use of de.lmu.ifi.dbs.elki.database.ids.ModifiableDoubleDBIDList in project elki by elki-project.
the class RdKNNTree method bulkReverseKNNQueryForID.
public List<ModifiableDoubleDBIDList> bulkReverseKNNQueryForID(DBIDs ids, int k, SpatialPrimitiveDistanceFunction<? super O> distanceFunction, KNNQuery<O> knnQuery) {
checkDistanceFunction(distanceFunction);
if (k > settings.k_max) {
throw new IllegalArgumentException("Parameter k is not supported, k > k_max: " + k + " > " + settings.k_max);
}
// get candidates
Map<DBID, ModifiableDoubleDBIDList> candidateMap = new HashMap<>();
for (DBIDIter iter = ids.iter(); iter.valid(); iter.advance()) {
DBID id = DBIDUtil.deref(iter);
candidateMap.put(id, DBIDUtil.newDistanceDBIDList());
}
doBulkReverseKNN(getRoot(), ids, candidateMap);
if (k == settings.k_max) {
List<ModifiableDoubleDBIDList> resultList = new ArrayList<>();
for (ModifiableDoubleDBIDList candidates : candidateMap.values()) {
candidates.sort();
resultList.add(candidates);
}
return resultList;
}
// refinement of candidates, if k < k_max
// perform a knn query for the candidates
ArrayModifiableDBIDs candidateIDs = DBIDUtil.newArray();
for (ModifiableDoubleDBIDList candidates : candidateMap.values()) {
candidateIDs.addDBIDs(candidates);
}
candidateIDs.sort();
List<? extends KNNList> knnLists = knnQuery.getKNNForBulkDBIDs(candidateIDs, k);
// and add candidate c to the result if o is a knn of c
List<ModifiableDoubleDBIDList> resultList = new ArrayList<>();
for (DBID id : candidateMap.keySet()) {
ModifiableDoubleDBIDList candidates = candidateMap.get(id);
ModifiableDoubleDBIDList result = DBIDUtil.newDistanceDBIDList();
for (DoubleDBIDListIter candidate = candidates.iter(); candidate.valid(); candidate.advance()) {
int pos = candidateIDs.binarySearch(candidate);
assert (pos >= 0);
for (DoubleDBIDListIter qr = knnLists.get(pos).iter(); qr.valid(); qr.advance()) {
if (DBIDUtil.equal(id, qr)) {
result.add(qr.doubleValue(), candidate);
break;
}
}
}
resultList.add(result);
}
return resultList;
}
use of de.lmu.ifi.dbs.elki.database.ids.ModifiableDoubleDBIDList in project elki by elki-project.
the class SimplifiedCoverTree method bulkConstruct.
/**
* Bulk-load the cover tree.
*
* This bulk-load is slightly simpler than the one used in the original
* cover-tree source: We do not look back into the "far" set of candidates.
*
* @param cur Current routing object
* @param maxScale Maximum scale
* @param elems Candidates
* @return Root node of subtree
*/
protected Node bulkConstruct(DBIDRef cur, int maxScale, ModifiableDoubleDBIDList elems) {
assert (!elems.contains(cur));
final double max = maxDistance(elems);
final int scale = Math.min(distToScale(max) - 1, maxScale);
final int nextScale = scale - 1;
// elements remaining:
if (max <= 0 || scale <= scaleBottom || elems.size() < truncate) {
return new Node(cur, max, elems);
}
// Find neighbors in the cover of the current object:
ModifiableDoubleDBIDList candidates = DBIDUtil.newDistanceDBIDList();
excludeNotCovered(elems, scaleToDist(scale), candidates);
// If no elements were not in the cover, build a compact tree:
if (candidates.size() == 0) {
LOG.warning("Scale not chosen appropriately? " + max + " " + scaleToDist(scale));
return bulkConstruct(cur, nextScale, elems);
}
// We will have at least one other child, so build the parent:
Node node = new Node(cur, max);
// Routing element now is a singleton:
final boolean curSingleton = elems.size() == 0;
if (!curSingleton) {
// Add node for the routing object:
node.children.add(bulkConstruct(cur, nextScale, elems));
}
final double fmax = scaleToDist(nextScale);
// Build additional cover nodes:
for (DoubleDBIDListIter it = candidates.iter(); it.valid(); ) {
assert (it.getOffset() == 0);
DBID t = DBIDUtil.deref(it);
// Recycle.
elems.clear();
collectByCover(it, candidates, fmax, elems);
assert (DBIDUtil.equal(t, it)) : "First element in candidates must not change!";
if (elems.size() == 0) {
// Singleton
node.singletons.add(it);
} else {
// Build a full child node:
node.children.add(bulkConstruct(it, nextScale, elems));
}
candidates.removeSwap(0);
}
assert (candidates.size() == 0);
// Routing object is not yet handled:
if (curSingleton) {
if (node.isLeaf()) {
// First in leaf is enough.
node.children = null;
} else {
// Add as regular singleton.
node.singletons.add(cur);
}
}
// TODO: improve recycling of lists?
return node;
}
use of de.lmu.ifi.dbs.elki.database.ids.ModifiableDoubleDBIDList in project elki by elki-project.
the class SimplifiedCoverTree method bulkLoad.
