Examples of MimeUtil

Examples of org.apache.nutch.util.MimeUtil

80
81
82
83
84
85
86
87
    this.url = url;
    this.base = base;
    this.content = content;
    this.metadata = metadata;

    this.mimeTypes = new MimeUtil(conf);
    this.contentType = getContentType(contentType, url, content);
  }
View Full Code Here

Examples of org.apache.nutch.util.MimeUtil

94
95
96
97
98
99
100
101
102
103
104
    this.orig = url.toString();
    this.base = url.toString();
    this.file = file;
    this.conf = conf;
   
    MIME = new MimeUtil(conf);

    if (!"file".equals(url.getProtocol()))
      throw new FileException("Not a file url:" + url);

    if (File.LOG.isTraceEnabled()) {
View Full Code Here

Examples of org.apache.nutch.util.MimeUtil

55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
    public void testIt() throws ProtocolException, ParseException, IOException {

  String urlString;
  Parse parse;
  Configuration conf = NutchConfiguration.create();
  MimeUtil mimeutil = new MimeUtil(conf);

  urlString = "file:" + sampleDir + fileSeparator + rtfFile;

  File file = new File(sampleDir + fileSeparator + rtfFile);
  byte[] bytes = new byte[(int) file.length()];
  DataInputStream in = new DataInputStream(new FileInputStream(file));
  in.readFully(bytes);
  in.close();

  WebPage page = new WebPage();
  page.setBaseUrl(new Utf8(urlString));
  page.setContent(ByteBuffer.wrap(bytes));
  String mtype = mimeutil.getMimeType(file);
  page.setContentType(new Utf8(mtype));

  parse = new ParseUtil(conf).parse(urlString, page);

  String title = parse.getTitle();
View Full Code Here

Examples of org.apache.nutch.util.MimeUtil

56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
    @Test
    public void testIt() throws ProtocolException, ParseException, IOException {
  String urlString;
  Parse parse;
  Configuration conf = NutchConfiguration.create();
  MimeUtil mimeutil = new MimeUtil(conf);

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

      File file = new File(sampleDir + fileSeparator + sampleFiles[i]);
      byte[] bytes = new byte[(int) file.length()];
      DataInputStream in = new DataInputStream(new FileInputStream(file));
      in.readFully(bytes);
      in.close();

      WebPage page = new WebPage();
      page.setBaseUrl(new Utf8(urlString));
      page.setContent(ByteBuffer.wrap(bytes));
      String mtype = mimeutil.getMimeType(file);
      page.setContentType(new Utf8(mtype));

      parse = new ParseUtil(conf).parse(urlString, page);

      int index = parse.getText().indexOf(expectedText);
View Full Code Here

Examples of org.apache.nutch.util.MimeUtil

71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
  public void testIt()throws ProtocolException, ParseException, IOException {
    String urlString;
    Parse parse;

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

      File file = new File(sampleDir + fileSeparator + sampleFiles[i]);
      byte[] bytes = new byte[(int) file.length()];
      DataInputStream in = new DataInputStream(new FileInputStream(file));
      in.readFully(bytes);
      in.close();

      WebPage page = new WebPage();
      page.setBaseUrl(new Utf8(urlString));
      page.setContent(ByteBuffer.wrap(bytes));
      String mtype = mimeutil.getMimeType(file);
      page.setContentType(new Utf8(mtype));

      parse = new ParseUtil(conf).parse(urlString, page);

      // check that there are 2 outlinks:
View Full Code Here

Examples of org.apache.nutch.util.MimeUtil

240
241
242
243
244
245
246
247
248
249
250
251
    // TikaParser parser = new TikaParser();
    // parser.setConf(conf);
    WebPage page = new WebPage();
    page.setBaseUrl(new Utf8(url));
    page.setContent(ByteBuffer.wrap(bytes));
    MimeUtil mimeutil = new MimeUtil(conf);
    String mtype = mimeutil.getMimeType(file);
    page.setContentType(new Utf8(mtype));
    // Parse parse = parser.getParse(url, page);

    Parse parse = new ParseUtil(conf).parse(url, page);
View Full Code Here

Examples of org.apache.nutch.util.MimeUtil

64
65
66
67
68
69
70
71
72
73
74
75
    Configuration conf = NutchConfiguration.create();

    WebPage page = new WebPage();
    page.setBaseUrl(new Utf8(url));
    page.setContent(ByteBuffer.wrap(bytes));
    MimeUtil mimeutil = new MimeUtil(conf);
    String mtype = mimeutil.getMimeType(file);
    page.setContentType(new Utf8(mtype));

    new ParseUtil(conf).parse(url, page);

    ByteBuffer bb = page.getFromMetadata(new Utf8("License-Url"));
View Full Code Here

Examples of org.apache.nutch.util.MimeUtil

74
75
76
77
78
79
80
81
82
83
84
85
  Parse parse;
  WebPage page = new WebPage();
  page.setBaseUrl(new Utf8("file:"+urlString));
  page.setContent(ByteBuffer.wrap(bytes));
  // set the content type?
  MimeUtil mimeutil = new MimeUtil(conf);
  String mtype = mimeutil.getMimeType(file);
  page.setContentType(new Utf8(mtype));
   
  parse = new ParseUtil(conf).parse("file:"+urlString, page);
  return parse.getText();
    }
View Full Code Here

Examples of org.apache.nutch.util.MimeUtil

58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
  @Test
  public void testIt() throws ProtocolException, ParseException, IOException {
    String urlString;
    Parse parse;
    Configuration conf = NutchConfiguration.create();
    MimeUtil mimeutil = new MimeUtil(conf);
 
    try {
      // read the test string
      FileInputStream fis = new FileInputStream(sampleDir + fileSeparator
        + sampleText);
      StringBuffer sb = new StringBuffer();
      int len = 0;
      InputStreamReader isr = new InputStreamReader(fis, "UTF-8");
      char[] buf = new char[1024];
      while ((len = isr.read(buf)) > 0) {
        sb.append(buf, 0, len);
      }
      isr.close();
      expectedText = sb.toString();
      // normalize space
      expectedText = expectedText.replaceAll("[ \t\r\n]+", " ");
    } catch (Exception e) {
      e.printStackTrace();
    }

    System.out.println("Expected : " + expectedText);

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

      if (sampleFiles[i].startsWith("ootest") == false)
      continue;

      File file = new File(sampleDir + fileSeparator + sampleFiles[i]);
      byte[] bytes = new byte[(int) file.length()];
      DataInputStream in = new DataInputStream(new FileInputStream(file));
      in.readFully(bytes);
      in.close();

      WebPage page = new WebPage();
      page.setBaseUrl(new Utf8(urlString));
      page.setContent(ByteBuffer.wrap(bytes));
      String mtype = mimeutil.getMimeType(file);
      page.setContentType(new Utf8(mtype));

      parse = new ParseUtil(conf).parse(urlString, page);

      String text = parse.getText().replaceAll("[ \t\r\n]+", " ").trim();
View Full Code Here

Examples of org.apache.nutch.util.MimeUtil

75
76
77
78
79
80
81
82
83
84
85
86
    dip.close();
   
    WebPage page = new WebPage();
    page.setBaseUrl(new Utf8(urlString));
    page.setContent(ByteBuffer.wrap(bytes));
    MimeUtil mutil = new MimeUtil(conf);
    String mime = mutil.getMimeType(file);
    page.setContentType(new Utf8(mime));
 
    parse = new ParseUtil(conf).parse(urlString, page);
    return parse.getOutlinks();
  }
View Full Code Here
TOP
Copyright © 2018 www.massapi.com. 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.