[1895] | 1 | // Copyright 2012 Georg-August-Universität Göttingen, Germany
|
---|
| 2 | //
|
---|
| 3 | // Licensed under the Apache License, Version 2.0 (the "License");
|
---|
| 4 | // you may not use this file except in compliance with the License.
|
---|
| 5 | // You may obtain a copy of the License at
|
---|
| 6 | //
|
---|
| 7 | // http://www.apache.org/licenses/LICENSE-2.0
|
---|
| 8 | //
|
---|
| 9 | // Unless required by applicable law or agreed to in writing, software
|
---|
| 10 | // distributed under the License is distributed on an "AS IS" BASIS,
|
---|
| 11 | // WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
---|
| 12 | // See the License for the specific language governing permissions and
|
---|
| 13 | // limitations under the License.
|
---|
| 14 |
|
---|
| 15 | package de.ugoe.cs.autoquest.commands.usability;
|
---|
| 16 |
|
---|
| 17 | import java.text.DecimalFormat;
|
---|
| 18 | import java.util.ArrayList;
|
---|
| 19 | import java.util.Collections;
|
---|
| 20 | import java.util.HashMap;
|
---|
| 21 | import java.util.IdentityHashMap;
|
---|
| 22 | import java.util.LinkedList;
|
---|
| 23 | import java.util.List;
|
---|
| 24 | import java.util.Map;
|
---|
| 25 | import java.util.logging.Level;
|
---|
| 26 |
|
---|
| 27 | import de.ugoe.cs.autoquest.CommandHelpers;
|
---|
| 28 | import de.ugoe.cs.autoquest.eventcore.IEventTarget;
|
---|
| 29 | import de.ugoe.cs.autoquest.tasktrees.taskequality.TaskEquality;
|
---|
| 30 | import de.ugoe.cs.autoquest.tasktrees.taskequality.TaskEqualityRuleManager;
|
---|
| 31 | import de.ugoe.cs.autoquest.tasktrees.temporalrelation.TaskComparator;
|
---|
| 32 | import de.ugoe.cs.autoquest.tasktrees.temporalrelation.utils.SimilarTasks;
|
---|
| 33 | import de.ugoe.cs.autoquest.tasktrees.treeifc.DefaultTaskTraversingVisitor;
|
---|
| 34 | import de.ugoe.cs.autoquest.tasktrees.treeifc.IEventTask;
|
---|
| 35 | import de.ugoe.cs.autoquest.tasktrees.treeifc.IEventTaskInstance;
|
---|
| 36 | import de.ugoe.cs.autoquest.tasktrees.treeifc.ISelection;
|
---|
| 37 | import de.ugoe.cs.autoquest.tasktrees.treeifc.ISequence;
|
---|
| 38 | import de.ugoe.cs.autoquest.tasktrees.treeifc.ITask;
|
---|
| 39 | import de.ugoe.cs.autoquest.tasktrees.treeifc.ITaskModel;
|
---|
| 40 | import de.ugoe.cs.autoquest.tasktrees.treeifc.TaskMetric;
|
---|
| 41 | import de.ugoe.cs.util.StopWatch;
|
---|
| 42 | import de.ugoe.cs.util.console.Command;
|
---|
| 43 | import de.ugoe.cs.util.console.Console;
|
---|
| 44 | import de.ugoe.cs.util.console.GlobalDataContainer;
|
---|
| 45 |
|
---|
| 46 | /**
|
---|
| 47 | * <p>
|
---|
| 48 | * compares a set of task models with each other and drops their similarity as results
|
---|
| 49 | * </p>
|
---|
| 50 | *
|
---|
| 51 | * @author Patrick Harms
|
---|
| 52 | * @version 1.0
|
---|
