Package org.cyberneko.html.parsers

Examples of org.cyberneko.html.parsers.DOMFragmentParser

132
133
134
135
136
137
138
139
140
141
142
    properties = OutputPropertiesFactory.getDefaultMethodProperties(Method.HTML);
    _HTMLserialiser = SerializerFactory.getSerializer(properties);
    _serializer = _HTMLserialiser;
    // serializer.setOutputStream(output);
    // _parser.setContentHandler(serializer.asContentHandler());
    viewStateParser = new DOMFragmentParser();
    // Set parser features
    try {
      viewStateParser
          .setProperty(
              "http://cyberneko.org/html/properties/names/elems",
View Full Code Here
123
124
125
126
127
128
129
130
131
132
133
    // OutputPropertiesFactory.getDefaultMethodProperties(Method.XML);
    // properties.put("encoding",_encoding);
    _serializer = SerializerFactory.getSerializer(properties);
    // serializer.setOutputStream(output);
    // _parser.setContentHandler(serializer.asContentHandler());
    viewStateParser = new DOMFragmentParser();
    // Set parser features
    try {
      viewStateParser
          .setProperty(
              "http://cyberneko.org/html/properties/names/elems",
View Full Code Here
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
      Document htmlDoc = documentProvider.createDocument(null, null, null);
      // Workaround for error check failure adding text node to entity ref as a child
      htmlDoc.setStrictErrorChecking(false);
      DocumentFragment fragment = htmlDoc.createDocumentFragment();
      InputSource input = new InputSource(new StringReader(source));
      DOMFragmentParser parser = new DOMFragmentParser();
      parser.setProperty("http://cyberneko.org/html/properties/names/elems", "default");
      parser.setFeature("http://cyberneko.org/html/features/document-fragment", true);
      parser.setProperty("http://cyberneko.org/html/properties/names/attrs", "no-change");
      parser.setFeature("http://apache.org/xml/features/scanner/notify-char-refs", true);
      parser.setFeature("http://cyberneko.org/html/features/scanner/notify-builtin-refs", true);
      parser.parse(input, fragment);
      return fragment;
    } catch (Exception e) {
      throw new GadgetException(GadgetException.Code.HTML_PARSE_ERROR, e);
    }
  }
View Full Code Here
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
  @Before
  public void setup() throws Exception {
    conf = NutchConfiguration.create();
    conf.setBoolean("parser.html.form.use_action", true);
    utils = new DOMContentUtils(conf);
    DOMFragmentParser parser = new DOMFragmentParser();
    parser.setFeature(
        "http://cyberneko.org/html/features/scanner/allow-selfclosing-iframe",
        true);
    for (int i = 0; i < testPages.length; i++) {
      DocumentFragment node = new HTMLDocumentImpl().createDocumentFragment();
      try {
        parser.parse(
            new InputSource(new ByteArrayInputStream(testPages[i].getBytes())),
            node);
        testBaseHrefURLs[i] = new URL(testBaseHrefs[i]);
      } catch (Exception e) {
        Assert.assertTrue("caught exception: " + e, false);
View Full Code Here
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
  private URL[][] currURLsAndAnswers;

  @Test
  public void testRobotsMetaProcessor() {
    DOMFragmentParser parser= new DOMFragmentParser();;

    try {
      currURLsAndAnswers= new URL[][] {
          {new URL("http://www.nutch.org"), null},
          {new URL("http://www.nutch.org"), null},
          {new URL("http://www.nutch.org"), null},
          {new URL("http://www.nutch.org"), null},
          {new URL("http://www.nutch.org"), null},
          {new URL("http://www.nutch.org"), null},
          {new URL("http://www.nutch.org"), null},
          {new URL("http://www.nutch.org/foo/"),
            new URL("http://www.nutch.org/")},
            {new URL("http://www.nutch.org"),
              new URL("http://www.nutch.org/base/")}
      };
    } catch (Exception e) {
      Assert.assertTrue("couldn't make test URLs!", false);
    }

    for (int i= 0; i < tests.length; i++) {
      byte[] bytes= tests[i].getBytes();

      DocumentFragment node = new HTMLDocumentImpl().createDocumentFragment();

      try {
        parser.parse(new InputSource(new ByteArrayInputStream(bytes)), node);
      } catch (Exception e) {
        e.printStackTrace();
      }

