- rotate merged indexes after merging

see: http://www.yacy-forum.de/viewtopic.php?t=1717
- fix -rwihashlist to correctly shutdown



git-svn-id: https://svn.berlios.de/svnroot/repos/yacy/trunk@1336 6c8d7289-2bf4-0310-a012-ef5d649a1542
pull/1/head
hermens 19 years ago
parent ec13ce9cdf
commit 971247b78f

@ -216,7 +216,8 @@ public final class plasmaWordIndex {
public Iterator wordHashes(String startHash, boolean up, boolean rot) { public Iterator wordHashes(String startHash, boolean up, boolean rot) {
//return ramCache.wordHashes(startHash, up); //return ramCache.wordHashes(startHash, up);
return new correctedWordIterator(up, rot, startHash); // use correction until bug is found if (rot) return new rotatingWordIterator(up, startHash);
else return new correctedWordIterator(up, rot, startHash); // use correction until bug is found
} }
private final class correctedWordIterator implements Iterator { private final class correctedWordIterator implements Iterator {
@ -224,7 +225,7 @@ public final class plasmaWordIndex {
String nextWord; String nextWord;
public correctedWordIterator(boolean up, boolean rotating, String firstWord) { public correctedWordIterator(boolean up, boolean rotating, String firstWord) {
iter = ramCache.wordHashes(firstWord, up); iter = ramCache.wordHashes(firstWord, up, rotating);
nextWord = (iter.hasNext()) ? (String) iter.next() : null; nextWord = (iter.hasNext()) ? (String) iter.next() : null;
boolean corrected = true; boolean corrected = true;
int cc = 0; // to avoid rotation loops int cc = 0; // to avoid rotation loops
@ -262,10 +263,40 @@ public final class plasmaWordIndex {
} }
public void remove() { public void remove() {
throw new java.lang.UnsupportedOperationException("kelondroTree: remove in kelondro Tables not yet supported"); throw new java.lang.UnsupportedOperationException("correctedWordIterator does not support remove");
} }
} // correctedWordIterator } // correctedWordIterator
private class rotatingWordIterator implements Iterator {
Iterator i;
boolean up;
public rotatingWordIterator(boolean up, String startWordHash) {
this.up = up;
i = new correctedWordIterator(up, false, startWordHash);
}
public void finalize() {
i = null;
}
public boolean hasNext() {
if (i.hasNext()) return true;
else {
i = new correctedWordIterator(up, false, (up)?"------------":"zzzzzzzzzzzz");
return i.hasNext();
}
}
public Object next() {
return i.next();
}
public void remove() {
throw new java.lang.UnsupportedOperationException("rotatingWordIterator does not support remove");
}
} // class rotatingWordIterator
public Iterator fileIterator(String startHash, boolean up, boolean deleteEmpty) { public Iterator fileIterator(String startHash, boolean up, boolean deleteEmpty) {
return new iterateFiles(startHash, up, deleteEmpty); return new iterateFiles(startHash, up, deleteEmpty);
} }

@ -215,8 +215,14 @@ public final class plasmaWordIndexAssortmentCluster {
} }
public Iterator hashConjunction(String startWordHash, boolean up) { public Iterator hashConjunction(String startWordHash, boolean up) {
// Old convention implies rot = true
return hashConjunction(startWordHash, up, true);
}
public Iterator hashConjunction(String startWordHash, boolean up, boolean rot) {
HashSet iterators = new HashSet(); HashSet iterators = new HashSet();
for (int i = 0; i < clusterCount; i++) iterators.add(assortments[i].hashes(startWordHash, up, true)); //if (rot) System.out.println("WARNING: kelondroMergeIterator does not work correctly when individual iterators rotate on their own!");
for (int i = 0; i < clusterCount; i++) iterators.add(assortments[i].hashes(startWordHash, up, rot));
return kelondroMergeIterator.cascade(iterators, kelondroNaturalOrder.naturalOrder, up); return kelondroMergeIterator.cascade(iterators, kelondroNaturalOrder.naturalOrder, up);
} }