| 53 | */
|
---|
| 54 | public class CMDgetTaskModelSimilarity implements Command {
|
---|
| 55 |
|
---|
| 56 | /*
|
---|
| 57 | * (non-Javadoc)
|
---|
| 58 | *
|
---|
| 59 | * @see de.ugoe.cs.util.console.Command#help()
|
---|
| 60 | */
|
---|
| 61 | @Override
|
---|
| 62 | public String help() {
|
---|
| 63 | return "getTaskModelSimilarity <tasktree1> <tasktree2> ...";
|
---|
| 64 | }
|
---|
| 65 |
|
---|
| 66 | /*
|
---|
| 67 | * (non-Javadoc)
|
---|
| 68 | *
|
---|
| 69 | * @see de.ugoe.cs.util.console.Command#run(java.util.List)
|
---|
| 70 | */
|
---|
| 71 | @Override
|
---|
| 72 | public void run(List<Object> parameters) {
|
---|
| 73 | List<String> inputTaskTreeNames = new LinkedList<>();
|
---|
| 74 |
|
---|
| 75 | try {
|
---|
| 76 | for (Object param : parameters) {
|
---|
| 77 | inputTaskTreeNames.add((String) param);
|
---|
| 78 | }
|
---|
| 79 | }
|
---|
| 80 | catch (Exception e) {
|
---|
| 81 | throw new IllegalArgumentException("must provide valid input task tree names");
|
---|
| 82 | }
|
---|
| 83 |
|
---|
| 84 | Map<ITaskModel, List<ISequence>> inputTaskModels = new IdentityHashMap<>();
|
---|
| 85 |
|
---|
| 86 | for (String inputTaskTreeName : inputTaskTreeNames) {
|
---|
| 87 | Object dataObject = GlobalDataContainer.getInstance().getData(inputTaskTreeName);
|
---|
| 88 | if (dataObject == null) {
|
---|
| 89 | CommandHelpers.objectNotFoundMessage(inputTaskTreeName);
|
---|
| 90 | return;
|
---|
| 91 | }
|
---|
| 92 | if (!(dataObject instanceof ITaskModel)) {
|
---|
| 93 | CommandHelpers.objectNotType(inputTaskTreeName, "ITaskModel");
|
---|
| 94 | return;
|
---|
| 95 | }
|
---|
| 96 |
|
---|
| 97 | List<ISequence> tasksToCompare = new LinkedList<>();
|
---|
| 98 |
|
---|
| 99 | for (ITask task : ((ITaskModel) dataObject).getTasks()) {
|
---|
| 100 | if (task instanceof ISequence) {
|
---|
| 101 | tasksToCompare.add((ISequence) task);
|
---|
| 102 | }
|
---|
| 103 | }
|
---|
| 104 |
|
---|
| 105 | inputTaskModels.put((ITaskModel) dataObject, tasksToCompare);
|
---|
| 106 | }
|
---|
| 107 |
|
---|
| 108 | SimilarityStatistics statistics = new SimilarityStatistics();
|
---|
| 109 |
|
---|
| 110 | getTaskModelSimilarity(inputTaskModels, statistics);
|
---|
| 111 |
|
---|
| 112 | Console.println("\nsimilarity statistics of all comparisons");
|
---|
| 113 | statistics.dump();
|
---|
| 114 | }
|
---|
| 115 |
|
---|
| 116 | /**
|
---|
| 117 | *
|
---|
| 118 | */
|
---|
| 119 | private void getTaskModelSimilarity(Map<ITaskModel, List<ISequence>> inputTaskModels,
|
---|
| 120 | SimilarityStatistics statistics)
|
---|
| 121 | {
|
---|
| 122 | // create the indexes to not do too many comparisons
|
---|
| 123 | Map<ITaskModel, Map<Integer, Map<IEventTarget, List<ISequence>>>> index1 =
|
---|
| 124 | new IdentityHashMap<>();
|
---|
| 125 |
|
---|
