Package bixo.datum

Examples of bixo.datum.ParsedDatum

187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
            Callable<ParsedDatum> c = new TikaCallable(_parser, _contentExtractor, _linkExtractor, is, metadata, isExtractLanguage(), _parseContext);
            FutureTask<ParsedDatum> task = new FutureTask<ParsedDatum>(c);
            Thread t = new Thread(task);
            t.start();
           
            ParsedDatum result;
            try {
                result = task.get(getParserPolicy().getMaxParseDuration(), TimeUnit.MILLISECONDS);
            } catch (TimeoutException e) {
                task.cancel(true);
                t.interrupt();
                throw e;
            } finally {
                t = null;
            }
           
            // TODO KKr Should there be a BaseParser to take care of copying
            // these two fields?
            result.setHostAddress(fetchedDatum.getHostAddress());
            result.setPayload(fetchedDatum.getPayload());
            return result;
        } finally {
            IoUtils.safeClose(is);
        }
    }
View Full Code Here
113
114
115
116
117
118
119
        super.prepare(process, opCall);
       
        _reader = new SAXReader(new Parser());
        _reader.setXMLFilter(new DowngradeXmlFilter(_removeNamespaces));
        _reader.setEncoding("UTF-8");
        _input = new ParsedDatum();
    }
View Full Code Here
121
122
123
124
125
126
127
128
129
130
131
                _parseContext = makeParseContext();
            }
            _parser.parse(_input, teeContentHandler, _metadata, _parseContext);
           
            String lang = _extractLanguage ? detectLanguage(_metadata, profilingHandler) : "";
            return new ParsedDatum(_metadata.get(Metadata.RESOURCE_NAME_KEY), null, _contentExtractor.getContent(), lang,
                            _metadata.get(Metadata.TITLE),
                            _linkExtractor.getLinks(), makeMap(_metadata));
        } catch (Exception e) {
            // Generic exception that's OK to re-throw
            throw e;
View Full Code Here
75
76
77
78
79
80
81
82
83
84
85
86
87
        @Override
        public void operate(FlowProcess flowProcess, FunctionCall<NullContext> functionCall) {
            FetchedDatum fetchedDatum = new FetchedDatum(functionCall.getArguments());
           
            try {
                ParsedDatum parseResult = _parser.parse(fetchedDatum);
                _flowProcess.increment(ParserCounters.DOCUMENTS_PARSED, 1);
                functionCall.getOutputCollector().add(BixoPlatform.clone(parseResult.getTuple(), flowProcess));
            } catch (Exception e) {
                LOGGER.warn("Error processing " + fetchedDatum.getUrl(), e);
                _flowProcess.increment(ParserCounters.DOCUMENTS_FAILED, 1);
                // TODO KKr - don't lose datums for documents that couldn't be parsed
            }
View Full Code Here
TOP

Related Classes of bixo.datum.ParsedDatum

  • bixo.examples.crawl.CreateUrlDatumFromOutlinksFunction
  • bixo.examples.crawl.CreateWritableSeqFileData
  • bixo.parser.DOMParser
  • bixo.parser.FakeParser
  • bixo.parser.SimpleParser
  • bixo.parser.SimpleParserTest
  • bixo.parser.TikaCallable
  • bixo.pipes.ParsePipe$ParseFunction
  • bixo.tools.FetchAndParseTool
  • cascading.tuple.Fields
    • 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.