1 | package de.ugoe.cs.quest.coverage;
|
---|
2 |
|
---|
3 | import java.util.Collection;
|
---|
4 | import java.util.List;
|
---|
5 | import java.util.Map;
|
---|
6 |
|
---|
7 | import de.ugoe.cs.quest.eventcore.Event;
|
---|
8 | import de.ugoe.cs.quest.usageprofiles.IStochasticProcess;
|
---|
9 |
|
---|
10 | /**
|
---|
11 | * <p>
|
---|
12 | * This class calculates various types of sequence coverage in relation to a stochastic process.
|
---|
13 | * </p>
|
---|
14 | *
|
---|
15 | * @author Steffen Herbold
|
---|
16 | * @version 1.0
|
---|
17 | */
|
---|
18 | public class CoverageCalculatorProcess {
|
---|
19 |
|
---|
20 | /**
|
---|
21 | * <p>
|
---|
22 | * Stochastic process that is the foundation for probabilistic coverages and coverages with
|
---|
23 | * reference to all possible sequences.
|
---|
24 | * </p>
|
---|
25 | */
|
---|
26 | private final IStochasticProcess process;
|
---|
27 |
|
---|
28 | /**
|
---|
29 | * <p>
|
---|
30 | * Sequences for which the coverage is calculated.
|
---|
31 | * </p>
|
---|
32 | */
|
---|
33 | private Collection<List<Event>> sequences;
|
---|
34 |
|
---|
35 | /**
|
---|
36 | * <p>
|
---|
37 | * Length of the subsequences in relation to which the coverage is calculated.
|
---|
38 | * </p>
|
---|
39 | */
|
---|
40 | private final int length;
|
---|
41 |
|
---|
42 | /**
|
---|
43 | * <p>
|
---|
44 | * All subsequences of {@link #length} of {@link #sequences}.
|
---|
45 | * </p>
|
---|
46 | */
|
---|
47 | private Collection<List<Event>> containedSubSeqs = null;
|
---|
48 |
|
---|
49 | /**
|
---|
50 | * <p>
|
---|
51 | * All subsequences of {@link #length} that can be generated by {@link #process}.
|
---|
52 | * </p>
|
---|
53 | */
|
---|
54 | private Collection<List<Event>> allPossibleSubSeqs = null;
|
---|
55 |
|
---|
56 | /**
|
---|
57 | * <p>
|
---|
58 | * The probabilities of all subsequences of {@link #length} according to {@link #process}.
|
---|
59 | * </p>
|
---|
60 | */
|
---|
61 | private Map<List<Event>, Double> subSeqWeights = null;
|
---|
62 |
|
---|
63 | /**
|
---|
64 | * <p>
|
---|
65 | * Constructor. Creates a new CoverageCalculatorProcess for a given stochastic process and
|
---|
66 | * generated sequences.
|
---|
67 | * </p>
|
---|
68 | *
|
---|
69 | * @param process
|
---|
70 | * stochastic process used for coverage calculations; must not be null
|
---|
71 | * @param sequences
|
---|
72 | * sequences for which the coverage is calculated; must not be null
|
---|
73 | * @param length
|
---|
74 | * length of the subsequences for which the coverage is analyzed; must be >0
|
---|
75 | * @throws IllegalArgumentException
|
---|
76 | * thrown if process or sequences is null or length less than or equal to 0
|
---|
77 | */
|
---|
78 | public CoverageCalculatorProcess(IStochasticProcess process,
|
---|
79 | Collection<List<Event>> sequences,
|
---|
80 | int length)
|
---|
81 | {
|
---|
82 | if (process == null) {
|
---|
83 | throw new IllegalArgumentException("process must not be null");
|
---|
84 | }
|
---|
85 | if (sequences == null) {
|
---|
86 | throw new IllegalArgumentException("sequences must not be null");
|
---|
87 | }
|
---|
88 | if (length <= 0) {
|
---|
89 | throw new IllegalArgumentException("length must be >0; actual value: " + length);
|
---|
90 | }
|
---|
91 | this.process = process;
|
---|
92 | this.sequences = sequences;
|
---|
93 | this.length = length;
|
---|
94 | }
|
---|
95 |
|
---|
96 | /**
|
---|
97 | * <p>
|
---|
98 | * Calculates the percentage of subsequences of length k that occur, including those that cannot
|
---|
99 | * be generated by {@link #process}.
