Laboratoire ICAR » Plateforme TXM

import javax.xml.stream.XMLStreamReader;

import javax.xml.stream.XMLStreamWriter;

/**

 * XML Identity transformation. extends the processXYZ methods to do something

 * @author mdecorde

 *

 */

public class StaxIdentityParser {

	/** The input */

	protected URL inputurl;

	protected InputStream inputData;

	protected XMLInputFactory factory;

	protected XMLStreamReader parser;

	/** The output. */

	protected XMLOutputFactory outfactory = XMLOutputFactory.newInstance();

	protected BufferedOutputStream output;

	protected XMLStreamWriter writer;

	public static String TXMNS = "http://textometrie.org/1.0";

	public static String TXM = "txm";

	public static String TEINS = "http://www.tei-c.org/ns/1.0";

	public static String TEI = "tei";

	protected PersonalNamespaceContext Nscontext = new PersonalNamespaceContext();

	// protected StringBuilder currentXPath = new StringBuilder("/")

	protected String localname;

	int processingXInclude = 0;

	public StaxIdentityParser(File infile) throws IOException, XMLStreamException {

		this(infile.toURI().toURL());

	}

	public StaxIdentityParser(URL inputurl) throws IOException, XMLStreamException {

		this.inputurl = inputurl;

		this.inputData = inputurl.openStream();

		this.parser = factory.createXMLStreamReader(inputData);

	}

	/**

	 * Helper method to get an attribute value

	 * 

	 */

	public String getParserAttributeValue(String name) {

		if (name == null) return null;

		int c = parser.getAttributeCount();

		for (int i = 0; i < c; i++) {

			if (name.equals(parser.getAttributeLocalName(i))) {

				return parser.getAttributeValue(i);

			}

		}

		return null;

	}

	protected void before() {

	}

	protected void after() throws XMLStreamException, IOException {

		factory = null;

		if (parser != null) parser.close();

		writer = null;

		parser = null;

	}

	protected void closeForError() throws XMLStreamException, IOException {

		if (parser != null) parser.close();

		if (inputData != null) inputData.close();

	}

	/**

	 * Creates the output.

	 *

	 * @param outfile the outfile

	 * @return true, if successful

	 */

		try {

			if (writer != null) // process from a file

				writer.close();

			if (output != null) // process from a file

				output.close();

			output = new BufferedOutputStream(new FileOutputStream(f), 16 * 1024);

			writer.setNamespaceContext(Nscontext);

			return true;

		}

			return false;

		}

	}

	public boolean process(File outfile) throws XMLStreamException, IOException {

		if (writer != null) {

			writer.close();

		}

				System.out.println("output excep: " + e);

			}

		}

		if (parser != null) {

			try {

				parser.close();

				System.out.println("parser excep: " + e);

			}

		}

		if (inputData != null) {

			try {

				inputData.close();

				System.out.println("inputData excep: " + e);

			}

		}

		return ret;

	}

	public final static String SLASH = "/";

	public boolean firstElementStarted = false;

	public boolean process(XMLStreamWriter awriter) throws XMLStreamException, IOException {

		firstElementStarted = false;

		this.writer = awriter;

		// if (processingXInclude == 0) {

		before(); // if you need to do something before parsing the XML

		// }

		try {

			for (int event = parser.getEventType(); event != XMLStreamConstants.END_DOCUMENT; event = parser.next()) {

				switch (event) {

				}

			}

		}

		// }

		return true;

	}

	/**

	 * The start element has already been written

	 * 

			for (int event = parser.next(); event != XMLStreamConstants.END_DOCUMENT; event = parser.next()) {

				// System.out.println("event "+event

				switch (event) {

				}

			}

		}

			return;

		}

	}

	public String getLocation() {

		if (parser != null)

			return "Line: " + parser.getLocation().getLineNumber() + " Col: " + parser.getLocation().getColumnNumber();

		return null;

	}

	protected void processNamespace() throws XMLStreamException {

		writer.writeNamespace(parser.getPrefix(), parser.getNamespaceURI());

	}

	public static final String INCLUDE = "include";

	public static final String XI = "xi";

	protected void processStartElement() throws XMLStreamException, IOException {

		String prefix = parser.getPrefix();

		if (INCLUDE == localname && XI == prefix) {

				writer.writeStartElement(Nscontext.getNamespaceURI(prefix), localname);

			else

				writer.writeStartElement(localname);

			for (int i = 0; i < parser.getNamespaceCount(); i++) {

				writer.writeNamespace(parser.getNamespacePrefix(i), parser.getNamespaceURI(i));

