Refactor: Rework parsing of CityGmlZipEntry

CityDoctorParent/CityDoctorModel/src/main/java/de/hft/stuttgart/citydoctor2/parser/CityGmlParser.java

@@ -26,6 +26,8 @@ import java.util.ArrayList;
 import java.util.List;
 import java.util.regex.Matcher;
 import java.util.regex.Pattern;
+import java.util.zip.ZipEntry;
+import java.util.zip.ZipFile;
 
 import javax.xml.XMLConstants;
 import javax.xml.namespace.QName;
@@ -189,28 +191,61 @@ public class CityGmlParser {
 		}
 	}
 
-	public static CityDoctorModel parseCityGmlStream(InputStream is, ParserConfiguration config)
-			throws CityGmlParseException, InvalidGmlFileException {
-		return parseCityGmlStream(is, config, null, null);
-	}
+	public static CityDoctorModel parseCityGmlZipEntry(ZipEntry entry, ZipFile archive, ParserConfiguration config)
+            throws CityGmlParseException, InvalidGmlFileException, IOException {
+		CityGMLContext context = getContext();
 
-	public static CityDoctorModel parseCityGmlStream(InputStream is, ParserConfiguration config, ProgressListener l)
-			throws CityGmlParseException, InvalidGmlFileException {
-		return parseCityGmlStream(is, config, l, null);
+		if (config.getValidate()) {
+			InputStream vis = archive.getInputStream(entry);
+			List<String> messages = validateStream(vis,context);
+			if (!messages.isEmpty()) {
+				throw new InvalidGmlFileException("Invalid GML File. First error: \n" + messages.get(0));
+			}
+		}
+		InputStream is = archive.getInputStream(entry);
+		return parseCityGmlStream(is, config, context);
 	}
 
-	public static CityDoctorModel parseCityGmlStream(InputStream is, ParserConfiguration config, ProgressListener l,
-			 GMLValidationHandler handler) throws CityGmlParseException, InvalidGmlFileException {
-		CityGMLContext context = getContext();
+	private static List<String> validateStream(InputStream vis, CityGMLContext context) throws CityGmlParseException {
+		GMLValidationHandler handler = new GMLValidationHandler();
 
-		if (config.getValidate()) {
-			//TODO: Think of something to XML-validate Inputstream
+		try {
+			BufferedInputStream bis = new BufferedInputStream(vis);
+			SchemaHandler schemaHandler = new ValidationSchemaHandler(context.getDefaultSchemaHandler());
+			readAdditionalSchemaDefinitions(context, bis, schemaHandler);
+			Source[] schemas = schemaHandler.getSchemas();
+			SchemaFactory schemaFactory = SchemaFactory.newInstance(XMLConstants.W3C_XML_SCHEMA_NS_URI);
+			schemaFactory.setFeature("http://apache.org/xml/features/disallow-doctype-decl", true);
+			Schema schema = schemaFactory.newSchema(schemas);
+			Validator validator = schema.newValidator();
+			validator.setErrorHandler(handler);
+			validator.validate(new StreamSource(bis));
+			return handler.getMessages();
+		} catch (SchemaHandlerException | SAXException | IOException e) {
+			throw new CityGmlParseException("Failed to validate CityGML file", e);
+		}
+	}
 
-			//XML validation requires looking at nesting of tags
-			//not a problem for small files, but big files will run into memory limit and crash
-			// Maybe do it dirty by writing stream to temp file and passing it to parseCityGmlFile()?
-			throw new InvalidGmlFileException("Invalid GML File.");
+	private static void readAdditionalSchemaDefinitions(CityGMLContext context, BufferedInputStream bis, SchemaHandler schemaHandler)
+			throws CityGmlParseException {
+		bis.mark(Integer.MAX_VALUE);
+		try (XMLReader reader = XMLReaderFactory.newInstance(context.getXMLObjects())
+				.withSchemaHandler(schemaHandler)
+				.createReader(bis)) {
+			reader.nextTag();
+			bis.reset();
+		} catch (Exception e) {
+			throw new CityGmlParseException("Failed to read Schema from stream.", e);
 		}
+	}
+
+	public static CityDoctorModel parseCityGmlStream(InputStream is, ParserConfiguration config, CityGMLContext context)
+			throws CityGmlParseException {
+		return parseCityGmlStream(is, config, null, context);
+	}
+
+	public static CityDoctorModel parseCityGmlStream(InputStream is, ParserConfiguration config, ProgressListener l, CityGMLContext context)
+			throws CityGmlParseException {
 
 		try {
 			BufferedInputStream bis = new BufferedInputStream(is);

CityDoctorParent/Extensions/CityDoctorZipLoader/src/main/java/de/hft/stuttgart/citydoctor2/ziploader/CityGmlZipEntry.java

@@ -13,7 +13,6 @@ import java.io.IOException;
 import java.io.InputStream;
 import java.util.zip.ZipEntry;
 import java.util.zip.ZipFile;
-import java.util.zip.ZipInputStream;
 
 public class CityGmlZipEntry {
 
@@ -31,9 +30,8 @@ public class CityGmlZipEntry {
             return new ErroneousEntry(entry, ZipEntryErrorType.EXCESSIVE_FILESIZE);
         } else {
             try {
-                InputStream is = archive.getInputStream(entry);
                 CityGmlParser.gagLogger(true);
-                CityDoctorModel model = CityGmlParser.parseCityGmlStream(is, config);
+                CityDoctorModel model = CityGmlParser.parseCityGmlZipEntry(entry, archive, config);
                 return new CityGmlZipEntry(entry.getName(), model);
             } catch (CityGmlParseException | InvalidGmlFileException e) {
                 logger.error(e);