org.jopendocument.dom.OOXML Maven / Gradle / Ivy

Go to download
/*
 * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS HEADER.
 * 
 * Copyright 2008 jOpenDocument, by ILM Informatique. All rights reserved.
 * 
 * The contents of this file are subject to the terms of the GNU
 * General Public License Version 3 only ("GPL").  
 * You may not use this file except in compliance with the License. 
 * You can obtain a copy of the License at http://www.gnu.org/licenses/gpl-3.0.html
 * See the License for the specific language governing permissions and limitations under the License.
 * 
 * When distributing the software, include this License Header Notice in each file.
 * 
 */

/*
 * Créé le 28 oct. 2004
 * 
 */
package org.jopendocument.dom;

import org.jopendocument.util.JDOMUtils;

import java.util.ArrayList;
import java.util.HashMap;
import java.util.Iterator;
import java.util.List;
import java.util.Map;
import java.util.Map.Entry;
import java.util.regex.Matcher;
import java.util.regex.Pattern;

import javax.xml.XMLConstants;

import org.apache.commons.collections.Transformer;
import org.apache.commons.collections.map.LazyMap;
import org.jdom.Content;
import org.jdom.Document;
import org.jdom.Element;
import org.jdom.JDOMException;
import org.jdom.Namespace;
import org.jdom.Parent;
import org.jdom.Text;
import org.jdom.xpath.XPath;
import org.xml.sax.SAXException;

/**
 * Various bits of OpenDocument XML.
 * 
 * @author Sylvain CUAZ
 * @see #get(String)
 */
public class OOXML {

    private static final Map instances = LazyMap.decorate(new HashMap(), new Transformer() {
        public Object transform(Object input) {
            return new OOXML((String) input);
        }
    });

    /**
     * Returns the instance that match the requested version.
     * 
     * @param version either {@link OOUtils#OOo} or {@link OOUtils#OOo}.
     * @return the corresponding instance.
     */
    public static OOXML get(String version) {
        return (OOXML) instances.get(version);
    }

    static public final String getTabS() {
        return "";
    }

    static public final String getLineBreakS() {
        return "";
    }

    static private final String rt2oo(String content, String tagName, String styleName) {
        return content.replaceAll("\\[" + tagName + "\\]", "").replaceAll("\\[/" + tagName + "\\]", "");
    }

    /**
     * Encode spaces for OpenOffice, and escape characters for XML.
     * 
     * @param s a string to encode, eg "hi\n 3<4".
     * @return the string encoded in XML, eg "hi3<4".
     * @see #encodeWS(String)
     */
    static public final String encodeOOWS(final String s) {
        String tmp = JDOMUtils.OUTPUTTER.escapeElementEntities(s).replaceAll("\n", getLineBreakS()).replaceAll("\t", getTabS());

        String res = "";
        // les séries de plus d'un espace consécutifs
        final Pattern p = Pattern.compile("  +");
        final Matcher m = p.matcher(tmp);
        int lastEnd = 0;
        while (m.find()) {
            // c == le nombre d'espaces
            res += tmp.substring(lastEnd, m.start()) + "";
            lastEnd = m.end();
        }
        res += tmp.substring(lastEnd);

        return res;
    }

    // *** instances

    private final String version;

    private OOXML(String version) {
        this.version = version;
    }

    public final String getVersion() {
        return this.version;
    }

    private NS getNS() {
        return NS.get(getVersion());
    }

    /**
     * Verify that the passed document is a valid OpenOffice.org 1 or ODF document.
     * 
     * @param doc the xml to test.
     * @return null if doc is valid, a String describing the pb
     *         otherwise.
     */
    public final String isValid(Document doc) {
        if (this.getVersion().equals(OOUtils.OD))
            try {
                return JDOMUtils.isValid(doc, XMLConstants.RELAXNG_NS_URI, getClass().getResource("oofficeDTDs/OpenDocument-strict-schema-v1.1.rng"));
            } catch (SAXException e) {
                throw new IllegalStateException("relaxNG schemas pb", e);
            }
        else {
            // DTDs are stubborn, xmlns have to be exactly where they want
            // in this case the root element
            for (final Namespace n : getNS().getALL())
                doc.getRootElement().addNamespaceDeclaration(n);
            return JDOMUtils.isValidDTD(doc, OOUtils.getBuilderLoadDTD());
        }
    }