/**
* Bulk-load the index.
*
* @param ids IDs to load
*/
public void bulkLoad(DBIDs ids) {
if (ids.size() == 0) {
return;
}
assert (root == null) : "Tree already initialized.";
DBIDIter it = ids.iter();
DBID first = DBIDUtil.deref(it);
// Compute distances to all neighbors:
ModifiableDoubleDBIDList candidates = DBIDUtil.newDistanceDBIDList(ids.size() - 1);
for (it.advance(); it.valid(); it.advance()) {
candidates.add(distance(first, it), it);
}
root = bulkConstruct(first, Integer.MAX_VALUE, candidates);
}
use of de.lmu.ifi.dbs.elki.database.ids.ModifiableDoubleDBIDList in project elki by elki-project.
the class InMemoryInvertedIndex method initialize.
@Override
public void initialize() {
if (index != null) {
LOG.warning("Index was already initialized!");
}
index = new ArrayList<>();
length = DataStoreUtil.makeDoubleStorage(relation.getDBIDs(), DataStoreFactory.HINT_DB);
for (DBIDIter iter = relation.iterDBIDs(); iter.valid(); iter.advance()) {
V obj = relation.get(iter);
if (obj instanceof SparseNumberVector) {
indexSparse(iter, (SparseNumberVector) obj);
} else {
indexDense(iter, obj);
}
}
// Sort indexes
long count = 0L;
for (ModifiableDoubleDBIDList column : index) {
column.sort();
count += column.size();
}
double sparsity = count / (index.size() * (double) relation.size());
if (sparsity > .2) {
LOG.warning("Inverted list indexes only perform well for very sparse data. Your data set has a sparsity of " + sparsity);
}
}
use of de.lmu.ifi.dbs.elki.database.ids.ModifiableDoubleDBIDList in project elki by elki-project.
the class CacheDoubleDistanceRangeQueries method run.
@Override
public void run() {
database.initialize();
Relation<O> relation = database.getRelation(distance.getInputTypeRestriction());
DistanceQuery<O> distanceQuery = database.getDistanceQuery(relation, distance);
RangeQuery<O> rangeQ = database.getRangeQuery(distanceQuery, radius, DatabaseQuery.HINT_HEAVY_USE);
LOG.verbose("Performing range queries with radius " + radius);
// open file.
try (RandomAccessFile file = new RandomAccessFile(out, "rw");
FileChannel channel = file.getChannel();
// and acquire a file write lock
FileLock lock = channel.lock()) {
// write magic header
file.writeInt(RANGE_CACHE_MAGIC);
// write the query radius.
file.writeDouble(radius);
// Initial size, enough for 100.
int bufsize = 100 * 12 * 2 + 10;
ByteBuffer buffer = ByteBuffer.allocateDirect(bufsize);
FiniteProgress prog = LOG.isVerbose() ? new FiniteProgress("Computing range queries", relation.size(), LOG) : null;
ModifiableDoubleDBIDList nn = DBIDUtil.newDistanceDBIDList();
DoubleDBIDListIter ni = nn.iter();
for (DBIDIter it = relation.iterDBIDs(); it.valid(); it.advance()) {
nn.clear();
rangeQ.getRangeForDBID(it, radius, nn);
nn.sort();
final int nnsize = nn.size();
// Grow the buffer when needed:
if (nnsize * 12 + 10 > bufsize) {
while (nnsize * 12 + 10 > bufsize) {
bufsize <<= 1;
}
LOG.verbose("Resizing buffer to " + bufsize + " to store " + nnsize + " results:");
buffer = ByteBuffer.allocateDirect(bufsize);
}
buffer.clear();
ByteArrayUtil.writeUnsignedVarint(buffer, it.internalGetIndex());
ByteArrayUtil.writeUnsignedVarint(buffer, nnsize);
int c = 0;
for (ni.seek(0); ni.valid(); ni.advance(), c++) {
ByteArrayUtil.writeUnsignedVarint(buffer, ni.internalGetIndex());
buffer.putDouble(ni.doubleValue());
}
if (c != nn.size()) {
throw new AbortException("Sizes did not agree. Cache is invalid.");
}
buffer.flip();
channel.write(buffer);
LOG.incrementProcessed(prog);
}
LOG.ensureCompleted(prog);
lock.release();
} catch (IOException e) {
LOG.exception(e);
}
// FIXME: close!
}
Aggregations