Package org.apache.nutch.storage

Examples of org.apache.nutch.storage.ProtocolStatus

510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
                final FetchItemQueue fiq = fetchQueues.getFetchItemQueue(fit.queueID);
                fiq.crawlDelay = rules.getCrawlDelay();
              }
            }
            final ProtocolOutput output = protocol.getProtocolOutput(fit.url, fit.page);
            final ProtocolStatus status = output.getStatus();
            final Content content = output.getContent();
            // unblock queue
            fetchQueues.finishFetchItem(fit);

            context.getCounter("FetcherStatus", ProtocolStatusUtils.getName(status.getCode())).increment(1);

            int length = 0;
            if (content!=null && content.getContent()!=null) length= content.getContent().length;
            updateStatus(length);

            switch(status.getCode()) {

            case ProtocolStatusCodes.WOULDBLOCK:
              // retry ?
              fetchQueues.addFetchItem(fit);
              break;

            case ProtocolStatusCodes.SUCCESS:        // got a page
              output(fit, content, status, CrawlStatus.STATUS_FETCHED);
              break;

            case ProtocolStatusCodes.MOVED:         // redirect
            case ProtocolStatusCodes.TEMP_MOVED:
              byte code;
              boolean temp;
              if (status.getCode() == ProtocolStatusCodes.MOVED) {
                code = CrawlStatus.STATUS_REDIR_PERM;
                temp = false;
              } else {
                code = CrawlStatus.STATUS_REDIR_TEMP;
                temp = true;
              }
              final String newUrl = ProtocolStatusUtils.getMessage(status);
              handleRedirect(fit.url, newUrl, temp,  FetcherJob.PROTOCOL_REDIR, fit.page);
              output(fit, content, status, code);
              break;
            case ProtocolStatusCodes.EXCEPTION:
              logFetchFailure(fit.url, ProtocolStatusUtils.getMessage(status));
              /* FALLTHROUGH */
            case ProtocolStatusCodes.RETRY:          // retry
            case ProtocolStatusCodes.BLOCKED:
              output(fit, null, status, CrawlStatus.STATUS_RETRY);
              break;

            case ProtocolStatusCodes.GONE:           // gone
            case ProtocolStatusCodes.NOTFOUND:
            case ProtocolStatusCodes.ACCESS_DENIED:
            case ProtocolStatusCodes.ROBOTS_DENIED:
              output(fit, null, status, CrawlStatus.STATUS_GONE);
              break;

            case ProtocolStatusCodes.NOTMODIFIED:
              output(fit, null, status, CrawlStatus.STATUS_NOTMODIFIED);
              break;

            default:
              if (LOG.isWarnEnabled()) {
                LOG.warn("Unknown ProtocolStatus: " + status.getCode());
              }
              output(fit, null, status, CrawlStatus.STATUS_RETRY);
            }

          } catch (final Throwable t) {                 // unexpected exception
View Full Code Here
135
136
137
138
139
140
141
142
143
144
        } else { // convert to exception
          throw new FtpError(code);
        }
      }
    } catch (Exception e) {
      ProtocolStatus ps = ProtocolStatusUtils.makeStatus(
          ProtocolStatusCodes.EXCEPTION, e.toString());
      return new ProtocolOutput(null, ps);
    }
  }
View Full Code Here
198
199
200
201
202
203
204
205
206
207
208
                  Bytes.toStringBinary(entry.getValue()));
            }
          }
          res.put(f, simpleMeta);
        } else if ("protocolStatus".equals(f)) {
          ProtocolStatus ps = page.getProtocolStatus();
          res.put(f, ProtocolStatusUtils.toString(ps));
        } else if ("parseStatus".equals(f)) {
          ParseStatus ps = page.getParseStatus();
          res.put(f, ParseStatusUtils.toString(ps));
        } else if ("signature".equals(f)) {
View Full Code Here
189
190
191
192
193
194
195
196
197
198
199
                  Bytes.toStringBinary(entry.getValue().array()));
            }
          }
          res.put(f, simpleMeta);
        } else if ("protocolStatus".equals(f)) {
          ProtocolStatus ps = page.getProtocolStatus();
          res.put(f, ProtocolStatusUtils.toString(ps));
        } else if ("parseStatus".equals(f)) {
          ParseStatus ps = page.getParseStatus();
          res.put(f, ParseStatusUtils.toString(ps));
        } else if ("signature".equals(f)) {
View Full Code Here
74
75
76
77
78
79
80
81
82
83
      return "BLOCKED";
    return "UNKNOWN_CODE_" + code;
  }

