Package org.apache.nutch.protocol

Examples of org.apache.nutch.protocol.ProtocolFactory

82
83
84
85
86
87
88
89
90
91
92
    Configuration conf = NutchConfiguration.create();

    for (int i = 0; i < sampleFiles.length; i++) {
      urlString = "file:" + sampleDir + fileSeparator + sampleFiles[i];

      protocol = new ProtocolFactory(conf).getProtocol(urlString);
      content = protocol.getProtocolOutput(new UTF8(urlString), new CrawlDatum()).getContent();

      parse = new ParseUtil(conf).parse(content);

      String text = parse.getText().replaceAll("[ \t\r\n]+", " ").trim();
View Full Code Here
65
66
67
68
69
70
71
72
73
74
75
    for (int i=0; i<sampleFiles.length; i++) {
      urlString = "file:" + sampleDir + fileSeparator + sampleFiles[i];

      Configuration conf = NutchConfiguration.create();
      protocol = new ProtocolFactory(conf).getProtocol(urlString);
      content = protocol.getProtocolOutput(new UTF8(urlString), new CrawlDatum()).getContent();
      parse = new ParseUtil(conf).parseByExtensionId("parse-pdf", content);

      int index = parse.getText().indexOf(expectedText);
      assertTrue(index > 0);
View Full Code Here
65
66
67
68
69
70
71
72
73
74
75
    Configuration conf = NutchConfiguration.create();
    for (int i = 0; i < sampleFiles.length; i++) {
      urlString = "file:" + sampleDir + fileSeparator + sampleFiles[i];

      protocol = new ProtocolFactory(conf).getProtocol(urlString);
      content = protocol.getProtocolOutput(new UTF8(urlString), new CrawlDatum()).getContent();
      parse = new ParseUtil(conf).parseByExtensionId("parse-zip",content);
      assertTrue(parse.getText().equals(expectedText));
    }
  }
View Full Code Here
341
342
343
344
345
346
347
348
349
350
351
   * @return contents of url
   */
  private static String getUrlContent(String url, Configuration conf) {
    Protocol protocol;
    try {
      protocol = new ProtocolFactory(conf).getProtocol(url);
      Content content = protocol.getProtocolOutput(new UTF8(url), new CrawlDatum()).getContent();
      Parse parse = new ParseUtil(conf).parse(content);
      System.out.println("text:" + parse.getText());
      return parse.getText();

View Full Code Here
65
66
67
68
69
70
71
72
73
74
75
    Content content;
    Parse parse;

    Configuration conf = NutchConfiguration.create();
    urlString = "file:" + sampleDir + fileSeparator + id3v2;
    protocol = new ProtocolFactory(conf).getProtocol(urlString);
    content = protocol.getProtocolOutput(new UTF8(urlString), new CrawlDatum())
                      .getContent();
    parse = new ParseUtil(conf).parseByExtensionId("parse-mp3", content);
    Metadata metadata = parse.getData().getParseMeta();
    assertEquals("postgresql comment id3v2", metadata.get("COMM-Text"));
View Full Code Here
97
98
99
100
101
102
103
104
105
106
107
    Content content;
    Parse parse;

    Configuration conf = NutchConfiguration.create();
    urlString = "file:" + sampleDir + fileSeparator + id3v1;
    protocol = new ProtocolFactory(conf).getProtocol(urlString);
    content = protocol.getProtocolOutput(new UTF8(urlString), new CrawlDatum())
                      .getContent();
    parse = new ParseUtil(conf).parseByExtensionId("parse-mp3", content);

    Metadata metadata = parse.getData().getParseMeta();
View Full Code Here
124
125
126
127
128
129
130
131
132
133
134
    Content content;
    Parse parse;

    Configuration conf = NutchConfiguration.create();
    urlString = "file:" + sampleDir + fileSeparator + none;
    protocol = new ProtocolFactory(conf).getProtocol(urlString);
    content = protocol.getProtocolOutput(new UTF8(urlString), new CrawlDatum())
                      .getContent();
    parse = new ParseUtil(conf).parseByExtensionId("parse-mp3", content);
//    Metadata metadata = parse.getData().getParseMeta();
    if (parse.getData().getStatus().isSuccess()) {
View Full Code Here
83
84
85
86
87
88
89
90
91
92
93
        Configuration conf = NutchConfiguration.create();
        for (int i = 0; i < sampleFiles.length; i++) {
            urlString = "file:" + sampleDir + fileSeparator + sampleFiles[i];

            protocol = new ProtocolFactory(conf).getProtocol(urlString);
            content = protocol.getProtocolOutput(new UTF8(urlString), new CrawlDatum()).getContent();
            parse = new ParseUtil(conf).parseByExtensionId("parse-rss", content);

            //check that there are 3 outlinks:
            //http://test.channel.com
View Full Code Here
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
    Content content;
    Parse parse;

    Configuration conf = NutchConfiguration.create();
    ParseUtil parser = new ParseUtil(conf);
    ProtocolFactory factory = new ProtocolFactory(conf);
    for (int i = 0; i < sampleFiles.length; i++) {
      urlString = "file:" + sampleDir + fileSeparator + sampleFiles[i];

      protocol = factory.getProtocol(urlString);
      content = protocol.getProtocolOutput(new UTF8(urlString),
                                           new CrawlDatum()).getContent();
      parse = parser.parseByExtensionId("parse-msexcel", content);

      assertTrue(parse.getText().equals(expectedText));
View Full Code Here
79
80
81
82
83
84
85
86
87
    FileOutputStream fos = new FileOutputStream(tempFile);
    fos.write(expectedText.getBytes());
    fos.close();

    // get nutch content
    Protocol protocol = new ProtocolFactory(NutchConfiguration.create()).getProtocol(urlString);
    content = protocol.getProtocolOutput(new UTF8(urlString), new CrawlDatum()).getContent();
    protocol = null;
  }
View Full Code Here
TOP

Related Classes of org.apache.nutch.protocol.ProtocolFactory

  • org.apache.nutch.analysis.lang.LanguageIdentifier
  • org.apache.nutch.fetcher.FetcherJob
  • org.apache.nutch.fetcher.FetcherReducer$FetcherThread
  • org.apache.nutch.indexer.IndexingFiltersChecker
  • org.apache.nutch.parse.ext.TestExtParser
  • org.apache.nutch.parse.feed.TestFeedParser
  • org.apache.nutch.parse.html.TestMetatagParser
  • org.apache.nutch.parse.metatags.TestMetatagParser
  • org.apache.nutch.parse.mp3.TestMP3Parser
  • org.apache.nutch.parse.msexcel.TestMSExcelParser
    • 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.