1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
|
package de.lmu.ifi.dbs.elki.index.tree.spatial.rstarvariants.query;
/*
This file is part of ELKI:
Environment for Developing KDD-Applications Supported by Index-Structures
Copyright (C) 2012
Ludwig-Maximilians-Universität München
Lehr- und Forschungseinheit für Datenbanksysteme
ELKI Development Team
This program is free software: you can redistribute it and/or modify
it under the terms of the GNU Affero General Public License as published by
the Free Software Foundation, either version 3 of the License, or
(at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU Affero General Public License for more details.
You should have received a copy of the GNU Affero General Public License
along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
import java.util.ArrayList;
import java.util.Collections;
import java.util.HashMap;
import java.util.List;
import java.util.Map;
import java.util.Map.Entry;
import de.lmu.ifi.dbs.elki.data.spatial.SpatialComparable;
import de.lmu.ifi.dbs.elki.database.ids.ArrayDBIDs;
import de.lmu.ifi.dbs.elki.database.ids.DBID;
import de.lmu.ifi.dbs.elki.database.ids.DBIDIter;
import de.lmu.ifi.dbs.elki.database.ids.DBIDRef;
import de.lmu.ifi.dbs.elki.database.ids.DBIDUtil;
import de.lmu.ifi.dbs.elki.database.ids.DBIDs;
import de.lmu.ifi.dbs.elki.database.ids.ModifiableDBIDs;
import de.lmu.ifi.dbs.elki.database.query.distance.SpatialDistanceQuery;
import de.lmu.ifi.dbs.elki.database.query.knn.AbstractDistanceKNNQuery;
import de.lmu.ifi.dbs.elki.database.query.knn.KNNResult;
import de.lmu.ifi.dbs.elki.distance.DistanceUtil;
import de.lmu.ifi.dbs.elki.distance.distancefunction.SpatialPrimitiveDistanceFunction;
import de.lmu.ifi.dbs.elki.distance.distancevalue.Distance;
import de.lmu.ifi.dbs.elki.index.tree.DirectoryEntry;
import de.lmu.ifi.dbs.elki.index.tree.DistanceEntry;
import de.lmu.ifi.dbs.elki.index.tree.LeafEntry;
import de.lmu.ifi.dbs.elki.index.tree.query.GenericDistanceSearchCandidate;
import de.lmu.ifi.dbs.elki.index.tree.spatial.SpatialEntry;
import de.lmu.ifi.dbs.elki.index.tree.spatial.rstarvariants.AbstractRStarTree;
import de.lmu.ifi.dbs.elki.index.tree.spatial.rstarvariants.AbstractRStarTreeNode;
import de.lmu.ifi.dbs.elki.utilities.datastructures.heap.Heap;
import de.lmu.ifi.dbs.elki.utilities.datastructures.heap.KNNHeap;
import de.lmu.ifi.dbs.elki.utilities.documentation.Reference;
/**
* Instance of a KNN query for a particular spatial index.
*
* Reference:
* <p>
* G. R. Hjaltason, H. Samet<br />
* Ranking in spatial databases<br />
* In: 4th Symposium on Advances in Spatial Databases, SSD'95
* </p>
*
* @author Erich Schubert
*
* @apiviz.uses AbstractRStarTree
* @apiviz.uses SpatialPrimitiveDistanceFunction
*/
@Reference(authors = "G. R. Hjaltason, H. Samet", title = "Ranking in spatial databases", booktitle = "Advances in Spatial Databases - 4th Symposium, SSD'95", url = "http://dx.doi.org/10.1007/3-540-60159-7_6")
public class GenericRStarTreeKNNQuery<O extends SpatialComparable, D extends Distance<D>> extends AbstractDistanceKNNQuery<O, D> {
/**
* The index to use
*/
protected final AbstractRStarTree<?, ?> tree;
/**
* Spatial primitive distance function
*/
protected final SpatialPrimitiveDistanceFunction<? super O, D> distanceFunction;
/**
* Constructor.
*
* @param tree Index to use
* @param distanceQuery Distance query to use
*/
public GenericRStarTreeKNNQuery(AbstractRStarTree<?, ?> tree, SpatialDistanceQuery<O, D> distanceQuery) {
super(distanceQuery);
this.tree = tree;
this.distanceFunction = distanceQuery.getDistanceFunction();
}
/**
* Performs a k-nearest neighbor query for the given NumberVector with the
* given parameter k and the according distance function. The query result is
* in ascending order to the distance to the query object.
*
* @param object the query object
* @param knnList the knn list containing the result
*/
protected void doKNNQuery(O object, KNNHeap<D> knnList) {
final Heap<GenericDistanceSearchCandidate<D>> pq = new Heap<GenericDistanceSearchCandidate<D>>(Math.min(knnList.getK() * 2, 20));
// push root
pq.add(new GenericDistanceSearchCandidate<D>(distanceFunction.getDistanceFactory().nullDistance(), tree.getRootID()));
D maxDist = distanceFunction.getDistanceFactory().infiniteDistance();
// search in tree
while(!pq.isEmpty()) {
GenericDistanceSearchCandidate<D> pqNode = pq.poll();
if(pqNode.mindist.compareTo(maxDist) > 0) {
return;
}
maxDist = expandNode(object, knnList, pq, maxDist, pqNode.nodeID);
}
}
private D expandNode(O object, KNNHeap<D> knnList, final Heap<GenericDistanceSearchCandidate<D>> pq, D maxDist, final Integer nodeID) {
AbstractRStarTreeNode<?, ?> node = tree.getNode(nodeID);
// data node
if(node.isLeaf()) {
for(int i = 0; i < node.getNumEntries(); i++) {
SpatialEntry entry = node.getEntry(i);
D distance = distanceFunction.minDist(entry, object);
tree.distanceCalcs++;
if(distance.compareTo(maxDist) <= 0) {
knnList.add(distance, ((LeafEntry) entry).getDBID());
maxDist = knnList.getKNNDistance();
}
}
}
// directory node
else {
for(int i = 0; i < node.getNumEntries(); i++) {
SpatialEntry entry = node.getEntry(i);
D distance = distanceFunction.minDist(entry, object);
tree.distanceCalcs++;
// Greedy expand, bypassing the queue
if(distance.isNullDistance()) {
expandNode(object, knnList, pq, maxDist, ((DirectoryEntry) entry).getPageID());
}
else {
if(distance.compareTo(maxDist) <= 0) {
pq.add(new GenericDistanceSearchCandidate<D>(distance, ((DirectoryEntry) entry).getPageID()));
}
}
}
}
return maxDist;
}
/**
* Performs a batch knn query.