    public final Element getLineBreak() {
        return new Element("line-break", getNS().getTEXT());
    }

    public final Element getTab() {
        return new Element("tab-stop", getNS().getTEXT());
    }

    protected final List encodeRT_L(String content, Map styles) {
        String res = JDOMUtils.OUTPUTTER.escapeElementEntities(content);
        final Iterator iter = styles.entrySet().iterator();
        while (iter.hasNext()) {
            final Entry e = (Entry) iter.next();
            res = rt2oo(res, (String) e.getKey(), (String) e.getValue());
        }
        try {
            return JDOMUtils.parseString(res, getNS().getALL());
        } catch (JDOMException e) {
            // should not happpen as we did escapeElementEntities which gives valid xml and then
            // rt2oo which introduce only static xml
            throw new IllegalStateException("could not parse " + res, e);
        }
    }

    /**
     * Convert rich text (with [] tags) into XML.
     * 
     * @param content the string to convert, eg "texte [b]gras[/b]".
     * @param styles the mapping from tagname (eg "b") to the name of the character style (eg
     *        "Gras").
     * @return the corresponding element.
     */
    public final Element encodeRT(String content, Map styles) {
        return new Element("span", getNS().getTEXT()).addContent(encodeRT_L(content, styles));
    }

    // create the necessary 
    private Element createSpaces(String spacesS) {
        return new Element("s", getNS().getTEXT()).setAttribute("c", spacesS.length() + "", getNS().getTEXT());
    }

    /**
     * Encode a String to OO XML. Handles substition of whitespaces to their OO equivalent.
     * 
     * @param s a plain ole String, eg "term\tdefinition".
     * @return an Element suitable to be inserted in an OO XML document, eg
     * 
     *      *     <text:span>term<text:tab-stop/>definition</text:span>
     * .
     */
    public final Element encodeWS(final String s) {
        return new Element("span", getNS().getTEXT()).setContent(encodeWSasList(s));
    }

    private final List encodeWSasList(final String s) {
        final List res = new ArrayList();
        final Matcher m = Pattern.compile("\n|\t| {2,}").matcher(s);
        int last = 0;
        while (m.find()) {
            res.add(new Text(s.substring(last, m.start())));
            switch (m.group().charAt(0)) {
            case '\n':
                res.add(getLineBreak());
                break;
            case '\t':
                res.add(getTab());
                break;
            case ' ':
                res.add(createSpaces(m.group()));
                break;

            default:
                throw new IllegalStateException("unknown item: " + m.group());
            }
            last = m.end();
        }
        res.add(new Text(s.substring(last)));
        return res;
    }

    @SuppressWarnings("unchecked")
    public final void encodeWS(final Text t) {
        final Parent parent = t.getParent();
        final int ind = parent.indexOf(t);
        t.detach();
        parent.getContent().addAll(ind, encodeWSasList(t.getText()));
    }

    @SuppressWarnings("unchecked")
    public final Element encodeWS(final Element elem) {
        final XPath path;
        try {
            path = OOUtils.getXPath(".//text()", NS.getVersion(elem));
        } catch (JDOMException e) {
            // static path, hence always valid
            throw new IllegalStateException("cannot create XPath", e);
        }
        try {
            final Iterator iter = new ArrayList(path.selectNodes(elem)).iterator();
            while (iter.hasNext()) {
                final Text t = (Text) iter.next();
                encodeWS(t);
            }
        } catch (JDOMException e) {
            throw new IllegalArgumentException("cannot find text nodes of " + elem, e);
        }
        return elem;
    }

}