support for XdmNode as Input

cleanup
This commit is contained in:
Andreas Penski (init) 2020-08-13 10:04:34 +02:00
parent f9c6248513
commit a41336e9cf
14 changed files with 227 additions and 38 deletions

View file

@ -29,6 +29,7 @@ import java.net.URI;
import java.net.URL;
import java.net.URLConnection;
import java.nio.file.Path;
import java.util.UUID;
import javax.xml.transform.Source;
import javax.xml.transform.stream.StreamSource;
@ -42,6 +43,9 @@ import de.kosit.validationtool.impl.input.ByteArrayInput;
import de.kosit.validationtool.impl.input.ResourceInput;
import de.kosit.validationtool.impl.input.SourceInput;
import de.kosit.validationtool.impl.input.StreamHelper;
import de.kosit.validationtool.impl.input.XdmNodeInput;
import net.sf.saxon.s9api.XdmNode;
/**
* Service zum Einlesen des Test-Objekts in den Speicher. Beim Einlesen wird gleichzeitig eine Prüfsumme ermittelt und
@ -54,6 +58,11 @@ public class InputFactory {
static final String DEFAULT_ALGORITH = "SHA-256";
/**
* Pseudo hashcode algorithm name, which indicates, thate the hashcode of the {@link Input} is actually the name.
*/
static final String PSEUDO_NAME_ALGORITHM = "NAME";
private static final String MESSAGE_OPEN_STREAM_ERROR = "Can not open stream from";
@Getter
@ -164,15 +173,18 @@ public class InputFactory {
}
/**
* Reads a test document from a {@link Source}. <br/>
* Note: computing the hashcode is only supported for {@link StreamSource}. You can not directly use other {@link Source
* Soures}. You need to supply the hashcode for identification then.
* Reads a test document from a {@link Source}. Note: computing the hashcode is only supported for {@link StreamSource}.
* You can not directly use other {@link Source Soures}. You need to supply the hashcode for identification then.
*
* @param source source
* @return an {@link Input}
*/
public static Input read(final StreamSource source) {
return read(source, DEFAULT_ALGORITH);
public static Input read(final Source source) {
if (source instanceof StreamSource) {
return read(source, source.getSystemId(), DEFAULT_ALGORITH);
}
final String name = UUID.randomUUID().toString();
return read(source, name, PSEUDO_NAME_ALGORITHM, name.getBytes());
}
/**
@ -182,11 +194,16 @@ public class InputFactory {
* Soures}. You need to supply the hashcode for identification then.
*
* @param source source
* @param digestAlgorithm the digest algorithm
* @param name the digest algorithm
* @return an {@link Input}
*/
public static Input read(final StreamSource source, final String digestAlgorithm) {
return read(source, digestAlgorithm, null);
public static Input read(final Source source, final String name) {
checkNotEmpty(name);
return read(source, name, PSEUDO_NAME_ALGORITHM, name.getBytes());
}
public static Input read(final Source source, final String name, final String digestAlgorithm) {
return read(source, name, digestAlgorithm, null);
}
/**
@ -198,7 +215,12 @@ public class InputFactory {
*/
public static Input read(final Source source, final String digestAlgorithm, final byte[] hashcode) {
checkNull(source);
return new SourceInput(source, source.getSystemId(), digestAlgorithm, hashcode);
return read(source, source.getSystemId(), digestAlgorithm, hashcode);
}
public static Input read(final Source source, final String name, final String digestAlgorithm, final byte[] hashcode) {
checkNull(source);
return new SourceInput(source, name, digestAlgorithm, hashcode);
}
/**
@ -281,4 +303,17 @@ public class InputFactory {
return read(new StreamSource(inputStream, name), digestAlgorithm);
}
/**
* Reads a saxon {@link XdmNode} with a given name. Hashcode identification is based on the name of the supplied input.
* Now real hashcode is computed.
*
* @param node the node to read
* @param name the name of the {@link Input}
* @return an {@link Input} to validate
*/
public static Input read(final XdmNode node, final String name) {
checkNull(node);
return new XdmNodeInput(node, name, PSEUDO_NAME_ALGORITHM, name.getBytes());
}
}

View file

@ -9,7 +9,7 @@ import org.w3c.dom.Document;
import net.sf.saxon.s9api.XdmNode;
/**
* API Rückgabe Objekt des Ergebnisses des Validierungsprozesses.
* API result object holding various information of the validation process results.
*
* @author Andreas Penski
*/