| 126 | Map<ISequence, Integer> index2 = new HashMap<>();
|
---|
| 127 | Map<ISequence, IEventTarget> index3 = new HashMap<>();
|
---|
| 128 |
|
---|
| 129 | final List<IEventTask> terminalNodes = new ArrayList<>();
|
---|
| 130 | final List<ISelection> selections = new ArrayList<>();
|
---|
| 131 |
|
---|
| 132 | for (Map.Entry<ITaskModel, List<ISequence>> entry : inputTaskModels.entrySet()) {
|
---|
| 133 | for (ISequence sequence : entry.getValue()) {
|
---|
| 134 | terminalNodes.clear();
|
---|
| 135 | selections.clear();
|
---|
| 136 |
|
---|
| 137 | sequence.accept(new DefaultTaskTraversingVisitor() {
|
---|
| 138 | @Override
|
---|
| 139 | public void visit(IEventTask eventTask) {
|
---|
| 140 | terminalNodes.add(eventTask);
|
---|
| 141 | }
|
---|
| 142 | @Override
|
---|
| 143 | public void visit(ISelection selection) {
|
---|
| 144 | selections.add(selection);
|
---|
| 145 | // no need to traverse children as a sequences containing a selection
|
---|
| 146 | // will be handled differently
|
---|
| 147 | }
|
---|
| 148 | });
|
---|
| 149 |
|
---|
| 150 | int length = terminalNodes.size();
|
---|
| 151 | IEventTarget firstTarget = ((IEventTaskInstance) terminalNodes.get(0).getInstances()
|
---|
| 152 | .iterator().next()).getEvent().getTarget();
|
---|
| 153 |
|
---|
| 154 | if (selections.size() > 0) {
|
---|
| 155 | length = -1;
|
---|
| 156 | firstTarget = null;
|
---|
| 157 | }
|
---|
| 158 |
|
---|
| 159 | Map<Integer, Map<IEventTarget, List<ISequence>>> lengthMap =
|
---|
| 160 | index1.get(entry.getKey());
|
---|
| 161 |
|
---|
| 162 | if (lengthMap == null) {
|
---|
| 163 | lengthMap = new HashMap<>();
|
---|
| 164 | index1.put(entry.getKey(), lengthMap);
|
---|
| 165 | }
|
---|
| 166 |
|
---|
| 167 | Map<IEventTarget, List<ISequence>> firstTaskMap = lengthMap.get(length);
|
---|
| 168 |
|
---|
| 169 | if (firstTaskMap == null) {
|
---|
| 170 | firstTaskMap = new HashMap<>();
|
---|
| 171 | lengthMap.put(length, firstTaskMap);
|
---|
| 172 | }
|
---|
| 173 |
|
---|
| 174 | List<ISequence> compareList = firstTaskMap.get(firstTarget);
|
---|
| 175 |
|
---|
| 176 | if (compareList == null) {
|
---|
| 177 | compareList = new LinkedList<ISequence>();
|
---|
| 178 | firstTaskMap.put(firstTarget, compareList);
|
---|
| 179 | }
|
---|
| 180 |
|
---|
| 181 | compareList.add(sequence);
|
---|
| 182 | index2.put(sequence, length);
|
---|
| 183 | index3.put(sequence, firstTarget);
|
---|
| 184 | }
|
---|
| 185 | }
|
---|
| 186 |
|
---|
| 187 | // create the comparison runnables
|
---|
| 188 | List<ComparisonRunnable> runnables = new LinkedList<>();
|
---|
| 189 |
|
---|
| 190 | for (Map.Entry<ITaskModel, List<ISequence>> model1 : inputTaskModels.entrySet()) {
|
---|
| 191 | for (Map.Entry<ITaskModel, List<ISequence>> model2 : inputTaskModels.entrySet()) {
