Method map

src/java/org/apache/nutch/parse/ParserJob.java:118–160

(String key, WebPage page, Context context)

    }

    @Override
    public void map(String key, WebPage page, Context context)
        throws IOException, InterruptedException {
      String unreverseKey = TableUtil.unreverseUrl(key);
      if (batchId.equals(REPARSE)) {
        LOG.debug("Reparsing " + unreverseKey);
      } else {
        if (Mark.FETCH_MARK.checkMark(page) == null) {
          if (LOG.isDebugEnabled()) {
            LOG.debug("Skipping {}, not fetched yet", unreverseKey);
          }
          return;
        }
        if (shouldResume && Mark.PARSE_MARK.checkMark(page) != null) {
          if (force) {
            LOG.info("Forced parsing " + unreverseKey + "; already parsed");
          } else {
            LOG.info("Skipping " + unreverseKey + "; already parsed");
            return;
          }
        } else {
          LOG.info("Parsing " + unreverseKey);
        }
      }

      if (skipTruncated && isTruncated(unreverseKey, page)) {
        return;
      }

      if (sitemap && URLFilters.isSitemap(page)) {
        LOG.info("Parsing for sitemap"); //TODO this log should be top line
        parseUtil.processSitemapParse(unreverseKey, page, context);
      } else {
        parseUtil.process(unreverseKey, page);
      }
      ParseStatus pstatus = page.getParseStatus();
      if (pstatus != null) {
        context.getCounter("ParserStatus",
            ParseStatusCodes.majorCodes[pstatus.getMajorCode()]).increment(1);
      }

      context.write(key, page);
    }
  }

  public ParserJob() {