1 | /*
|
---|
2 | * NumbersFilter.java
|
---|
3 | * Copyright (C) 2000 Eibe Frank
|
---|
4 | *
|
---|
5 | * This program is free software; you can redistribute it and/or modify
|
---|
6 | * it under the terms of the GNU General Public License as published by
|
---|
7 | * the Free Software Foundation; either version 2 of the License, or
|
---|
8 | * (at your option) any later version.
|
---|
9 | *
|
---|
10 | * This program is distributed in the hope that it will be useful,
|
---|
11 | * but WITHOUT ANY WARRANTY; without even the implied warranty of
|
---|
12 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
---|
13 | * GNU General Public License for more details.
|
---|
14 | *
|
---|
15 | * You should have received a copy of the GNU General Public License
|
---|
16 | * along with this program; if not, write to the Free Software
|
---|
17 | * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
|
---|
18 | */
|
---|
19 |
|
---|
20 | import weka.core.*;
|
---|
21 | import weka.filters.*;
|
---|
22 |
|
---|
23 | import java.util.*;
|
---|
24 |
|
---|
25 | /**
|
---|
26 | * Removes all numbers from all the string attributes in the given
|
---|
27 | * dataset. Assumes that words are separated by whitespace.
|
---|
28 | *
|
---|
29 | * @author Eibe Frank ([email protected])
|
---|
30 | * @version 1.0
|
---|
31 | */
|
---|
32 | public class NumbersFilter extends Filter {
|
---|
33 |
|
---|
34 | /**
|
---|
35 | * Returns a string describing this filter
|
---|
36 | *
|
---|
37 | * @return a description of the filter suitable for
|
---|
38 | * displaying in the explorer/experimenter gui
|
---|
39 | */
|
---|
40 | public String globalInfo() {
|
---|
41 | return "Removes all numbers from all the string attributes in " +
|
---|
42 | "the given dataset. Assumes that words are separated by whitespace.";
|
---|
43 | }
|
---|
44 |
|
---|
45 | /**
|
---|
46 | * Signify that this batch of input to the filter is finished. If
|
---|
47 | * the filter requires all instances prior to filtering, output()
|
---|
48 | * may now be called to retrieve the filtered instances. Any
|
---|
49 | * subsequent instances filtered should be filtered based on setting
|
---|
50 | * obtained from the first batch (unless the inputFormat has been
|
---|
51 | * re-assigned or new options have been set). This default
|
---|
52 | * implementation assumes all instance processing occurs during
|
---|
53 | * inputFormat() and input().
|
---|
54 | *
|
---|
55 | * @return true if there are instances pending output
|
---|
56 | * @exception NullPointerException if no input structure has been defined,
|
---|
57 | * @exception Exception if there was a problem finishing the batch.
|
---|
58 | */
|
---|
59 | public boolean batchFinished() throws Exception {
|
---|
60 |
|
---|
61 | if (getInputFormat() == null) {
|
---|
62 | throw new NullPointerException("No input instance format defined");
|
---|
63 | }
|
---|
64 | m_NewBatch = true;
|
---|
65 | return (numPendingOutput() != 0);
|
---|
66 | }
|
---|
67 |
|
---|
68 | /**
|
---|
69 | * Sets the format of the input instances.
|
---|
70 | *
|
---|
71 | * @param instanceInfo an Instances object containing the input
|
---|
72 | * instance structure (any instances contained in the object are
|
---|
73 | * ignored - only the structure is required).
|
---|
74 | * @return true if the outputFormat may be collected immediately
|
---|
75 | */
|
---|
76 | public boolean setInputFormat(Instances instanceInfo) throws Exception {
|
---|
77 |
|
---|
78 | super.setInputFormat(instanceInfo);
|
---|
79 | setOutputFormat(instanceInfo);
|
---|
80 | return true;
|
---|
81 | }
|
---|
82 |
|
---|
83 | /**
|
---|
84 | * Input an instance for filtering. Ordinarily the instance is processed
|
---|
85 | * and made available for output immediately. Some filters require all
|
---|
86 | * instances be read before producing output.