@ -259,15 +259,23 @@ public final class plasmaWordIndexCache implements plasmaWordIndexInterface {
} }
public Iterator wordHashes(String startWordHash, boolean up) { public Iterator wordHashes(String startWordHash, boolean up) {
// Old convention implies rot = true
//return new rotatingWordHashes(startWordHash, up);
return wordHashes(startWordHash, up, true);
}
public Iterator wordHashes(String startWordHash, boolean up, boolean rot) {
// here we merge 3 databases into one view: // here we merge 3 databases into one view:
// - the RAM Cache // - the RAM Cache
// - the assortmentCluster File Cache // - the assortmentCluster File Cache
// - the backend // - the backend
if (!(up)) throw new RuntimeException("plasmaWordIndexCache.wordHashes can only count up"); if (!(up)) throw new RuntimeException("plasmaWordIndexCache.wordHashes can only count up");
//if (rot) System.out.println("WARNING: wordHashes does not work correctly when individual Assotments rotate on their own!");
//return new rotatingWordHashes(startWordHash, up);
return new kelondroMergeIterator( return new kelondroMergeIterator(
new kelondroMergeIterator( new kelondroMergeIterator(
cache.tailMap(startWordHash).keySet().iterator(), cache.tailMap(startWordHash).keySet().iterator(),
assortmentCluster.hashConjunction(startWordHash, true), assortmentCluster.hashConjunction(startWordHash, true, rot),
kelondroNaturalOrder.naturalOrder, kelondroNaturalOrder.naturalOrder,
true), true),
backend.wordHashes(startWordHash, true), backend.wordHashes(startWordHash, true),
@ -552,4 +560,4 @@ public final class plasmaWordIndexCache implements plasmaWordIndexInterface {
} }
} }
} }

@ -1265,6 +1265,7 @@ public final class yacy {
} }
private static void RWIHashList(String homePath, String targetName, String resource) { private static void RWIHashList(String homePath, String targetName, String resource) {
plasmaWordIndex WordIndex = null;
serverLog log = new serverLog("HASHLIST"); serverLog log = new serverLog("HASHLIST");
File homeDBroot = new File(new File(homePath), "DATA/PLASMADB"); File homeDBroot = new File(new File(homePath), "DATA/PLASMADB");
String wordChunkStartHash = "------------"; String wordChunkStartHash = "------------";
@ -1276,11 +1277,11 @@ public final class yacy {
BufferedOutputStream bos = new BufferedOutputStream(new FileOutputStream(file)); BufferedOutputStream bos = new BufferedOutputStream(new FileOutputStream(file));
Iterator WordHashIterator = null; Iterator WordHashIterator = null;
if (resource.equals("all")) { if (resource.equals("all")) {
plasmaWordIndex WordIndex = new plasmaWordIndex(homeDBroot, 8*1024*1024, log); WordIndex = new plasmaWordIndex(homeDBroot, 8*1024*1024, log);
WordHashIterator = WordIndex.wordHashes(wordChunkStartHash, true, false); WordHashIterator = WordIndex.wordHashes(wordChunkStartHash, true, false);
} else if (resource.equals("assortments")) { } else if (resource.equals("assortments")) {
plasmaWordIndexAssortmentCluster assortmentCluster = new plasmaWordIndexAssortmentCluster(new File(homeDBroot, "ACLUSTER"), 64, 16*1024*1024, log); plasmaWordIndexAssortmentCluster assortmentCluster = new plasmaWordIndexAssortmentCluster(new File(homeDBroot, "ACLUSTER"), 64, 16*1024*1024, log);
WordHashIterator = assortmentCluster.hashConjunction(wordChunkStartHash, true); WordHashIterator = assortmentCluster.hashConjunction(wordChunkStartHash, true, false);
} else if (resource.startsWith("assortment")) { } else if (resource.startsWith("assortment")) {
int a = Integer.parseInt(resource.substring(10)); int a = Integer.parseInt(resource.substring(10));
plasmaWordIndexAssortment assortment = new plasmaWordIndexAssortment(new File(homeDBroot, "ACLUSTER"), a, 8*1024*1024, null); plasmaWordIndexAssortment assortment = new plasmaWordIndexAssortment(new File(homeDBroot, "ACLUSTER"), a, 8*1024*1024, null);
@ -1299,9 +1300,14 @@ public final class yacy {
log.logInfo("Found " + counter + " Hashs until now. Last found Hash: " + wordHash); log.logInfo("Found " + counter + " Hashs until now. Last found Hash: " + wordHash);
} }
} }
bos.close();
} catch (IOException e) { } catch (IOException e) {
e.printStackTrace(); e.printStackTrace();
} }
if (WordIndex != null) {
WordIndex.close(60);
WordIndex = null;
}
} }

Loading…
Cancel
Save