Package org.apache.nutch.protocol

Examples of org.apache.nutch.protocol.ProtocolFactory

53
54
55
56
57
58
59
60
61
62
63
    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 Text(urlString), new CrawlDatum()).getContent();
      parse = new ParseUtil(conf).parseByExtensionId("parse-tika", content).get(content.getUrl());

      Assert.assertEquals("121", parse.getData().getMeta("width"));
      Assert.assertEquals("48", parse.getData().getMeta("height"));
View Full Code Here
60
61
62
63
64
65
66
67
68
69
70
    Content content;
    Parse parse;

    Configuration conf = NutchConfiguration.create();
    urlString = "file:" + sampleDir + fileSeparator + rtfFile;
    protocol = new ProtocolFactory(conf).getProtocol(urlString);
    content = protocol.getProtocolOutput(new Text(urlString),
        new CrawlDatum()).getContent();
    parse = new ParseUtil(conf).parseByExtensionId("parse-tika", content)
        .get(content.getUrl());
    String text = parse.getText();
View Full Code Here
60
61
62
63
64
65
66
67
68
69
70
    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 Text(urlString), new CrawlDatum()).getContent();
      parse = new ParseUtil(conf).parseByExtensionId("parse-tika", content).get(content.getUrl());

      int index = parse.getText().indexOf(expectedText);
      Assert.assertTrue(index > 0);
View Full Code Here
75
76
77
78
79
80
81
82
83
84
85
    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).parseByExtensionId("parse-tika",
          content).get(content.getUrl());
View Full Code Here
124
125
126
127
128
129
130
131
132
133
134
135
      if (value == null)
        value = "";
      cd.getMetaData().put(new Text(key), new Text(value));
    }

    ProtocolFactory factory = new ProtocolFactory(conf);
    Protocol protocol = factory.getProtocol(url);
    Text turl = new Text(url);
    ProtocolOutput output = protocol.getProtocolOutput(turl, cd);
   
    if (!output.getStatus().isSuccess()) {
      System.err.println("Fetch failed with protocol status: " + output.getStatus());
View Full Code Here
72
73
74
75
76
77
78
79
80
81
82
83
    }

    if (LOG.isInfoEnabled()) { LOG.info("fetching: "+url); }

    Configuration conf = NutchConfiguration.create();
    ProtocolFactory factory = new ProtocolFactory(conf);
    Protocol protocol = factory.getProtocol(url);
    Content content = protocol.getProtocolOutput(new Text(url), new CrawlDatum()).getContent();

    if (force) {
      content.setContentType(contentType);
    } else {
View Full Code Here
66
67
68
69
70
71
72
73
74
75
76
    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 Text(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
214
215
216
217
218
219
220
221
222
223
224
    //LOG.setLevel(Level.FINE);
    String url = args[0];
    Configuration conf = NutchConfiguration.create();
    RSSParser parser = new RSSParser();
    parser.setConf(conf);
    Protocol protocol = new ProtocolFactory(conf).getProtocol(url);
    Content content = protocol.getProtocolOutput(new Text(url), new CrawlDatum()).getContent();
    Parse parse = parser.getParse(content);
    System.out.println("data: "+ parse.getData());
    System.out.println("text: "+parse.getText());
  }
View Full Code Here
66
67
68
69
70
71
72
73
74
75
76
    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).parseByExtensionId("parse-zip",content);
      assertTrue(parse.getText().equals(expectedText));
    }
  }
View Full Code Here
80
81
82
83
84
85
86
87
88
    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 Text(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.