|
---|
| 192 | if (model1.getKey() != model2.getKey()) {
|
---|
| 193 | runnables.add(new ComparisonRunnable(model1.getKey(), model1.getValue(),
|
---|
| 194 | model2.getKey(), model2.getValue(),
|
---|
| 195 | Collections.unmodifiableMap
|
---|
| 196 | (index1.get(model2.getKey())),
|
---|
| 197 | Collections.unmodifiableMap(index2),
|
---|
| 198 | Collections.unmodifiableMap(index3),
|
---|
| 199 | runnables));
|
---|
| 200 | }
|
---|
| 201 | }
|
---|
| 202 | }
|
---|
| 203 |
|
---|
| 204 | // execute the threads
|
---|
| 205 | Console.traceln(Level.FINE, "scheduling " + runnables.size() + " comparison threads");
|
---|
| 206 |
|
---|
| 207 | synchronized (runnables) {
|
---|
| 208 | int startedRunnables = 0;
|
---|
| 209 | for (ComparisonRunnable runnable : runnables) {
|
---|
| 210 | while (startedRunnables >= Math.max(1, Runtime.getRuntime().availableProcessors()))
|
---|
| 211 | {
|
---|
| 212 | try {
|
---|
| 213 | Console.traceln(Level.FINER, "waiting for next thread to finish");
|
---|
| 214 | runnables.wait();
|
---|
| 215 | startedRunnables--;
|
---|
| 216 | Console.traceln(Level.FINER, "thread finished");
|
---|
| 217 | }
|
---|
| 218 | catch (InterruptedException e) {
|
---|
| 219 | // should not happen
|
---|
| 220 | Console.logException(e);
|
---|
| 221 | }
|
---|
| 222 | }
|
---|
| 223 |
|
---|
| 224 | Console.traceln(Level.FINER, "starting next thread");
|
---|
| 225 | startedRunnables++;
|
---|
| 226 | Console.traceln(Level.FINE, "comparing " + runnable.tasks1.size()
|
---|
| 227 | + " tasks of model " + runnable.model1 + " with " +
|
---|
| 228 | runnable.tasks2.size() + " tasks of model " + runnable.model2);
|
---|
| 229 |
|
---|
| 230 | new Thread(runnable).start();
|
---|
| 231 | Console.traceln(Level.FINER, "started next thread " + runnable);
|
---|
| 232 | }
|
---|
| 233 |
|
---|
| 234 | while (startedRunnables > 0) {
|
---|
| 235 | try {
|
---|
| 236 | Console.traceln(Level.FINER, "waiting for next thread to finish");
|
---|
| 237 | runnables.wait();
|
---|
| 238 | startedRunnables--;
|
---|
| 239 | Console.traceln(Level.FINER, "thread finished");
|
---|
| 240 | }
|
---|
| 241 | catch (InterruptedException e) {
|
---|
| 242 | // should not happen
|
---|
| 243 | Console.logException(e);
|
---|
| 244 | }
|
---|
| 245 | }
|
---|
| 246 | }
|
---|
| 247 |
|
---|
| 248 | // merge the results
|
---|
| 249 | Console.traceln(Level.FINER, "all threads finished, mergin results");
|
---|
| 250 |
|
---|
| 251 | for (ComparisonRunnable runnable : runnables) {
|
---|
| 252 | Console.println("\n similarity statistics of comparison between " + runnable.model1 +
|
---|
| 253 | " and " + runnable.model2);
|
---|
| 254 | runnable.getStatistics().dump();
|
---|
| 255 | statistics.mergeWith(runnable.getStatistics());
|
---|
| 256 | }
|
---|
| 257 | }
|
