saxwriterfilter.java

一个自然语言处理的Java开源工具包。LingPipe目前已有很丰富的功能

            mWriter.flush(); 
        } catch (IOException e) { 
            // do nothing
        }
    }

    /**
     * Prints the start element, using the qualified name, and sorting
     * the attributes using the underlying string ordering.  Namespace
     * URI and local names are ignored, and qualified name must not be
     * <code>null</code>.
     *
     * @param namespaceURI The URI of the namespace for this element.
     * @param localName The local name (without prefix) for this
     * element.
     * @param qName The qualified name (with prefix, if any) for this
     * element.
     * @param atts The attributes for this element.
     */
    public void startElement(String namespaceURI, String localName,
                             String qName, Attributes atts)
        throws SAXException {

        super.startElement(namespaceURI,localName,qName,atts);

        if (mStartedElement) {
            mPrinter.print(">");
            mStartedElement=false;
        }
        mPrinter.print('<');
        mPrinter.print(qName);
        printAttributes(atts);
        mStartedElement = true;
        // close is picked up later in implicit continuation
        // mPrinter.print('>');
    }

    /**
     * Prints the end element, using the qualified name.  Namespace
     * URI and local name parameters are ignored, and the qualified
     * name must not be <code>null</code>
     *
     * @param namespaceURI The URI of the namespace for this element.
     * @param localName The local name (without prefix) for this
     * element.
     * @param qName The qualified name (with prefix, if any) for this
     * element.
     */
    public void endElement(String namespaceURI, String localName,
                           String qName)
        throws SAXException{

        super.endElement(namespaceURI,localName,qName);
        if (mStartedElement) {
            mPrinter.print("/>");
            mStartedElement = false;
            return;
        }
        mPrinter.print('<');
        mPrinter.print('/');
        mPrinter.print(qName);
        mPrinter.print('>');
    }

    /**
     * Prints the characters in the specified range.
     *
     * @param ch Character array from which to draw characters.
     * @param start Index of first character to print.
     * @param length Number of characters to print.
     */
    public void characters(char[] ch, int start, int length)
        throws SAXException {
        super.characters(ch,start,length);
        if (mStartedElement) {
            mPrinter.print('>');
            mStartedElement = false;
        }
        printCharacters(ch,start,length);
    }

    /**
     * Does not print ignorable whitespace, nor does it send the
     * event to the contained handler.
     *
     * @param ch Character array from which to draw characters.
     * @param start Index of first character to print.
     * @param length Number of characters to print.
     */
    public void ignorableWhitespace(char[] ch, int start, int length)
        throws SAXException {
        // super.ignorableWhitespace(ch,start,length);
    }

    /**
     * Print a representation of the proecssing instruction.  This
     * will be <code>&langle;?<i>Target</i>&rangle;</code> if there is
     * no data, or * <code>&lt;?<i>Target</i>
     * <i>Data</i>&gt;</code> if there is data.
     *
     * @param target Target of the instruction.
     * @param data Value of the instruction, or <code>null</code>.
     */
    public void processingInstruction(String target, String data)
        throws SAXException {

        super.processingInstruction(target,data);
        if (mStartedElement) {
            mPrinter.print('>');
            mStartedElement = false;
        }
        mPrinter.print("<?");
        mPrinter.print(target);
        if (data != null && data.length() > 0) {
            mPrinter.print(' ');
            mPrinter.print(data);
        }
        mPrinter.print("?>");
    }

    // prints atts and outstanding namespace decls
    private void printAttributes(Attributes atts) {
        if (mPrefixMap.size() > 0) {
            Iterator it = mPrefixMap.entrySet().iterator();
            while (it.hasNext()) {
                Map.Entry entry = (Map.Entry) it.next();
                String key = entry.getKey().toString();
                String value = entry.getValue().toString();
                printAttribute("xmlns:" + key,value);
            }
            mPrefixMap.clear();
        }
        TreeSet orderedAtts = new TreeSet();
        for (int i = 0; i < atts.getLength(); ++i)
            orderedAtts.add(atts.getQName(i));
        Iterator attsIterator = orderedAtts.iterator();
        while (attsIterator.hasNext()) {
            String attQName = attsIterator.next().toString();
            printAttribute(attQName,atts.getValue(attQName));
        }
    }

    /**
     * Prints an attribute and value, with value properly
     * escaped. Prints leading space before attribute and value pair.
     *
     * @param att Attribute name.
     * @param val Attribute value.
     */
    private void printAttribute(String att, String val) {
        mPrinter.print(' ');
        mPrinter.print(att);
        mPrinter.print('=');
        mPrinter.print('"');
        printCharacters(val);
        mPrinter.print('"');
    }

    /**
     * Print the specified string, with appropriate escapes.
     *
     * @param s Print the characters in the specified string.
     */
    private void printCharacters(String s) {
        printCharacters(s.toCharArray(),0,s.length());
    }

    /**
     * Print the specified range of characters, with escapes.
     *
     * @param ch Array of characters from which to draw.
     * @param start Index of first character to print.
     * @param length Number of characters to print.
     */
    private void printCharacters(char[] ch, int start, int length) {
        for (int i = start; i < start+length; ++i)
            printCharacter(ch[i]);
    }

    /**
     * Print the specified character, rendering it as an entity if
     * necessary (see class doc).
     *
     * @param c Character to print.
     */
    private void printCharacter(char c) {
        // note: does not catch illegal conjugate pairs
        if (!mCharsetEncoder.canEncode(c)) {
            printEntity("#x" + Integer.toHexString((int)c));
            return;
        }
        switch (c) {
        case '<':  { printEntity("lt"); break; }
        case '>':  { printEntity("gt"); break; }
        case '&':  { printEntity("amp"); break; }
        case '"':  { printEntity("quot"); break; }
        default:   { mPrinter.print(c); }
        }
    }

    /**
     * Print the specified entity.
     *
     * @param entity Name of entity to print.
     */
    private void printEntity(String entity) {
        mPrinter.print('&');
        mPrinter.print(entity);
        mPrinter.print(';');
    }

    /**
     * Prints the XML declaration, including the character set
     * declaration and the DTD if any is defined.  The declaration
     * printed is:
     * <blockquote>
     *   <code>&lt;?xml version="1.0" encoding="<i>CharSet</i>"?&gt;</code>.
     * </blockquote>
     * where <code><i>CharSet</i></code> is the string representation
     * of the character set being used.  No spaces are included after
     *  the XML declaration or the DTD declaration, if any.
     */
    private void printXMLDeclaration() {
        mPrinter.print("<?xml");
        printAttribute("version","1.0");
        printAttribute("encoding",mCharsetName);
        mPrinter.print("?>");
        if (mDtdString != null) {
            mPrinter.print(mDtdString);
        }
    }

}