*
* @param node the node for which the query should be performed
* @param knnLists a map containing the knn lists for each query objects
*/
protected void batchNN(AbstractRStarTreeNode<?, ?> node, Map<DBID, KNNHeap<D>> knnLists) {
if(node.isLeaf()) {
for(int i = 0; i < node.getNumEntries(); i++) {
SpatialEntry p = node.getEntry(i);
for(Entry<DBID, KNNHeap<D>> ent : knnLists.entrySet()) {
final DBID q = ent.getKey();
final KNNHeap<D> knns_q = ent.getValue();
D knn_q_maxDist = knns_q.getKNNDistance();
DBID pid = ((LeafEntry) p).getDBID();
// FIXME: objects are NOT accessible by DBID in a plain rtree context!
D dist_pq = distanceQuery.distance(pid, q);
if(dist_pq.compareTo(knn_q_maxDist) <= 0) {
knns_q.add(dist_pq, pid);
}
}
}
}
else {
ModifiableDBIDs ids = DBIDUtil.newArray(knnLists.size());
for(DBID id : knnLists.keySet()) {
ids.add(id);
}
List<DistanceEntry<D, SpatialEntry>> entries = getSortedEntries(node, ids);
for(DistanceEntry<D, SpatialEntry> distEntry : entries) {
D minDist = distEntry.getDistance();
for(Entry<DBID, KNNHeap<D>> ent : knnLists.entrySet()) {
final KNNHeap<D> knns_q = ent.getValue();
D knn_q_maxDist = knns_q.getKNNDistance();
if(minDist.compareTo(knn_q_maxDist) <= 0) {
SpatialEntry entry = distEntry.getEntry();
AbstractRStarTreeNode<?, ?> child = tree.getNode(((DirectoryEntry) entry).getPageID());
batchNN(child, knnLists);
break;
}
}
}
}
}
@Override
public void getKNNForBulkHeaps(Map<DBID, KNNHeap<D>> heaps) {
AbstractRStarTreeNode<?, ?> root = tree.getRoot();
batchNN(root, heaps);
}
/**
* Sorts the entries of the specified node according to their minimum distance
* to the specified objects.
*
* @param node the node
* @param ids the id of the objects
* @return a list of the sorted entries
*/
protected List<DistanceEntry<D, SpatialEntry>> getSortedEntries(AbstractRStarTreeNode<?, ?> node, DBIDs ids) {
List<DistanceEntry<D, SpatialEntry>> result = new ArrayList<DistanceEntry<D, SpatialEntry>>();
for(int i = 0; i < node.getNumEntries(); i++) {
SpatialEntry entry = node.getEntry(i);
D minMinDist = distanceQuery.getDistanceFactory().infiniteDistance();
for (DBIDIter iter = ids.iter(); iter.valid(); iter.advance()) {
D minDist = distanceFunction.minDist(entry, relation.get(iter));
minMinDist = DistanceUtil.min(minDist, minMinDist);
}
result.add(new DistanceEntry<D, SpatialEntry>(entry, minMinDist, i));
}
Collections.sort(result);
return result;
}
@Override
public KNNResult<D> getKNNForObject(O obj, int k) {
if(k < 1) {
throw new IllegalArgumentException("At least one enumeration has to be requested!");
}
final KNNHeap<D> knnList = new KNNHeap<D>(k, distanceFunction.getDistanceFactory().infiniteDistance());
doKNNQuery(obj, knnList);
return knnList.toKNNList();
}
@Override
public KNNResult<D> getKNNForDBID(DBIDRef id, int k) {
return getKNNForObject(relation.get(id), k);
}
@Override
public List<KNNResult<D>> getKNNForBulkDBIDs(ArrayDBIDs ids, int k) {
if(k < 1) {
throw new IllegalArgumentException("At least one enumeration has to be requested!");
}
// While this works, it seems to be slow at least for large sets!
final Map<DBID, KNNHeap<D>> knnLists = new HashMap<DBID, KNNHeap<D>>(ids.size());
for (DBIDIter iter = ids.iter(); iter.valid(); iter.advance()) {
knnLists.put(iter.getDBID(), new KNNHeap<D>(k, distanceFunction.getDistanceFactory().infiniteDistance()));
}
batchNN(tree.getRoot(), knnLists);
List<KNNResult<D>> result = new ArrayList<KNNResult<D>>();
for (DBIDIter iter = ids.iter(); iter.valid(); iter.advance()) {
result.add(knnLists.get(iter.getDBID()).toKNNList());
}
return result;
}
}
|