---|
| 258 |
|
---|
| 259 | /**
|
---|
| 260 | *
|
---|
| 261 | */
|
---|
| 262 | private static class SimilarityStatistics {
|
---|
| 263 |
|
---|
| 264 | /** */
|
---|
| 265 | private int taskCounter = 0;
|
---|
| 266 |
|
---|
| 267 | /** */
|
---|
| 268 | private int maxCoverage = 0;
|
---|
| 269 |
|
---|
| 270 | /** */
|
---|
| 271 | private Map<Integer, Integer> coverageCounters = new HashMap<>();
|
---|
| 272 |
|
---|
| 273 | /** */
|
---|
| 274 | private Map<Integer, Map<Integer, Map<TaskEquality, Integer>>> equalityCounters =
|
---|
| 275 | new HashMap<>();
|
---|
| 276 |
|
---|
| 277 | /**
|
---|
| 278 | *
|
---|
| 279 | */
|
---|
| 280 | private void store(ITask task,
|
---|
| 281 | ITaskModel model,
|
---|
| 282 | ITask other,
|
---|
| 283 | ITaskModel otherModel,
|
---|
| 284 | TaskEquality equality)
|
---|
| 285 | {
|
---|
| 286 | int coverageRatio1 =
|
---|
| 287 | model.getTaskInfo(task).getMeasureValue(TaskMetric.EVENT_COVERAGE);
|
---|
| 288 | int coverageRatio2 =
|
---|
| 289 | otherModel.getTaskInfo(other).getMeasureValue(TaskMetric.EVENT_COVERAGE);
|
---|
| 290 |
|
---|
| 291 | addToMaps(coverageRatio1, 1);
|
---|
| 292 | addToMaps(coverageRatio1, coverageRatio2, equality, 1);
|
---|
| 293 | taskCounter++;
|
---|
| 294 | maxCoverage = Math.max(maxCoverage, coverageRatio1);
|
---|
| 295 | }
|
---|
| 296 |
|
---|
| 297 | /**
|
---|
| 298 | *
|
---|
| 299 | */
|
---|
| 300 | private void store(ITask task, ITaskModel model) {
|
---|
| 301 | int coverageRatio1 =
|
---|
| 302 | model.getTaskInfo(task).getMeasureValue(TaskMetric.EVENT_COVERAGE);
|
---|
| 303 |
|
---|
| 304 | addToMaps(coverageRatio1, 1);
|
---|
| 305 | addToMaps(coverageRatio1, 0, TaskEquality.UNEQUAL, 1);
|
---|
| 306 | taskCounter++;
|
---|
| 307 | maxCoverage = Math.max(maxCoverage, coverageRatio1);
|
---|
| 308 | }
|
---|
| 309 |
|
---|
| 310 | /**
|
---|
| 311 | *
|
---|
| 312 | */
|
---|
| 313 | private void mergeWith(SimilarityStatistics statistics) {
|
---|
| 314 | for (Map.Entry<Integer, Map<Integer, Map<TaskEquality, Integer>>> entry1 :
|
---|
| 315 | statistics.equalityCounters.entrySet())
|
---|
| 316 | {
|
---|
| 317 | for (Map.Entry<Integer, Map<TaskEquality, Integer>> entry2 :
|
---|
| 318 | entry1.getValue().entrySet())
|
---|
| 319 | {
|
---|
| 320 | for (Map.Entry<TaskEquality, Integer> entry3 : entry2.getValue().entrySet()) {
|
---|
| 321 | addToMaps
|
---|
| 322 | (entry1.getKey(), entry2.getKey(), entry3.getKey(), entry3.getValue());
|
---|
| 323 | }
|
---|
| 324 | }
|
---|
| 325 | }
|
---|
| 326 |
|
---|
| 327 | for (Map.Entry<Integer, Integer> entry : statistics.coverageCounters.entrySet()) {
|
---|
| 328 | addToMaps(entry.getKey(), entry.getValue());
|
---|
| 329 | }
|
---|
| 330 |
|
---|
| 331 | taskCounter += statistics.taskCounter;
|
---|