  public static ProtocolStatus makeStatus(int code) {
    ProtocolStatus pstatus = new ProtocolStatus();
    pstatus.setCode(code);
    pstatus.setLastModified(0);
    return pstatus;
  }
View Full Code Here
81
82
83
84
85
86
87
88
89
    pstatus.setLastModified(0);
    return pstatus;
  }

  public static ProtocolStatus makeStatus(int code, String message) {
    ProtocolStatus pstatus = makeStatus(code);
    pstatus.addToArgs(new Utf8(message));
    return pstatus;
  }
View Full Code Here
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
                  final FetchItemQueue fiq = fetchQueues.getFetchItemQueue(fit.queueID);
                  fiq.crawlDelay = rules.getCrawlDelay();
                }
              }
              final ProtocolOutput output = protocol.getProtocolOutput(fit.url, fit.page);
              final ProtocolStatus status = output.getStatus();
              final Content content = output.getContent();
              // unblock queue
              fetchQueues.finishFetchItem(fit);

              context.getCounter("FetcherStatus", ProtocolStatusUtils.getName(status.getCode())).increment(1);

              int length = 0;
              if (content!=null && content.getContent()!=null) length= content.getContent().length;
              updateStatus(length);

              switch(status.getCode()) {

              case ProtocolStatusCodes.WOULDBLOCK:
                // retry ?
                fetchQueues.addFetchItem(fit);
                break;

              case ProtocolStatusCodes.SUCCESS:        // got a page
                output(fit, content, status, CrawlStatus.STATUS_FETCHED);
                break;

              case ProtocolStatusCodes.MOVED:         // redirect
              case ProtocolStatusCodes.TEMP_MOVED:
                byte code;
                boolean temp;
                if (status.getCode() == ProtocolStatusCodes.MOVED) {
                  code = CrawlStatus.STATUS_REDIR_PERM;
                  temp = false;
                } else {
                  code = CrawlStatus.STATUS_REDIR_TEMP;
                  temp = true;
                }
                output(fit, content, status, code);
                final String newUrl = ProtocolStatusUtils.getMessage(status);
                handleRedirect(fit.url, newUrl, temp,  FetcherJob.PROTOCOL_REDIR);
                redirecting = false;
                break;
              case ProtocolStatusCodes.EXCEPTION:
                logError(fit.url, ProtocolStatusUtils.getMessage(status));
                /* FALLTHROUGH */
              case ProtocolStatusCodes.RETRY:          // retry
              case ProtocolStatusCodes.BLOCKED:
                output(fit, null, status, CrawlStatus.STATUS_RETRY);
                break;

              case ProtocolStatusCodes.GONE:           // gone
              case ProtocolStatusCodes.NOTFOUND:
              case ProtocolStatusCodes.ACCESS_DENIED:
              case ProtocolStatusCodes.ROBOTS_DENIED:
                output(fit, null, status, CrawlStatus.STATUS_GONE);
                break;

              case ProtocolStatusCodes.NOTMODIFIED:
                output(fit, null, status, CrawlStatus.STATUS_NOTMODIFIED);
                break;

              default:
                if (LOG.isWarnEnabled()) {
                  LOG.warn("Unknown ProtocolStatus: " + status.getCode());
                }
                output(fit, null, status, CrawlStatus.STATUS_RETRY);
              }

