Pull request #275: Tolerate wrongly reported content types for remote shape files if file name can be used

Merge in ITB/shacl-validator from development to master

* commit '28c6aa12dac195f269e1c23da3695e520d168926':
  Tolerate wrongly reported content types for remote shape files if file name can be used

Author: costas80

shaclvalidator-common/src/main/java/eu/europa/ec/itb/shacl/util/ShaclValidatorUtils.java

@@ -19,15 +19,19 @@
 import eu.europa.ec.itb.shacl.validation.FileManager;
 import eu.europa.ec.itb.shacl.validation.ReportSpecs;
 import eu.europa.ec.itb.shacl.validation.SHACLReportHandler;
+import eu.europa.ec.itb.validation.commons.FileInfo;
 import eu.europa.ec.itb.validation.commons.LocalisationHelper;
 import eu.europa.ec.itb.validation.commons.ReportPair;
+import org.apache.jena.atlas.web.ContentType;
 import org.apache.jena.query.Query;
 import org.apache.jena.query.QueryExecution;
 import org.apache.jena.query.QueryExecutionFactory;
 import org.apache.jena.query.QueryFactory;
 import org.apache.jena.rdf.model.Model;
 import org.apache.jena.rdf.model.RDFNode;
 import org.apache.jena.rdf.model.Statement;
+import org.apache.jena.riot.Lang;
+import org.apache.jena.riot.RDFLanguages;
 import org.slf4j.Logger;
 import org.slf4j.LoggerFactory;
 import org.springframework.util.MimeType;
@@ -115,6 +119,31 @@ public static String handleEquivalentContentSyntaxes(String contentSyntax) {
         return contentSyntax;
     }
 
+    /**
+     * Determine the RDF language for the provided file.
+     *
+     * @param fileInfo The file information.
+     * @return The language.
+     */
+    public static Lang determineRdfLanguage(FileInfo fileInfo) {
+        // First check content type.
+        String contentType = handleEquivalentContentSyntaxes(fileInfo.getType());
+        if (isRdfContentSyntax(contentType)) {
+            // This is an RDF content type.
+            return RDFLanguages.contentTypeToLang(contentType);
+        } else {
+            // Not an RDF content type.
+            ContentType guessedContentType = RDFLanguages.guessContentType(fileInfo.getFile().getName());
+            if (guessedContentType != null) {
+                // We guessed the content type from the file name.
+                return RDFLanguages.contentTypeToLang(guessedContentType);
+            } else {
+                // Use the original content type for a best-effort guess.
+                return RDFLanguages.contentTypeToLang(fileInfo.getType());
+            }
+        }
+    }
+
     /**
      * Make sure badly reported content types are correctly handled.
      *

shaclvalidator-common/src/main/java/eu/europa/ec/itb/shacl/validation/SHACLValidator.java

@@ -66,8 +66,7 @@
 import java.io.*;
 import java.util.*;
 
-import static eu.europa.ec.itb.shacl.util.ShaclValidatorUtils.getStatementSafe;
-import static eu.europa.ec.itb.shacl.util.ShaclValidatorUtils.handleEquivalentContentSyntaxes;
+import static eu.europa.ec.itb.shacl.util.ShaclValidatorUtils.*;
 import static eu.europa.ec.itb.shacl.validation.SHACLResources.VALIDATION_REPORT;
 import static org.apache.jena.riot.lang.LangJSONLD11.JSONLD_OPTIONS;
 
@@ -414,7 +413,7 @@ private Model getShapesModel(List<FileInfo> shaclFiles) {
             if (specs.isLogProgress()) {
                 LOG.info("Validating against [{}]", shaclFile.getFile().getName());
             }
-            Lang rdfLanguage = processRdfLanguage(RDFLanguages.contentTypeToLang(handleEquivalentContentSyntaxes(shaclFile.getType())));
+            Lang rdfLanguage = processRdfLanguage(determineRdfLanguage(shaclFile));
             if (rdfLanguage == null) {
                 throw new ValidatorException("validator.label.exception.unableToDetermineShaclContentType");
             }