Package org.apache.nutch.protocol

Examples of org.apache.nutch.protocol.Protocol

324
325
326
327
328
329
330
331
332
333
334
335
336
337
  /**
   * @param url
   * @return contents of url
   */
  private static String getUrlContent(String url) {
    Protocol protocol;
    try {
      protocol = ProtocolFactory.getProtocol(url);
      Content content = protocol.getProtocolOutput(url).getContent();
      String contentType = content.getContentType();
      Parser parser = ParserFactory.getParser(contentType, url);
      Parse parse = parser.getParse(content);
      System.out.println("text:" + parse.getText());
      return parse.getText();
View Full Code Here
48
49
50
51
52
53
54
55
56
57
58
59
 
  public Metadata parseMeta(String fileName, Configuration conf) {
    Metadata metadata = null;
    try {
      String urlString = "file:" + sampleDir + fileSeparator + fileName;    
      Protocol protocol = new ProtocolFactory(conf).getProtocol(urlString);
      Content content = protocol.getProtocolOutput(new Text(urlString),
          new CrawlDatum()).getContent();
      Parse parse = new ParseUtil(conf).parse(content).get(content.getUrl());
      metadata = parse.getData().getParseMeta();
    } catch (Exception e) {
      e.printStackTrace();
View Full Code Here
97
98
99
100
101
102
103
104
105
106
107
108
    if (LOG.isInfoEnabled()) {
      LOG.info("fetching: " + url);
    }

    ProtocolFactory factory = new ProtocolFactory(conf);
    Protocol protocol = factory.getProtocol(url);
    ProtocolOutput output = protocol.getProtocolOutput(new Text(url), new CrawlDatum());
   
    if (!output.getStatus().isSuccess()) {
      System.err.println("Fetch failed with protocol status: " + output.getStatus());
      return (-1);
    }
View Full Code Here
78
79
80
81
82
83
84
85
86
87
88
89
90
91
    }

    IndexingFilters indexers = new IndexingFilters(conf);

    ProtocolFactory factory = new ProtocolFactory(conf);
    Protocol protocol = factory.getProtocol(url);
    CrawlDatum datum = new CrawlDatum();

    ProtocolOutput output = protocol.getProtocolOutput(new Text(url), datum);
   
    if (!output.getStatus().isSuccess()) {
      System.out.println("Fetch failed with protocol status: " + output.getStatus());
      return 0;
    }
View Full Code Here
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
  protected void tearDown() {}

  public void testIt() throws ProtocolException, ParseException {
    String urlString;
    Protocol protocol;
    Content content;
    Parse parse;
    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 Text(urlString), new CrawlDatum()).getContent();

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

      String text = parse.getText().replaceAll("[ \t\r\n]+", " ").trim();
      assertTrue(sampleTexts[i].equals(text));
View Full Code Here
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
  }

  public void testIt() throws ProtocolException, ParseException {

    String urlString;
    Protocol protocol;
    Content content;
    Parse parse;

    Configuration conf = NutchConfiguration.create();
    urlString = "file:" + sampleDir + fileSeparator + rtfFile;
    protocol = new ProtocolFactory(conf).getProtocol(urlString);
    content = protocol.getProtocolOutput(new UTF8(urlString), new CrawlDatum())
                      .getContent();
    parse = new ParseUtil(conf).parseByExtensionId("parse-rtf", content);
    String text = parse.getText();
    assertEquals("The quick brown fox jumps over the lazy dog", text.trim());
View Full Code Here
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
  protected void tearDown() {}

  public void testIt() throws ProtocolException, ParseException {
    String urlString;
    Protocol protocol;
    Content content;
    Parse parse;

    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-msword", content);

      assertTrue(parse.getText().startsWith(expectedText));
    }
  }
View Full Code Here
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
  protected void tearDown() {}

  public void testIt() throws ProtocolException, ParseException {
    String urlString;
    Protocol protocol;
    Content content;
    Parse parse;
    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();
      assertTrue(sampleTexts[i].equals(text));
View Full Code Here
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
  protected void tearDown() {}

  public void testIt() throws ProtocolException, ParseException {
    String urlString;
    Protocol protocol;
    Content content;
    Parse parse;

    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
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
  protected void tearDown() {}

  public void testIt() throws ProtocolException, ParseException {
    String urlString;
    Protocol protocol;
    Content content;
    Parse parse;

    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
TOP

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

  • org.apache.nutch.analysis.lang.LanguageIdentifier
  • 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
  • org.apache.nutch.parse.msword.TestMSWordParser
    • 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.