource = ResourceFactory.createResource();
+                // TODO make sure normalized names does not clash
+                String columnNameNormalized = normalize(columnName);
 
                 outputModel.add(
                         T_Schema,
@@ -114,12 +117,17 @@ ExecutionContext executeSelf() {
                 outputModel.add(
                         columnResource,
                         CSVW.name,
-                        ResourceFactory.createStringLiteral(columnName)
+                        ResourceFactory.createStringLiteral(columnNameNormalized)
+                );
+                outputModel.add(
+                    columnResource,
+                    CSVW.title,
+                    ResourceFactory.createStringLiteral(columnName)
                 );
                 outputModel.add(
                         columnResource,
                         CSVW.aboutUrl,
-                        outputModel.createTypedLiteral(sourceResource.getUri() + "/columns/" + columnName + "-{_row}", CSVW.uriTemplate)
+                        outputModel.createTypedLiteral(sourceResource.getUri() + "/columns/" + columnNameNormalized + "-{_row}", CSVW.uriTemplate)
                 );
             }
 
@@ -195,16 +203,17 @@ ExecutionContext executeSelf() {
                     // 4.6.8.3
                     String propertyUrl = null; //TODO get from user
                     String columnName = header[i];
+                    String normalizedColumnName = normalize(columnName);
 
                     Property P;
                     if (propertyUrl != null && !propertyUrl.isEmpty()) {
                         P = ResourceFactory.createProperty(propertyUrl);
                     } else if (dataPrefix != null && !dataPrefix.isEmpty()) {
                         P = ResourceFactory.createProperty(
-                                dataPrefix + URLEncoder.encode(columnName, "UTF-8"));
+                                dataPrefix + URLEncoder.encode(normalizedColumnName, "UTF-8"));
                     } else {
                         P = ResourceFactory.createProperty(
-                                sourceResource.getUri() + "#" + URLEncoder.encode(columnName, "UTF-8")); //TODO should be URL (according to specification) not URI
+                                sourceResource.getUri() + "#" + URLEncoder.encode(normalizedColumnName, "UTF-8")); //TODO should be URL (according to specification) not URI
                     }
 
                     String valueUrl = null; //TODO get from user
@@ -343,6 +352,10 @@ private void onTable(String tableUri) {
         }
     }
 
+    private String normalize(String label) {
+        return label.replaceAll("[^\\w]", "_");
+    }
+
     private Reader getReader() {
         return new StringReader(new String(sourceResource.getContent()));
     }