[123] | 1 | package de.ugoe.cs.eventbench.coverage;
|
---|
| 2 |
|
---|
[328] | 3 | import java.security.InvalidParameterException;
|
---|
[123] | 4 | import java.util.Collection;
|
---|
| 5 | import java.util.LinkedHashMap;
|
---|
| 6 | import java.util.LinkedHashSet;
|
---|
| 7 | import java.util.LinkedList;
|
---|
| 8 | import java.util.List;
|
---|
| 9 | import java.util.Map;
|
---|
| 10 | import java.util.Set;
|
---|
| 11 |
|
---|
| 12 | import de.ugoe.cs.eventbench.data.Event;
|
---|
| 13 | import de.ugoe.cs.eventbench.models.IStochasticProcess;
|
---|
| 14 |
|
---|
| 15 | /**
|
---|
| 16 | * <p>
|
---|
| 17 | * This class gathers some helper functions to work with sequences.
|
---|
| 18 | * </p>
|
---|
| 19 | *
|
---|
| 20 | * @author Steffen Herbold
|
---|
| 21 | * @version 1.0
|
---|
| 22 | */
|
---|
| 23 | public class SequenceTools {
|
---|
| 24 |
|
---|
| 25 | /**
|
---|
| 26 | * <p>
|
---|
| 27 | * Calculates the weights for all sequences passed to this function as
|
---|
| 28 | * defined by the stochastic process and stores them in a {@link Map}. The
|
---|
| 29 | * weights are normalized, i.e., the sum of all weights contained in this
|
---|
| 30 | * map is 1.
|
---|
| 31 | * </p>
|
---|
| 32 | *
|
---|
| 33 | * @param process
|
---|
| 34 | * process used for weight calculation
|
---|
| 35 | * @param sequences
|
---|
| 36 | * sequences for which the weights are calculated
|
---|
| 37 | * @return {@link Map} of weights
|
---|
| 38 | */
|
---|
| 39 | public static Map<List<? extends Event<?>>, Double> generateWeights(
|
---|
| 40 | IStochasticProcess process,
|
---|
| 41 | Collection<List<? extends Event<?>>> sequences) {
|
---|
| 42 | Map<List<? extends Event<?>>, Double> subSeqWeights = new LinkedHashMap<List<? extends Event<?>>, Double>();
|
---|
[328] | 43 | if (sequences != null && !sequences.isEmpty()) {
|
---|
| 44 | if (process != null) {
|
---|
| 45 | double sum = 0.0;
|
---|
| 46 | for (List<? extends Event<?>> sequence : sequences) {
|
---|
| 47 | double prob = process.getProbability(sequence);
|
---|
| 48 | subSeqWeights.put(sequence, prob);
|
---|
| 49 | sum += prob;
|
---|
| 50 | }
|
---|
| 51 | if (sum < 1.0) {
|
---|
| 52 | for (Map.Entry<List<? extends Event<?>>, Double> entry : subSeqWeights
|
---|
| 53 | .entrySet()) {
|
---|
| 54 | entry.setValue(entry.getValue() / sum);
|
---|
| 55 | }
|
---|
| 56 | }
|
---|
| 57 | } else {
|
---|
| 58 | for( List<? extends Event<?>> sequence : sequences ) {
|
---|
| 59 | subSeqWeights.put(sequence, 0.0d);
|
---|
| 60 | }
|
---|
[123] | 61 | }
|
---|
| 62 | }
|
---|
| 63 | return subSeqWeights;
|
---|
| 64 | }
|
---|
| 65 |
|
---|
| 66 | /**
|
---|
| 67 | * <p>
|
---|
| 68 | * Calculates the number of all existing sequences of a given length,
|
---|
| 69 | * regardless whether they are possible or not.
|
---|
| 70 | * </p>
|
---|
| 71 | *
|
---|
| 72 | * @param process
|
---|
| 73 | * stochastic process whose symbols are the basis for this
|
---|
| 74 | * calculation
|
---|
| 75 | * @param length
|
---|
| 76 | * lenght of the sequences
|
---|
| 77 | * @return numStates^length
|
---|
[328] | 78 | * @throws InvalidParameterException
|
---|
| 79 | * thrown if length less or equal to 0
|
---|
[123] | 80 | */
|
---|
| 81 | public static long numSequences(IStochasticProcess process, int length) {
|
---|
[328] | 82 | if (length <= 0) {
|
---|
| 83 | throw new InvalidParameterException(
|
---|
| 84 | "length must be a positive integer");
|
---|
| 85 | }
|
---|
| 86 | long result = 0;
|
---|
| 87 | if (process != null) {
|
---|
| 88 | result = (long) Math.pow(process.getNumSymbols(), length);
|
---|
| 89 | }
|
---|
| 90 | return result;
|
---|
[123] | 91 | }
|
---|
| 92 |
|
---|
| 93 | /**
|
---|
| 94 | * <p>
|
---|
| 95 | * Creates a {@link Set} of all subsequences of a given length that are
|
---|
| 96 | * contained in a sequence collection.
|
---|
| 97 | * </p>
|
---|
| 98 | *
|
---|
| 99 | * @param sequences
|
---|
| 100 | * sequences from which the subsequences are extracted
|
---|
| 101 | * @param length
|
---|
| 102 | * length of the subsequences
|
---|
| 103 | * @return {@link Set} of all subsequences
|
---|
[328] | 104 | * @throws InvalidParameterException
|
---|
| 105 | * thrown if length less or equal to 0
|
---|
[123] | 106 | */
|
---|
| 107 | public static Set<List<? extends Event<?>>> containedSubSequences(
|
---|
| 108 | Collection<List<? extends Event<?>>> sequences, int length) {
|
---|
[328] | 109 | if (length <= 0) {
|
---|
| 110 | throw new InvalidParameterException(
|
---|
| 111 | "length must be a positive integer");
|
---|
| 112 | }
|
---|
[123] | 113 | Set<List<? extends Event<?>>> containedSubSeqs = new LinkedHashSet<List<? extends Event<?>>>();
|
---|
[328] | 114 | if (sequences != null) {
|
---|
| 115 | for (List<? extends Event<?>> sequence : sequences) {
|
---|
| 116 | List<Event<?>> subSeq = new LinkedList<Event<?>>();
|
---|
| 117 | boolean minLengthReached = false;
|
---|
| 118 | for (Event<?> event : sequence) {
|
---|
| 119 | subSeq.add(event);
|
---|
| 120 | if (!minLengthReached) {
|
---|
| 121 | if (subSeq.size() == length) {
|
---|
| 122 | minLengthReached = true;
|
---|
| 123 | }
|
---|
| 124 | } else {
|
---|
| 125 | subSeq.remove(0);
|
---|
[123] | 126 | }
|
---|
[328] | 127 | if (minLengthReached) {
|
---|
| 128 | if (!containedSubSeqs.contains(subSeq)) {
|
---|
| 129 | containedSubSeqs.add(new LinkedList<Event<?>>(
|
---|
| 130 | subSeq));
|
---|
| 131 | }
|
---|
[123] | 132 | }
|
---|
| 133 | }
|
---|
| 134 | }
|
---|
| 135 | }
|
---|
| 136 | return containedSubSeqs;
|
---|
| 137 | }
|
---|
| 138 | }
|
---|