			}

			writeAttributes();

		}

	}

	private void _processStartElement() throws XMLStreamException, IOException {

		String prefix = parser.getPrefix();

		if (INCLUDE == localname && XI == prefix) {

				writer.writeStartElement(Nscontext.getNamespaceURI(prefix), localname);

			else

				writer.writeStartElement(localname);

			for (int i = 0; i < parser.getNamespaceCount(); i++) {

				writer.writeNamespace(parser.getNamespacePrefix(i), parser.getNamespaceURI(i));

			}

			writeAttributes();

		}

	}

	protected void writeAttributes() throws XMLStreamException {

		for (int i = 0; i < parser.getAttributeCount(); i++) {

			writeAttribute(parser.getAttributePrefix(i), parser.getAttributeLocalName(i), parser.getAttributeValue(i));

		}

	}

	protected void writeAttribute(String prefix, String name, String value) throws XMLStreamException {

		if (prefix != null && prefix.length() > 0)

			writer.writeAttribute(prefix + ":" + name, value);

		else

			writer.writeAttribute(name, value);

	}

	protected void processCharacters() throws XMLStreamException {

		writer.writeCharacters(parser.getText());

	}

	protected void processProcessingInstruction() throws XMLStreamException {

		writer.writeProcessingInstruction(parser.getPITarget(), parser.getPIData());

		if (!firstElementStarted) writer.writeCharacters("\n"); // new lines are not reported before the first element is parsed

	}

	protected void processDTD() throws XMLStreamException {

		writer.writeDTD(parser.getText());

	}

	protected void processCDATA() throws XMLStreamException {

		writer.writeCData(parser.getText());

	}

	protected void processComment() throws XMLStreamException {

		writer.writeComment(parser.getText());

	}

	protected void processEndElement() throws XMLStreamException {

		if (localname == INCLUDE && parser.getPrefix() == XI) {

			// nothing !!

			writer.writeEndElement();

		}

	}

	protected void processEndDocument() throws XMLStreamException {

		writer.writeEndDocument();

	}

	protected void processEntityReference() throws XMLStreamException {

		writer.writeEntityRef(parser.getLocalName());

	}

	/**

	 * Process the XInclude elements

	 * 

			System.out.println("Warning referenced file: $ref does not exists");

		}

	}

	public static void main(String[] args) {

		try {

			File input = new File("/home/mdecorde/xml/identity/test.xml");

			if (!(input.exists() && input.canRead())) {

				System.out.println("cannot found $input");

				return;

			}

			StaxIdentityParser builder;

			builder = new StaxIdentityParser(input.toURI().toURL());

			}

			else {

				System.out.println("failure !");

		cwbEn.setDebug(debug);

		cwbMa.setDebug(debug);

		String textAttr ="text:0+base+project"

		if (trans != null) {

			for (String key : trans.keySet()) {

					cqpFile.getAbsolutePath(),

					registryFile.getAbsolutePath(), pAttributes, sAttributes);

			if (!registryFile.exists()) {

				return false;

			}

			cwbMa.run(corpusname, registryFile.getParent());

								String name = parser.getAttributeLocalName(i).replace("_","").toLowerCase()

								output.write(" "+name+"=\""+parser.getAttributeValue(i).replace("\"", """)+"\"");

								if (name == "time") formatedTime = parser.getAttributeValue(i)

							}

							output.write ">\n"

							break;

				writer.writeEndElement() // teiHeader

				writer.writeStartElement("text")

				writer.writeAttribute("id", textid)

				beforeBody = false; // end of info parsing

				break;

		}

		def attributes = speakers.get(vSpeaker)

		if (attributes == null) { // in case of Who@n wrong number

			if (vSpeaker.startsWith("#") && vSpeaker.endsWith("?")) { // don't show "N/A" vSpeaker

				writer.writeAttribute("spkid", vSpeaker)

				writeAttributes();

			}

//				println " write attribute "+p.getFirst()+" "+p.getSecond()

				String attrn = p.getFirst().toString();

				if (attrn == "name") { // rename @name to @spk

					u_name = p.getSecond();

				} else if (attrn == "id") { // rename @id to @spkid

					attrn = "spkid"

				}

			}

		}

		uOpened = true;

	}

	private String formatTime(String time) {

		try {

			return formatTime(Float.parseFloat(time))