Examples of org.cyberneko.html.HTMLConfiguration

            remover,
            writer,
        };


        // create HTML parser
        XMLParserConfiguration parser = new HTMLConfiguration();
        parser.setProperty("http://cyberneko.org/html/properties/filters", filters);


        // parse documents
        for (int i = 0; i < argv.length; i++) {
            String systemId = argv[i];
            XMLInputSource source = new XMLInputSource(null, systemId, null);
            parser.parse(source);
        }


    } // main(String[])

    // MAIN
    //


    /** Main. */
    public static void main(String[] argv) throws Exception {
        HTMLConfiguration parser = new HTMLConfiguration();
        parser.setFeature(AUGMENTATIONS, true);
        XMLDocumentFilter[] filters = { new Script(parser), new Identity(), new Writer() };
        parser.setProperty(FILTERS, filters);
        for (int i = 0; i < argv.length; i++) {
            parser.parse(new XMLInputSource(null, argv[i], null));
        }
    } // main(String[])

    // Constructors
    //


    /** Default constructor. */
    public SAXParser() {
        super(new HTMLConfiguration());
    } // <init>()

    // Constructors
    //


    /** Default constructor. */
    public DOMFragmentParser() {
        fParserConfiguration = new HTMLConfiguration();
        fParserConfiguration.addRecognizedFeatures(RECOGNIZED_FEATURES);
        fParserConfiguration.addRecognizedProperties(RECOGNIZED_PROPERTIES);
        fParserConfiguration.setFeature(DOCUMENT_FRAGMENT, true);
        fParserConfiguration.setDocumentHandler(this);
    } // <init>()

    // Constructors
    //


    /** Default constructor. */
    public DOMParser() {
        super(new HTMLConfiguration());
        /*** extending DOMParser ***/
        try {
            setProperty("http://apache.org/xml/properties/dom/document-class-name",
                                       "org.apache.html.dom.HTMLDocumentImpl");
        }

    public static void main(String[] argv) throws Exception {
        if (argv.length == 0) {
            printUsage();
            System.exit(1);
        }
        XMLParserConfiguration parser = new HTMLConfiguration();
        parser.setFeature(NOTIFY_CHAR_REFS, true);
        parser.setFeature(NOTIFY_HTML_BUILTIN_REFS, true);
        String iencoding = null;
        String oencoding = "Windows-1252";
        boolean identity = false;
        boolean purify = false;
        for (int i = 0; i < argv.length; i++) {
            String arg = argv[i];
            if (arg.equals("-ie")) {
                iencoding = argv[++i];
                continue;
            }
            if (arg.equals("-e") || arg.equals("-oe")) {
                oencoding = argv[++i];
                continue;
            }
            if (arg.equals("-i")) {
                identity = true;
                continue;
            }
            if (arg.equals("-p")) {
                purify = true;
                continue;
            }
            if (arg.equals("-h")) {
                printUsage();
                System.exit(1);
            }
            java.util.Vector filtersVector = new java.util.Vector(2);
            if (identity) {
                filtersVector.addElement(new Identity());
            }
            else if (purify) {
                filtersVector.addElement(new Purifier());
            }
            filtersVector.addElement(new Writer(System.out, oencoding));
            XMLDocumentFilter[] filters = 
                new XMLDocumentFilter[filtersVector.size()];
            filtersVector.copyInto(filters);
            parser.setProperty(FILTERS, filters);
            XMLInputSource source = new XMLInputSource(null, arg, null);
            source.setEncoding(iencoding);
            parser.parse(source);
        }
    } // main(String[])

    // Constructors
    //


    /** Default constructor. */
    public HTMLSAXParser() {
        super(new HTMLConfiguration());
    } // <init>()


    /**
     * Constructor
     */
    public FavIconSaxParser() {
        super(new HTMLConfiguration());
    }

    public static void main(String[] argv) throws Exception {
        if (argv.length == 0) {
            printUsage();
            System.exit(1);
        }
        XMLParserConfiguration parser = new HTMLConfiguration();
        parser.setFeature(NOTIFY_CHAR_REFS, true);
        parser.setFeature(NOTIFY_HTML_BUILTIN_REFS, true);
        String encoding = "Windows-1252";
        boolean identity = false;
        boolean purify = false;
        for (int i = 0; i < argv.length; i++) {
            String arg = argv[i];
            if (arg.equals("-e")) {
                encoding = argv[++i];
                continue;
            }
            if (arg.equals("-i")) {
                identity = true;
                continue;
            }
            if (arg.equals("-p")) {
                purify = true;
                continue;
            }
            if (arg.equals("-h")) {
                printUsage();
                System.exit(1);
            }
            java.util.Vector filtersVector = new java.util.Vector(2);
            if (identity) {
                filtersVector.addElement(new Identity());
            }
            else if (purify) {
                filtersVector.addElement(new Purifier());
            }
            filtersVector.addElement(new Writer(System.out, encoding));
            XMLDocumentFilter[] filters = 
                new XMLDocumentFilter[filtersVector.size()];
            filtersVector.copyInto(filters);
            parser.setProperty(FILTERS, filters);
            parser.parse(new XMLInputSource(null, arg, null));
        }
    } // main(String[])

    // Constructors
    //


    /** Default constructor. */
    public DOMParser() {
        super(new HTMLConfiguration());
        /*** extending DOMParser ***/
        try {
            setProperty("http://apache.org/xml/properties/dom/document-class-name",
                                       "org.apache.html.dom.HTMLDocumentImpl");
        }

Examples of org.cyberneko.html.HTMLConfiguration

Related Classes of org.cyberneko.html.HTMLConfiguration