View file

@ -1,5 +1,6 @@
package de.kosit.validationtool.daemon;
import static de.kosit.validationtool.impl.Printer.writeOut;
import static org.apache.commons.lang3.StringUtils.defaultIfBlank;
import java.io.IOException;
@ -10,7 +11,6 @@ import java.util.concurrent.Executors;
import com.sun.net.httpserver.HttpHandler;
import com.sun.net.httpserver.HttpServer;
import lombok.AccessLevel;
import lombok.RequiredArgsConstructor;
import lombok.Setter;
import lombok.extern.slf4j.Slf4j;
@ -48,7 +48,7 @@ public class Daemon {
* @param port the port to expose
* @param threadCount the number of working threads
*/
public Daemon(String hostname, int port, int threadCount) {
public Daemon(final String hostname, final int port, final int threadCount) {
this.bindAddress = hostname;
this.port = port;
this.threadCount = threadCount;
@ -73,17 +73,20 @@ public class Daemon {
server.setExecutor(createExecutor());
server.start();
log.info("Server {} started", server.getAddress());
if (!log.isInfoEnabled()) {
writeOut("Server {0} started", server.getAddress());
}
} catch (final IOException e) {
log.error("Error starting HttpServer for Valdidator: {}", e.getMessage(), e);
}
}
private HttpHandler createRootHandler(Configuration config) {
HttpHandler rootHandler;
private HttpHandler createRootHandler(final Configuration config) {
final HttpHandler rootHandler;
final DefaultCheck check = new DefaultCheck(config);
final CheckHandler checkHandler = new CheckHandler(check, config.getContentRepository().getProcessor());
if (guiEnabled) {
GuiHandler gui = new GuiHandler();
if (this.guiEnabled) {
final GuiHandler gui = new GuiHandler();
rootHandler = new RoutingHandler(checkHandler, gui);
} else {
rootHandler = checkHandler;

View file

@ -62,6 +62,7 @@ public class DefaultCheck implements Check {
@Getter
private final ConversionService conversionService;
@Getter
private final Configuration configuration;
@Getter

View file

@ -0,0 +1,21 @@
package de.kosit.validationtool.impl;
import java.text.MessageFormat;
/**
* Wrapper for {@link System Systems} printing capability.
*
* @author Andreas Penski
*/
public class Printer {
/**
* Writes to standard output channel.
*
* @param message the message with placeholders
* @param params the params.
*/
public static void writeOut(final String message, final Object... params) {
System.out.println(MessageFormat.format(message, params));
}
}

View file

@ -1,9 +1,13 @@
package de.kosit.validationtool.impl.input;
import static org.apache.commons.lang3.StringUtils.defaultIfBlank;
import java.io.IOException;
import java.nio.charset.Charset;
import javax.xml.bind.util.JAXBSource;
import javax.xml.transform.Source;
import javax.xml.transform.dom.DOMSource;
import javax.xml.transform.stream.StreamSource;
import org.apache.commons.io.input.ReaderInputStream;
@ -11,21 +15,23 @@ import org.apache.commons.io.input.ReaderInputStream;
import lombok.Getter;
import lombok.extern.slf4j.Slf4j;
import net.sf.saxon.om.TreeInfo;
/**
* A validator {@link de.kosit.validationtool.api.Input} based on a {@link Source}. <br/>
* A validator {@link de.kosit.validationtool.api.Input} based on a {@link Source}.
* <p>
* Note: The various implementations of {@link Source} varies wether the can be read twice or no. This implementation
* tries to handle this with respect document identification (hashcode).
*
* This class is know to work with:
* This class is known to work with:
* <ul>
* <li>{@link StreamSource} - both {@link java.io.InputStream} based and {@link java.io.Reader} based</li>
* <li>{@link javax.xml.transform.dom.DOMSource}</li>
* <li>{@link javax.xml.bind.util.JAXBSource}</li>
* <li>{@link TreeInfo}</li>
* </ul>
*
* Other {@link Source Sources} may work as well, please try and let us know.
* </p>
*
* @author Andreas Penski
*/
@ -51,28 +57,33 @@ public class SourceInput extends AbstractInput {
validate();
}
@Override
public String getName() {
return defaultIfBlank(this.name, this.source.getClass().getSimpleName());
}
private void validate() {
if (!isHashcodeComputed() && isNotSupported()) {
if (!isHashcodeComputed() && !isHashcodeComputationSupported()) {
throw new IllegalStateException("Unsupported source. Only StreamSource supported yet");
}
if (!isHashcodeComputed() && ((StreamSource) this.source).getInputStream() == null) {
log.warn("No hashcode supplied, will wrap the reader using system default charset");
}
if (!(isTreeInfo() || isDomSource() || isStreamSource() || isJaxbSource())) {
log.warn("No known to be working Source implementation provided.");
}
}
@Override
public Source getSource() throws IOException {
if (!isHashcodeComputed() && isNotSupported()) {
throw new IllegalStateException("Unsupported source. Only InputStream-based StreamSource supported yet");
}
if (isConsumed()) {
throw new IllegalStateException("A SourceInput can only read once");
}
return isHashcodeComputed() ? this.source : wrappedSource();
}
private boolean isNotSupported() {
return !isStreamSource();
private boolean isHashcodeComputationSupported() {
return isStreamSource();
}
private boolean isConsumed() throws IOException {
@ -94,6 +105,18 @@ public class SourceInput extends AbstractInput {
return this.source instanceof StreamSource;
}
private boolean isDomSource() {
return this.source instanceof DOMSource;
}
public boolean isTreeInfo() {
return this.source instanceof TreeInfo;
}
private boolean isJaxbSource() {
return this.source instanceof JAXBSource;
}
private Source wrappedSource() {
Source result = this.source;
if (isStreamSource()) {
@ -107,11 +130,9 @@ public class SourceInput extends AbstractInput {
return result;
}
@Override
public boolean supportsMultipleReads() {
return false;
return isDomSource() || isTreeInfo();
}
}