              if (redirecting && redirectCount > maxRedirect) {
View Full Code Here
105
106
107
108
109
110
111
112
113
114
          throw new FileError(code);
        }
      }
    } catch (Exception e) {
      e.printStackTrace();
      ProtocolStatus ps = ProtocolStatusUtils.makeStatus(
          ProtocolStatusCodes.EXCEPTION, e.toString());
      return new ProtocolOutput(null, ps);
    }
  }
View Full Code Here
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
                  LOG.info("Crawl delay for queue: " + fit.queueID + " is set to " + fiq.crawlDelay + " as per robots.txt. url: " + fit.url);
                }
              }
            }
            final ProtocolOutput output = protocol.getProtocolOutput(fit.url, fit.page);
            final ProtocolStatus status = output.getStatus();
            final Content content = output.getContent();
            // unblock queue
            fetchQueues.finishFetchItem(fit);

            context.getCounter("FetcherStatus", ProtocolStatusUtils.getName(status.getCode())).increment(1);

            int length = 0;
            if (content!=null && content.getContent()!=null) length= content.getContent().length;
            updateStatus(length);

            switch(status.getCode()) {

            case ProtocolStatusCodes.WOULDBLOCK:
              // retry ?
              fetchQueues.addFetchItem(fit);
              break;

            case ProtocolStatusCodes.SUCCESS:        // got a page
              output(fit, content, status, CrawlStatus.STATUS_FETCHED);
              break;

            case ProtocolStatusCodes.MOVED:         // redirect
            case ProtocolStatusCodes.TEMP_MOVED:
              byte code;
              boolean temp;
              if (status.getCode() == ProtocolStatusCodes.MOVED) {
                code = CrawlStatus.STATUS_REDIR_PERM;
                temp = false;
              } else {
                code = CrawlStatus.STATUS_REDIR_TEMP;
                temp = true;
              }
              final String newUrl = ProtocolStatusUtils.getMessage(status);
              handleRedirect(fit.url, newUrl, temp,  FetcherJob.PROTOCOL_REDIR, fit.page);
              output(fit, content, status, code);
              break;
            case ProtocolStatusCodes.EXCEPTION:
              logFetchFailure(fit.url, ProtocolStatusUtils.getMessage(status));
              /* FALLTHROUGH */
            case ProtocolStatusCodes.RETRY:          // retry
            case ProtocolStatusCodes.BLOCKED:
              output(fit, null, status, CrawlStatus.STATUS_RETRY);
              break;

            case ProtocolStatusCodes.GONE:           // gone
            case ProtocolStatusCodes.NOTFOUND:
            case ProtocolStatusCodes.ACCESS_DENIED:
            case ProtocolStatusCodes.ROBOTS_DENIED:
              output(fit, null, status, CrawlStatus.STATUS_GONE);
              break;

            case ProtocolStatusCodes.NOTMODIFIED:
              output(fit, null, status, CrawlStatus.STATUS_NOTMODIFIED);
              break;

            default:
              if (LOG.isWarnEnabled()) {
                LOG.warn("Unknown ProtocolStatus: " + status.getCode());
              }
              output(fit, null, status, CrawlStatus.STATUS_RETRY);
            }

          } catch (final Throwable t) {                 // unexpected exception
View Full Code Here
198
199
200
201
202
203
204
205
206
207
208
                  Bytes.toStringBinary(entry.getValue().array()));
            }
          }
          res.put(f, simpleMeta);
        } else if ("protocolStatus".equals(f)) {
          ProtocolStatus ps = page.getProtocolStatus();
          res.put(f, ProtocolStatusUtils.toString(ps));
        } else if ("parseStatus".equals(f)) {
          ParseStatus ps = page.getParseStatus();
          res.put(f, ParseStatusUtils.toString(ps));
        } else if ("signature".equals(f)) {
View Full Code Here
TOP

Related Classes of org.apache.nutch.storage.ProtocolStatus

  • org.apache.gora.persistency.impl.StateManagerImpl
  • org.apache.nutch.api.DbReader$DbIterator
  • org.apache.nutch.fetcher.FetcherReducer$FetcherThread
  • org.apache.nutch.protocol.file.File
  • org.apache.nutch.protocol.ftp.Ftp
  • org.apache.nutch.protocol.ProtocolStatusUtils
  • org.apache.avro.AvroRuntimeException
    • 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.