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
|
package de.lmu.ifi.dbs.elki.datasource.parser;
/*
This file is part of ELKI:
Environment for Developing KDD-Applications Supported by Index-Structures
Copyright (C) 2011
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.io.BufferedReader;
import java.io.IOException;
import java.io.InputStream;
import java.io.InputStreamReader;
import java.util.ArrayList;
import java.util.BitSet;
import java.util.Iterator;
import java.util.List;
import java.util.regex.Pattern;
import de.lmu.ifi.dbs.elki.data.LabelList;
import de.lmu.ifi.dbs.elki.data.NumberVector;
import de.lmu.ifi.dbs.elki.data.type.TypeUtil;
import de.lmu.ifi.dbs.elki.data.type.VectorFieldTypeInformation;
import de.lmu.ifi.dbs.elki.datasource.bundle.MultipleObjectsBundle;
import de.lmu.ifi.dbs.elki.datasource.bundle.SingleObjectBundle;
import de.lmu.ifi.dbs.elki.utilities.optionhandling.OptionID;
import de.lmu.ifi.dbs.elki.utilities.optionhandling.parameterization.Parameterization;
import de.lmu.ifi.dbs.elki.utilities.optionhandling.parameters.IntListParameter;
import de.lmu.ifi.dbs.elki.utilities.pairs.Pair;
/**
* <p>
* Provides a parser for parsing one point per line, attributes separated by
* whitespace.
* </p>
* <p>
* Several labels may be given per point. A label must not be parseable as
* double. Lines starting with "#" will be ignored.
* </p>
* <p>
* An index can be specified to identify an entry to be treated as class label.
* This index counts all entries (numeric and labels as well) starting with 0.
* </p>
*
* @author Arthur Zimek
* @param <V> the type of NumberVector used
*/
public abstract class NumberVectorLabelParser<V extends NumberVector<?, ?>> extends AbstractParser implements LinebasedParser, Parser {
/**
* A comma separated list of the indices of labels (may be numeric), counting
* whitespace separated entries in a line starting with 0. The corresponding
* entries will be treated as a label.
* <p>
* Key: {@code -parser.labelIndices}
* </p>
*/
public static final OptionID LABEL_INDICES_ID = OptionID.getOrCreateOptionID("parser.labelIndices", "A comma separated list of the indices of labels (may be numeric), counting whitespace separated entries in a line starting with 0. The corresponding entries will be treated as a label.");
/**
* Keeps the indices of the attributes to be treated as a string label.
*/
protected BitSet labelIndices;
/**
* Constructor
*
* @param colSep
* @param quoteChar
* @param labelIndices
*/
public NumberVectorLabelParser(Pattern colSep, char quoteChar, BitSet labelIndices) {
super(colSep, quoteChar);
this.labelIndices = labelIndices;
}
@Override
public MultipleObjectsBundle parse(InputStream in) {
BufferedReader reader = new BufferedReader(new InputStreamReader(in));
int lineNumber = 1;
int dimensionality = -1;
List<V> vectors = new ArrayList<V>();
List<LabelList> labels = new ArrayList<LabelList>();
try {
for(String line; (line = reader.readLine()) != null; lineNumber++) {
if(!line.startsWith(COMMENT) && line.length() > 0) {
Pair<V, LabelList> objectAndLabels = parseLineInternal(line);
if(dimensionality < 0) {
dimensionality = objectAndLabels.getFirst().getDimensionality();
}
else if(dimensionality != objectAndLabels.getFirst().getDimensionality()) {
throw new IllegalArgumentException("Differing dimensionality in line " + lineNumber + ":" + objectAndLabels.getFirst().getDimensionality() + " != " + dimensionality);
}
vectors.add(objectAndLabels.first);
labels.add(objectAndLabels.second);
}
}
}
catch(IOException e) {
throw new IllegalArgumentException("Error while parsing line " + lineNumber + ".");
}
return MultipleObjectsBundle.makeSimple(getTypeInformation(dimensionality), vectors, TypeUtil.LABELLIST, labels);
}
@Override
public SingleObjectBundle parseLine(String line) {
Pair<V, LabelList> objectAndLabels = parseLineInternal(line);
SingleObjectBundle pkg = new SingleObjectBundle();
pkg.append(getTypeInformation(objectAndLabels.first.getDimensionality()), objectAndLabels.first);
pkg.append(TypeUtil.LABELLIST, objectAndLabels.second);
return pkg;
}
/**
* Internal method for parsing a single line. Used by both line based parsig
* as well as block parsing. This saves the building of meta data for each
* line.
*
* @param line Line to process
* @return parsing result
*/
protected Pair<V, LabelList> parseLineInternal(String line) {
List<String> entries = tokenize(line);
// Split into numerical attributes and labels
List<Double> attributes = new ArrayList<Double>(entries.size());
LabelList labels = new LabelList();
Iterator<String> itr = entries.iterator();
for(int i = 0; itr.hasNext(); i++) {
String ent = itr.next();
if(!labelIndices.get(i)) {
try {
Double attribute = Double.valueOf(ent);
attributes.add(attribute);
}
catch(NumberFormatException e) {
labels.add(ent);
}
}
else {
labels.add(ent);
}
}
Pair<V, LabelList> objectAndLabels;
V vec = createDBObject(attributes);
objectAndLabels = new Pair<V, LabelList>(vec, labels);
return objectAndLabels;
}
/**
* <p>
* Creates a database object of type V.
* </p>
*
* @param attributes the attributes of the vector to create.
* @return a RalVector of type V containing the given attribute values
*/
protected abstract V createDBObject(List<Double> attributes);
/**
* Get a prototype object for the given dimensionality.
*
* @param dimensionality Dimensionality
* @return Prototype object
*/
abstract protected VectorFieldTypeInformation<V> getTypeInformation(int dimensionality);
/**
* Parameterization class.
*
* @author Erich Schubert
*
* @apiviz.exclude
*/
public static abstract class Parameterizer<V extends NumberVector<?, ?>> extends AbstractParser.Parameterizer {
/**
* Keeps the indices of the attributes to be treated as a string label.
*/
protected BitSet labelIndices = null;
@Override
protected void makeOptions(Parameterization config) {
super.makeOptions(config);
IntListParameter labelIndicesP = new IntListParameter(LABEL_INDICES_ID, true);
labelIndices = new BitSet();
if(config.grab(labelIndicesP)) {
List<Integer> labelcols = labelIndicesP.getValue();
for(Integer idx : labelcols) {
labelIndices.set(idx);
}
}
}
@Override
protected abstract NumberVectorLabelParser<V> makeInstance();
}
}
|