|
---|
87 | *
|
---|
88 | * @param instance the input instance
|
---|
89 | * @return true if the filtered instance may now be
|
---|
90 | * collected with output().
|
---|
91 | * @exception Exception if the input instance was not of the correct
|
---|
92 | * format or if there was a problem with the filtering.
|
---|
93 | */
|
---|
94 | public boolean input(Instance instance) throws Exception {
|
---|
95 |
|
---|
96 | if (getInputFormat() == null) {
|
---|
97 | throw new Exception("No input instance format defined");
|
---|
98 | }
|
---|
99 | if (m_NewBatch) {
|
---|
100 | resetQueue();
|
---|
101 | m_NewBatch = false;
|
---|
102 | }
|
---|
103 | convertInstance(instance);
|
---|
104 | return true;
|
---|
105 | }
|
---|
106 |
|
---|
107 | /**
|
---|
108 | * Main method for testing this class.
|
---|
109 | *
|
---|
110 | * @param argv should contain arguments to the filter: use -h for help
|
---|
111 | */
|
---|
112 | public static void main(String [] argv) {
|
---|
113 |
|
---|
114 | try {
|
---|
115 | if (Utils.getFlag('b', argv)) {
|
---|
116 | Filter.batchFilterFile(new NumbersFilter(), argv);
|
---|
117 | } else {
|
---|
118 | Filter.filterFile(new NumbersFilter(), argv);
|
---|
119 | }
|
---|
120 | } catch (Exception ex) {
|
---|
121 | System.out.println(ex.getMessage());
|
---|
122 | }
|
---|
123 | }
|
---|
124 |
|
---|
125 | /**
|
---|
126 | * Converts an instance. A phrase boundary is inserted where
|
---|
127 | * a number is found.
|
---|
128 | */
|
---|
129 | private void convertInstance(Instance instance) throws Exception {
|
---|
130 |
|
---|
131 | double[] instVals = new double[instance.numAttributes()];
|
---|
132 |
|
---|
133 | for (int i = 0; i < instance.numAttributes(); i++) {
|
---|
134 | if ((!instance.attribute(i).isString()) || instance.isMissing(i)) {
|
---|
135 | instVals[i] = instance.value(i);
|
---|
136 | } else {
|
---|
137 | String str = instance.stringValue(i);
|
---|
138 | StringBuffer resultStr = new StringBuffer();
|
---|
139 | StringTokenizer tok = new StringTokenizer(str, " \t\n", true);
|
---|
140 | while (tok.hasMoreTokens()) {
|
---|
141 | String token = tok.nextToken();
|
---|
142 |
|
---|
143 | // Everything that doesn't contain at least
|
---|
144 | // one letter is considered to be a number
|
---|
145 | boolean isNumber = true;
|
---|
146 | for (int j = 0; j < token.length(); j++) {
|
---|
147 | if (Character.isLetter(token.charAt(j))) {
|
---|
148 | isNumber = false;
|
---|
149 | break;
|
---|
150 | }
|
---|
151 | }
|
---|
152 | if (!isNumber) {
|
---|
153 | resultStr.append(token);
|
---|
154 | } else {
|
---|
155 | if (token.equals(" ") || token.equals("\t") ||
|
---|
156 | token.equals("\n")) {
|
---|
157 | resultStr.append(token);
|
---|
158 | } else {
|
---|
159 | resultStr.append(" \n ");
|
---|
160 | }
|
---|
161 | }
|
---|
162 | }
|
---|
163 | int index = getOutputFormat().attribute(i).addStringValue(resultStr.toString());
|
---|
164 | instVals[i] = (double)index;
|
---|
165 | }
|
---|
166 | }
|
---|
167 | Instance inst = new Instance(instance.weight(), instVals);
|
---|
168 | inst.setDataset(getOutputFormat());
|
---|
169 | push(inst);
|
---|
170 | }
|
---|
171 | }
|
---|
172 |
|
---|
173 |
|
---|
174 |
|
---|
175 |
|
---|
176 |
|
---|
177 |
|
---|
178 |
|
---|
179 |
|
---|
180 |
|
---|
181 |
|
---|
182 |
|
---|