Package org.apache.nutch.protocol

Examples of org.apache.nutch.protocol.Content

216
217
218
219
220
221
222
223
224
225
226
        try {
            Configuration conf = NutchConfiguration.create();

      byte[] raw = getRawBytes(new File(file));
            Metadata meta = new Metadata();
            Content content = new Content(file, file, raw, "application/pdf", meta, conf);

      //Protocol protocol = new ProtocolFactory(conf).getProtocol(file);
      //Content content = protocol.getProtocolOutput(new Text(file), new CrawlDatum()).getContent();
      //Parse parse = new ParseUtil(conf).parseByExtensionId("parse-pdf", content);
View Full Code Here
334
335
336
337
338
339
340
341
342
343
344
   */
  public void reduce(WritableComparable key, Iterator values, OutputCollector output, Reporter reporter) throws IOException {
    CrawlDatum lastG = null;
    CrawlDatum lastF = null;
    CrawlDatum lastSig = null;
    Content lastC = null;
    ParseData lastPD = null;
    ParseText lastPT = null;
    String lastGname = null;
    String lastFname = null;
    String lastSigname = null;
View Full Code Here
216
217
218
219
220
221
222
223
224
225
226
        try {
            Configuration conf = NutchConfiguration.create();

      byte[] raw = getRawBytes(new File(file));
            Metadata meta = new Metadata();
            Content content = new Content(file, file, raw, "application/pdf", meta, conf);

      //Protocol protocol = new ProtocolFactory(conf).getProtocol(file);
      //Content content = protocol.getProtocolOutput(new Text(file), new CrawlDatum()).getContent();
      //Parse parse = new ParseUtil(conf).parseByExtensionId("parse-pdf", content);
View Full Code Here
154
155
156
157
158
159
160
161
162
163
164
        try {
            Configuration conf = NutchConfiguration.create();

            byte[] raw = getRawBytes(new File(file));
            Metadata meta = new Metadata();
            Content content = new Content(file, file, raw, "trec/plain", meta, conf);

            //Protocol protocol = new ProtocolFactory(conf).getProtocol(file);                                                                                                                                                             
            //Content content = protocol.getProtocolOutput(new Text(file), new CrawlDatum()).getContent();                                                                                                                                 
            //Parse parse = new ParseUtil(conf).parseByExtensionId("parse-pdf", content);                                                                                                                                                  
View Full Code Here
254
255
256
257
258
259
260
261
262
263
264
      DataInputStream in = new DataInputStream(new FileInputStream(file));
     
      try
      {
        in.readFully(bytes);
        Parse parse = parser.getParse(new Content(url, url, bytes,
          "application/pdf", new Metadata(), conf));
        System.out.println(parse.getData().getTitle());
      }
      finally
      {
View Full Code Here
154
155
156
157
158
159
160
161
162
163
164
        try {
            Configuration conf = NutchConfiguration.create();

            byte[] raw = getRawBytes(new File(file));
            Metadata meta = new Metadata();
            Content content = new Content(file, file, raw, "trec/plain", meta, conf);

            //Protocol protocol = new ProtocolFactory(conf).getProtocol(file);                                                                                                                                                             
            //Content content = protocol.getProtocolOutput(new Text(file), new CrawlDatum()).getContent();                                                                                                                                 
            //Parse parse = new ParseUtil(conf).parseByExtensionId("parse-pdf", content);                                                                                                                                                  
View Full Code Here
154
155
156
157
158
159
160
161
162
163
164
        try {
            Configuration conf = NutchConfiguration.create();

            byte[] raw = getRawBytes(new File(file));
            Metadata meta = new Metadata();
            Content content = new Content(file, file, raw, "trec/plain", meta, conf);

