Package de.anomic.search

Examples of de.anomic.search.Segment.termIndex()


                }

                // prepare index
                ReferenceContainer<WordReference> index;
                final long starttime = System.currentTimeMillis();
                index = segment.termIndex().get(keyhash, null);
                // built urlCache
                final Iterator<WordReference> urlIter = index.entries();
                final TreeMap<byte[], URIMetadataRow> knownURLs = new TreeMap<byte[], URIMetadataRow>(Base64Order.enhancedCoder);
                final HandleSet unknownURLEntries = new HandleSet(WordReferenceRow.urlEntryRow.primaryKeyLength, WordReferenceRow.urlEntryRow.objectOrder, index.size());
                Reference iEntry;
View Full Code Here


                Log.logException(e);
            }

            // generate list
            if (post.containsKey("keyhashsimilar")) try {
                final Iterator<ReferenceContainer<WordReference>> containerIt = segment.termIndex().references(keyhash, true, 256, false).iterator();
                    ReferenceContainer<WordReference> container;
                    i = 0;
                    int rows = 0, cols = 0;
                    prop.put("keyhashsimilar", "1");
                    while (containerIt.hasNext() && i < 256) {
View Full Code Here

                        pw.close();
                    } catch (final IOException e) {
                    }
                }
                try {
                    segment.termIndex().remove(keyhash, urlHashes);
                } catch (final IOException e) {
                    Log.logException(e);
                }
            }
View Full Code Here

            if (prop.getInt("searchresult", 0) == 3) listHosts(prop, keyhash, sb);
        }


        // insert constants
        prop.putNum("wcount", segment.termIndex().sizesMax());
        // return rewrite properties
        return prop;
    }

    public static void genURLList(final serverObjects prop, final byte[] keyhash, final String keystring, final RankingProcess ranked, final Bitfield flags, final int maxlines) {
View Full Code Here

            final Segment wordIndex = new Segment(
                    log,
                    new File(new File(indexPrimaryRoot, "freeworld"), "TEXT"),
                    10000,
                    (long) Integer.MAX_VALUE, false, false);
            final Iterator<ReferenceContainer<WordReference>> indexContainerIterator = wordIndex.termIndex().references("AAAAAAAAAAAA".getBytes(), false, false);

            long urlCounter = 0, wordCounter = 0;
            long wordChunkStart = System.currentTimeMillis(), wordChunkEnd = 0;
            String wordChunkStartHash = "AAAAAAAAAAAA", wordChunkEndHash;
View Full Code Here

                WordIndex = new Segment(
                        log,
                        new File(new File(indexPrimaryRoot, "freeworld"), "TEXT"),
                        10000,
                        (long) Integer.MAX_VALUE, false, false);
                indexContainerIterator = WordIndex.termIndex().references(wordChunkStartHash.getBytes(), false, false);
            }
            int counter = 0;
            ReferenceContainer<WordReference> container = null;
            if (format.equals("zip")) {
                log.logInfo("Writing Hashlist to ZIP-file: " + targetName + ".zip");
View Full Code Here

       
        yacySeed initiator;
       
        // index size
        prop.putNum("urlpublictextSize", segment.urlMetadata().size());
        prop.putNum("rwipublictextSize", segment.termIndex().sizesMax());

        // loader queue
        prop.putNum("loaderSize", sb.crawlQueues.workerSize());       
        prop.putNum("loaderMax", sb.getConfigLong(SwitchboardConstants.CRAWLER_THREADS_ACTIVE_MAX, 10));
        if (sb.crawlQueues.workerSize() == 0) {
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.