View file

@ -0,0 +1,34 @@
package de.kosit.validationtool.impl.input;
import javax.xml.transform.Source;
import lombok.Getter;
import lombok.RequiredArgsConstructor;
import de.kosit.validationtool.api.Input;
import net.sf.saxon.s9api.XdmNode;
/**
* An {@link Input} implementation holding saxon's {@link XdmNode} object.
*
* @author Andreas Penski
*/
@RequiredArgsConstructor
@Getter
public class XdmNodeInput implements Input {
private final XdmNode node;
private final String name;
private final String digestAlgorithm;
private final byte[] hashCode;
@Override
public Source getSource() {
// usually not neccessary to be called.
return this.node.getUnderlyingNode();
}
}

View file

@ -134,7 +134,7 @@ public interface CheckAction {
* Prüfschritt bedingt auszuführen.
*
* @param results die bisher gesammelten Information
* @return <tt>true</tt> wenn der Schritt ausgelassen werden soll
* @return <code>true</code> wenn der Schritt ausgelassen werden soll
*/
default boolean isSkipped(final Bag results) {
return false;

View file

@ -27,6 +27,7 @@ import lombok.RequiredArgsConstructor;
import lombok.extern.slf4j.Slf4j;
import de.kosit.validationtool.api.Input;
import de.kosit.validationtool.impl.input.XdmNodeInput;
import de.kosit.validationtool.impl.model.Result;
import de.kosit.validationtool.model.reportInput.ValidationResultsWellformedness;
import de.kosit.validationtool.model.reportInput.XMLSyntaxError;
@ -47,6 +48,7 @@ import net.sf.saxon.s9api.XdmNode;
public class DocumentParseAction implements CheckAction {
private final Processor processor;
/**
* Parsed und überprüft ein übergebenes Dokument darauf ob es well-formed ist. Dies stellt den ersten
* Verarbeitungsschritt des Prüf-Tools dar. Diese Funktion verzichtet explizit auf die Validierung gegenüber einem
@ -60,11 +62,17 @@ public class DocumentParseAction implements CheckAction {
throw new IllegalArgumentException("Input may not be null");
}
Result<XdmNode, XMLSyntaxError> result;
try {
final DocumentBuilder builder = this.processor.newDocumentBuilder();
builder.setLineNumbering(true);
final XdmNode doc = builder.build(content.getSource());
result = new Result<>(doc, Collections.emptyList());
if (content instanceof XdmNodeInput && hasCompatibleConfiguration((XdmNodeInput) content)) {
// parsing not neccessary
result = new Result<>(((XdmNodeInput) content).getNode());
} else {
final DocumentBuilder builder = this.processor.newDocumentBuilder();
builder.setLineNumbering(true);
final XdmNode doc = builder.build(content.getSource());
result = new Result<>(doc, Collections.emptyList());
}
} catch (final SaxonApiException | IOException e) {
log.debug("Exception while parsing {}", content.getName(), e);
final XMLSyntaxError error = new XMLSyntaxError();
@ -76,6 +84,10 @@ public class DocumentParseAction implements CheckAction {
return result;
}
private boolean hasCompatibleConfiguration(final XdmNodeInput content) {
return content.getNode().getProcessor().getUnderlyingConfiguration().isCompatible(this.processor.getUnderlyingConfiguration());
}
@Override
public void check(final Bag results) {
final Result<XdmNode, XMLSyntaxError> parserResult = parseDocument(results.getInput());