            //Protocol protocol = new ProtocolFactory(conf).getProtocol(file);                                                                                                                                                             
            //Content content = protocol.getProtocolOutput(new Text(file), new CrawlDatum()).getContent();                                                                                                                                 
            //Parse parse = new ParseUtil(conf).parseByExtensionId("parse-pdf", content);                                                                                                                                                  
View Full Code Here
124
125
126
127
128
129
130
131
132
133
    String file = args[0];
    byte[] raw = getRawBytes(new File(file));
       
    Metadata meta = new Metadata();
    meta.set(Response.CONTENT_LENGTH, "" + raw.length);
    Content content = new Content(file, file, raw, mime, meta,
                                  NutchConfiguration.create());

    System.out.println(parser.getParse(content).getText());
  }
View Full Code Here
446
447
448
449
450
451
452
453
454
455
456
   
    // Score at this stage is 1.0f.
    metaData.set(Nutch.SCORE_KEY, Float.toString(datum.getScore()));

    final long startTime = System.currentTimeMillis();
    final Content content = new Content(url, url, contentBytes, mimetype,
      metaData, getConf());
    datum.setFetchTime(Nutchwax.getDate(arcData.getDate()));

    MapWritable mw = datum.getMetaData();
   
View Full Code Here
43
44
45
46
47
48
49
50
51
52
53
54
55
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
81
82
83
84
85
86
    // first disable auto detection
    conf.setInt(EncodingDetector.MIN_CONFIDENCE_KEY, -1);

    Metadata metadata = new Metadata();
    EncodingDetector detector;
    Content content;
    String encoding;

    content = new Content("http://www.example.com", "http://www.example.com/",
        contentInOctets, "text/plain", metadata, conf);
    detector = new EncodingDetector(conf);
    detector.autoDetectClues(content, true);
    encoding = detector.guessEncoding(content, "windows-1252");
    // no information is available, so it should return default encoding
    Assert.assertEquals("windows-1252", encoding.toLowerCase());

    metadata.clear();
    metadata.set(Response.CONTENT_TYPE, "text/plain; charset=UTF-16");
    content = new Content("http://www.example.com", "http://www.example.com/",
        contentInOctets, "text/plain", metadata, conf);
    detector = new EncodingDetector(conf);
    detector.autoDetectClues(content, true);
    encoding = detector.guessEncoding(content, "windows-1252");
    Assert.assertEquals("utf-16", encoding.toLowerCase());

    metadata.clear();
    content = new Content("http://www.example.com", "http://www.example.com/",
        contentInOctets, "text/plain", metadata, conf);
    detector = new EncodingDetector(conf);
    detector.autoDetectClues(content, true);
    detector.addClue("windows-1254", "sniffed");
    encoding = detector.guessEncoding(content, "windows-1252");
    Assert.assertEquals("windows-1254", encoding.toLowerCase());

    // enable autodetection
    conf.setInt(EncodingDetector.MIN_CONFIDENCE_KEY, 50);
    metadata.clear();
    metadata.set(Response.CONTENT_TYPE, "text/plain; charset=UTF-16");
    content = new Content("http://www.example.com", "http://www.example.com/",
        contentInOctets, "text/plain", metadata, conf);
    detector = new EncodingDetector(conf);
    detector.autoDetectClues(content, true);
    detector.addClue("utf-32", "sniffed");
    encoding = detector.guessEncoding(content, "windows-1252");
View Full Code Here
TOP

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

  • com.atlantbh.nutch.filter.xpath.XPathHtmlParserFilterTest
  • com.flaptor.hounder.crawler.Nutch9Fetcher$NutchSegment$SegmentIterator
  • nise.ExtractNutch
  • org.apache.commons.cli.Options
  • org.apache.hadoop.conf.Configuration
  • org.apache.hadoop.fs.FileSystem
  • org.apache.hadoop.fs.Path
  • org.apache.hadoop.util.GenericOptionsParser
  • org.apache.nutch.analysis.lang.LanguageIdentifier
  • org.apache.nutch.analysis.lang.TestHTMLLanguageParser
    • 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.