|
---|
100 | * </p>
|
---|
101 | *
|
---|
102 | * @return coverage percentage
|
---|
103 | */
|
---|
104 | public double getCoverageAllNoWeight() {
|
---|
105 | if (containedSubSeqs == null) {
|
---|
106 | containedSubSeqs = SequenceTools.containedSubSequences(sequences, length);
|
---|
107 | }
|
---|
108 | return ((double) containedSubSeqs.size()) / SequenceTools.numSequences(process, length);
|
---|
109 | }
|
---|
110 |
|
---|
111 | /**
|
---|
112 | * <p>
|
---|
113 | * Calculates the percentage of subsequences of length k that occur and can generated by
|
---|
114 | * {@link #process}.
|
---|
115 | * </p>
|
---|
116 | *
|
---|
117 | * @return coverage percentage
|
---|
118 | */
|
---|
119 | public double getCoveragePossibleNoWeight() {
|
---|
120 | if (containedSubSeqs == null) {
|
---|
121 | containedSubSeqs = SequenceTools.containedSubSequences(sequences, length);
|
---|
122 | }
|
---|
123 | if (allPossibleSubSeqs == null) {
|
---|
124 | allPossibleSubSeqs = process.generateSequences(length);
|
---|
125 | }
|
---|
126 | return ((double) containedSubSeqs.size()) / allPossibleSubSeqs.size();
|
---|
127 | }
|
---|
128 |
|
---|
129 | /**
|
---|
130 | * <p>
|
---|
131 | * Calculates the weight of the subsequences that occur with relation to {@link #process}, i.e.,
|
---|
132 | * the mass of the subsequence probability covered by the subsequences.
|
---|
133 | * </p>
|
---|
134 | *
|
---|
135 | * @return coverage weight
|
---|
136 | */
|
---|
137 | public double getCoveragePossibleWeight() {
|
---|
138 | if (containedSubSeqs == null) {
|
---|
139 | containedSubSeqs = SequenceTools.containedSubSequences(sequences, length);
|
---|
140 | }
|
---|
141 | if (allPossibleSubSeqs == null) {
|
---|
142 | allPossibleSubSeqs = process.generateSequences(length);
|
---|
143 | }
|
---|
144 | if (subSeqWeights == null) {
|
---|
145 | subSeqWeights = SequenceTools.generateWeights(process, allPossibleSubSeqs);
|
---|
146 | }
|
---|
147 | double weight = 0.0;
|
---|
148 | for (List<Event> subSeq : containedSubSeqs) {
|
---|
149 | Double curWeight = subSeqWeights.get(subSeq);
|
---|
150 | if (curWeight != null) {
|
---|
151 | weight += curWeight;
|
---|
152 | }
|
---|
153 | }
|
---|
154 | return weight;
|
---|
155 | }
|
---|
156 |
|
---|
157 | /**
|
---|
158 | * <p>
|
---|
159 | * Returns the number of covered subsequences of length k.
|
---|
160 | * </p>
|
---|
161 | *
|
---|
162 | * @return number of covered subsequences
|
---|
163 | */
|
---|
164 | public int getNumCovered() {
|
---|
165 | if (containedSubSeqs == null) {
|
---|
166 | containedSubSeqs = SequenceTools.containedSubSequences(sequences, length);
|
---|
167 | }
|
---|
168 | return containedSubSeqs.size();
|
---|
169 | }
|
---|
170 |
|
---|
171 | /**
|
---|
172 | * <p>
|
---|
173 | * Returns the number of possible subsequences of length k according to the stochastic process.
|
---|
174 | * </p>
|
---|
175 | *
|
---|
176 | * @return number of possible subsequences
|
---|
177 | */
|
---|
178 | public int getNumPossible() {
|
---|
179 | if (allPossibleSubSeqs == null) {
|
---|
180 | allPossibleSubSeqs = process.generateSequences(length);
|
---|
181 | }
|
---|
182 | return allPossibleSubSeqs.size();
|
---|
183 | }
|
---|
184 |
|
---|
185 | /**
|
---|
186 | * <p>
|
---|
187 | * Sets a new collection of sequences for which the coverage is analyzed.
|
---|
188 | * </p>
|
---|
189 | *
|
---|
190 | * @param newSequences
|
---|
191 | * new collection of sequences
|
---|
192 | * @throws IllegalArgumentException
|
---|
193 | * thrown is newSequences is null
|
---|
194 | */
|
---|
195 | public void setSequences(Collection<List<Event>> newSequences) {
|
---|
196 | if (newSequences == null) {
|
---|
197 | throw new IllegalArgumentException("sequences must not be null");
|
---|
198 | }
|
---|
199 | this.sequences = newSequences;
|
---|
200 | containedSubSeqs = null;
|
---|
201 | }
|
---|
202 |
|
---|
203 | }
|
---|