View file

@ -11,5 +11,5 @@ Currently, there are two public third party validation configurations available.
* Source code is available on [GitHub](https://github.com/itplr-kosit/validator-configuration-xgewerbeanzeige)
* [Releases](https://github.com/itplr-kosit/validator-configuration-xgewerbeanzeige/releases) can also be downloaded
For creating custom configurations see [configaration documentation](https://github.com/itplr-kosit/validator/blob/master/docs/configurations.md)
For creating custom configurations see [configuration documentation](https://github.com/itplr-kosit/validator/blob/master/docs/configurations.md)
for details

View file

@ -19,6 +19,7 @@
package de.kosit.validationtool.api;
import static de.kosit.validationtool.impl.Helper.Simple.SIMPLE_VALID;
import static de.kosit.validationtool.impl.input.StreamHelper.drain;
import static org.assertj.core.api.Assertions.assertThat;
@ -31,7 +32,6 @@ import java.net.URISyntaxException;
import java.net.URL;
import java.nio.file.Paths;
import javax.xml.transform.Source;
import javax.xml.transform.dom.DOMSource;
import javax.xml.transform.stream.StreamSource;
@ -188,8 +188,7 @@ public class InputFactoryTest {
final Document dom = NodeOverNodeInfo.wrap(handler.getDocumentNode().getUnderlyingNode()).getOwnerDocument();
final Input domInput = InputFactory.read(new DOMSource(dom), "MD5", "id".getBytes());
assertThat(domInput).isNotNull();
final Source source = domInput.getSource();
assertThat(source).isNotNull();
assertThat(domInput.getSource()).isNotNull();
final Result<XdmNode, XMLSyntaxError> parsed = Helper.parseDocument(domInput);
assertThat(parsed.isValid()).isTrue();
@ -197,4 +196,17 @@ public class InputFactoryTest {
assertThat(Helper.parseDocument(domInput).getObject()).isNotNull();
}
@Test
public void testXdmNode() throws Exception {
final XdmNode node = TestObjectFactory.createProcessor().newDocumentBuilder().build(new StreamSource(SIMPLE_VALID.toASCIIString()));
final Input nodeInput = InputFactory.read(node, "node test");
assertThat(nodeInput).isNotNull();
assertThat(nodeInput.getSource()).isNotNull();
final Result<XdmNode, XMLSyntaxError> parsed = Helper.parseDocument(nodeInput);
assertThat(parsed.isValid()).isTrue();
// read twice
assertThat(Helper.parseDocument(nodeInput).getObject()).isNotNull();
}
}

View file

@ -35,6 +35,8 @@ import java.util.List;
import java.util.stream.Collectors;
import java.util.stream.IntStream;
import javax.xml.transform.stream.StreamSource;
import org.junit.Before;
import org.junit.Test;
import org.w3c.dom.Document;
@ -42,9 +44,12 @@ import org.w3c.dom.Document;
import de.kosit.validationtool.api.AcceptRecommendation;
import de.kosit.validationtool.api.Configuration;
import de.kosit.validationtool.api.Input;
import de.kosit.validationtool.api.InputFactory;
import de.kosit.validationtool.api.Result;
import de.kosit.validationtool.impl.Helper.Simple;
import net.sf.saxon.s9api.XdmNode;
/**
* Test das Check-Interface
*
@ -238,4 +243,18 @@ public class DefaultCheckTest {
assertThat(result.getProcessingErrors()).hasSize(1);
}
@Test
public void testXdmNode() throws Exception {
XdmNode node = TestObjectFactory.createProcessor().newDocumentBuilder().build(new StreamSource(SIMPLE_VALID.toASCIIString()));
Input domInput = InputFactory.read(node, "node test");
Result result = this.validCheck.checkInput(domInput);
assertThat(result.isProcessingSuccessful()).isEqualTo(true);
// test compatible configuration
node = this.validCheck.getConfiguration().getContentRepository().getProcessor().newDocumentBuilder()
.build(new StreamSource(SIMPLE_VALID.toASCIIString()));
domInput = InputFactory.read(node, "node test");
result = this.validCheck.checkInput(domInput);
assertThat(result.isProcessingSuccessful()).isEqualTo(true);
}
}