| 332 | maxCoverage = Math.max(maxCoverage, statistics.taskCounter);
|
---|
| 333 | }
|
---|
| 334 |
|
---|
| 335 | /**
|
---|
| 336 | *
|
---|
| 337 | */
|
---|
| 338 | private void addToMaps(int ratio, int increment) {
|
---|
| 339 | Integer counter = coverageCounters.get(ratio);
|
---|
| 340 |
|
---|
| 341 | if (counter == null) {
|
---|
| 342 | coverageCounters.put(ratio, increment);
|
---|
| 343 | }
|
---|
| 344 | else {
|
---|
| 345 | coverageCounters.put(ratio, counter + increment);
|
---|
| 346 | }
|
---|
| 347 | }
|
---|
| 348 |
|
---|
| 349 | /**
|
---|
| 350 | *
|
---|
| 351 | */
|
---|
| 352 | private void addToMaps(Integer ratio1, Integer ratio2, TaskEquality equality, int value) {
|
---|
| 353 | Map<Integer, Map<TaskEquality, Integer>> counters1 = equalityCounters.get(ratio1);
|
---|
| 354 |
|
---|
| 355 | if (counters1 == null) {
|
---|
| 356 | counters1 = new HashMap<>();
|
---|
| 357 | equalityCounters.put(ratio1, counters1);
|
---|
| 358 | }
|
---|
| 359 |
|
---|
| 360 | Map<TaskEquality, Integer> counters2 = counters1.get(ratio2);
|
---|
| 361 |
|
---|
| 362 | if (counters2 == null) {
|
---|
| 363 | counters2 = new HashMap<>();
|
---|
| 364 | counters1.put(ratio2, counters2);
|
---|
| 365 | }
|
---|
| 366 |
|
---|
| 367 | Integer counter = counters2.get(equality);
|
---|
| 368 |
|
---|
| 369 | if (counter == null) {
|
---|
| 370 | counters2.put(equality, value);
|
---|
| 371 | }
|
---|
| 372 | else {
|
---|
| 373 | counters2.put(equality, counter + value);
|
---|
| 374 | }
|
---|
| 375 | }
|
---|
| 376 |
|
---|
| 377 | /**
|
---|
| 378 | *
|
---|
| 379 | */
|
---|
| 380 | private void dump() {
|
---|
| 381 | Console.println("Statistics of Similarity");
|
---|
| 382 | Console.println("========================");
|
---|
| 383 |
|
---|
| 384 | int[][] bins = getBinsOfAlmostEqualSize();
|
---|
| 385 | int lowerBorder1;
|
---|
| 386 | int higherBorder1;
|
---|
| 387 | int lowerBorder2;
|
---|
| 388 | int higherBorder2;
|
---|
| 389 | int allEqualitiesOfBin = 0;
|
---|
| 390 | int allEqualities = 0;
|
---|
| 391 |
|
---|
| 392 | for (int i = bins.length - 1; i >= 0 ; i--) {
|
---|
| 393 | if (i <= 0) {
|
---|
| 394 | lowerBorder1 = 0;
|
---|
| 395 | }
|
---|
| 396 | else {
|
---|
| 397 | lowerBorder1 = bins[i - 1][0] + 1;
|
---|
| 398 | }
|
---|
| 399 |
|
---|
| 400 | higherBorder1 = bins[i][0];
|
---|
| 401 |
|
---|
| 402 | allEqualitiesOfBin = 0;
|
---|
| 403 |
|
---|
| 404 | Console.println("similarities of " + bins[i][1] + " tasks with " + lowerBorder1 +
|
---|
| 405 | " to " + higherBorder1 + "‰ coverage");
|
---|
| 406 |
|
---|
| 407 | for (int j = bins.length - 1; j >= 0; j--) {
|
---|
| 408 | if (j <= 0) {
|
---|
| 409 | lowerBorder2 = 0;
|
---|
| 410 | }
|
---|
| 411 | else {
|
---|
| 412 | lowerBorder2 = bins[j - 1][0] + 1;
|
---|
| 413 | }
|
---|