      HTMLMetaTags robotsMeta= new HTMLMetaTags();
View Full Code Here
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
  private static void setup() throws Exception {
    conf = NutchConfiguration.create();
    conf.setBoolean("parser.html.form.use_action", true);
    utils = new DOMContentUtils(conf);
    DOMFragmentParser parser = new DOMFragmentParser();
    parser.setFeature(
        "http://cyberneko.org/html/features/scanner/allow-selfclosing-iframe",
        true);
    for (int i = 0; i < testPages.length; i++) {
      DocumentFragment node = new HTMLDocumentImpl().createDocumentFragment();
      try {
        parser.parse(
            new InputSource(new ByteArrayInputStream(testPages[i].getBytes())),
            node);
        testBaseHrefURLs[i] = new URL(testBaseHrefs[i]);
      } catch (Exception e) {
        assertTrue("caught exception: " + e, false);
View Full Code Here
37
38
39
40
41
42
43
44
45
46
47
48
49
    // MAIN
    //

    /** Main. */
    public static void main(String[] argv) throws Exception {
        DOMFragmentParser parser = new DOMFragmentParser();
        HTMLDocument document = new HTMLDocumentImpl();
        for (int i = 0; i < argv.length; i++) {
            DocumentFragment fragment = document.createDocumentFragment();
            parser.parse(argv[i], fragment);
            print(fragment, "");
        }
    } // main(String[])
View Full Code Here
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
    public void testInvalidAttributeName() throws Exception {
      doTest("<html 9='id'></html>", "<HTML/>");
    }

    private void doTest(final String html, final String expected) throws Exception {
        DOMFragmentParser parser = new DOMFragmentParser();
        HTMLDocument document = new HTMLDocumentImpl();

        DocumentFragment fragment = document.createDocumentFragment();
        InputSource source = new InputSource(new StringReader(html));
        parser.parse(source, fragment);
//        final OutputFormat of = new OutputFormat();
//        of.setOmitXMLDeclaration(true);
//        XMLSerializer s = new XMLSerializer(of);
//        StringWriter sw = new StringWriter();
//        s.setOutputCharStream(sw);
View Full Code Here
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
    reader.parse(input);
    return frag;
  }

  private DocumentFragment parseNeko(InputSource input) throws Exception {
    DOMFragmentParser parser = new DOMFragmentParser();
    try {
      parser.setFeature("http://cyberneko.org/html/features/augmentations",
          true);
      parser.setProperty("http://cyberneko.org/html/properties/default-encoding",
          defaultCharEncoding);
      parser.setFeature("http://cyberneko.org/html/features/scanner/ignore-specified-charset",
          true);
      parser.setFeature("http://cyberneko.org/html/features/balance-tags/ignore-outside-content",
          false);
      parser.setFeature("http://cyberneko.org/html/features/balance-tags/document-fragment",
          true);
      parser.setFeature("http://cyberneko.org/html/features/report-errors",
          LOG.isTraceEnabled());
    } catch (SAXException e) {}
    // convert Document to DocumentFragment
    HTMLDocumentImpl doc = new HTMLDocumentImpl();
    doc.setErrorChecking(false);
    DocumentFragment res = doc.createDocumentFragment();
    DocumentFragment frag = doc.createDocumentFragment();
    parser.parse(input, frag);
    res.appendChild(frag);

    try {
      while(true) {
        frag = doc.createDocumentFragment();
        parser.parse(input, frag);
        if (!frag.hasChildNodes()) break;
        if (LOG.isInfoEnabled()) {
          LOG.info(" - new frag, " + frag.getChildNodes().getLength() + " nodes.");
        }
        res.appendChild(frag);
View Full Code Here
123
124
125
126
127
128
129
130
131
132
133
    // OutputPropertiesFactory.getDefaultMethodProperties(Method.XML);
    // properties.put("encoding",_encoding);
    _serializer = SerializerFactory.getSerializer(properties);
    // serializer.setOutputStream(output);
    // _parser.setContentHandler(serializer.asContentHandler());
    viewStateParser = new DOMFragmentParser();
    // Set parser features
    try {
      viewStateParser
          .setProperty(
              "http://cyberneko.org/html/properties/names/elems",
View Full Code Here
TOP

Related Classes of org.cyberneko.html.parsers.DOMFragmentParser

  • org.ajax4jsf.webapp.nekko.NekkoParser
  • org.apache.kato.anttasks.xmlgen.XMLWriter
  • org.apache.myfaces.tobago.example.test.SeleniumScript
  • org.apache.nutch.parse.html.HtmlParser
  • org.apache.nutch.tika.TestDOMContentUtils
  • org.apache.nutch.tika.TestRobotsMetaProcessor
  • org.apache.shindig.gadgets.parse.nekohtml.NekoHtmlParser
  • org.apache.xerces.util.ErrorHandlerWrapper
  • org.apache.xerces.xni.parser.XMLErrorHandler
  • org.apache.xerces.xni.parser.XMLInputSource
    • Copyright © 2018 www.massapicom. All rights reserved.
    All source code are property of their respective owners. Java is a trademark of Sun Microsystems, Inc and owned by ORACLE Inc. Contact coftware#gmail.com.