| 414 |
|
---|
| 415 | higherBorder2 = bins[j][0];
|
---|
| 416 |
|
---|
| 417 | Console.print(" --> to " + bins[j][1] + " tasks with " + lowerBorder2 +
|
---|
| 418 | " to " + higherBorder2 + "‰ coverage");
|
---|
| 419 |
|
---|
| 420 | Console.print(" | ");
|
---|
| 421 | int allInBin = countAllInBin(lowerBorder1, higherBorder1);
|
---|
| 422 |
|
---|
| 423 | int count1 = countEqualities(lowerBorder1, higherBorder1, lowerBorder2,
|
---|
| 424 | higherBorder2, TaskEquality.LEXICALLY_EQUAL);
|
---|
| 425 |
|
---|
| 426 | dump(count1, allInBin, "LEX");
|
---|
| 427 | Console.print("\t| ");
|
---|
| 428 |
|
---|
| 429 | int count2 = countEqualities(lowerBorder1, higherBorder1, lowerBorder2,
|
---|
| 430 | higherBorder2, TaskEquality.SYNTACTICALLY_EQUAL);
|
---|
| 431 |
|
---|
| 432 | dump(count2, allInBin, "SYN");
|
---|
| 433 | Console.print("\t| ");
|
---|
| 434 |
|
---|
| 435 | int count3 = countEqualities(lowerBorder1, higherBorder1, lowerBorder2,
|
---|
| 436 | higherBorder2, TaskEquality.SEMANTICALLY_EQUAL);
|
---|
| 437 |
|
---|
| 438 | dump(count3, allInBin, "SEM");
|
---|
| 439 | Console.print("\t|| ");
|
---|
| 440 |
|
---|
| 441 | int count4 = countEqualities(lowerBorder1, higherBorder1, lowerBorder2,
|
---|
| 442 | higherBorder2, null);
|
---|
| 443 |
|
---|
| 444 | dump(count4, allInBin, "SIM");
|
---|
| 445 | Console.print("\t|| ");
|
---|
| 446 |
|
---|
| 447 | dump(count1 + count2 + count3 + count4, allInBin, "ALL");
|
---|
| 448 | Console.println("");
|
---|
| 449 |
|
---|
| 450 | allEqualitiesOfBin += count1 + count2 + count3 + count4;
|
---|
| 451 | }
|
---|
| 452 |
|
---|
| 453 | Console.print(" --> to all other tasks\t");
|
---|
| 454 | dump(allEqualitiesOfBin, taskCounter, "ALL");
|
---|
| 455 | Console.println("");
|
---|
| 456 |
|
---|
| 457 | allEqualities += allEqualitiesOfBin;
|
---|
| 458 | }
|
---|
| 459 |
|
---|
| 460 | Console.println("");
|
---|
| 461 | Console.print("complete recall is ");
|
---|
| 462 | dump(allEqualities, taskCounter, "ALL");
|
---|
| 463 | Console.println("");
|
---|
| 464 | }
|
---|
| 465 |
|
---|
| 466 | /**
|
---|
| 467 | *
|
---|
| 468 | */
|
---|
| 469 | private int[][] getBinsOfAlmostEqualSize() {
|
---|
| 470 | int[][] result = new int[5][];
|
---|
| 471 |
|
---|
| 472 | for (int i = 0; i < result.length; i++) {
|
---|
| 473 | result[i] = new int[2];
|
---|
| 474 | }
|
---|
| 475 |
|
---|
| 476 | int averageBinSize = taskCounter / result.length;
|
---|
| 477 | int aimedBinSize = averageBinSize;
|
---|
| 478 |
|
---|
| 479 | int index = 0;
|
---|
| 480 |
|
---|
| 481 | for (int i = 0; i < maxCoverage; i++) {
|
---|
| 482 | if (result[index][1] > aimedBinSize) {
|
---|
| 483 | // try to compensate, if the previous bin was a little too large
|
---|
| 484 | aimedBinSize = averageBinSize + averageBinSize - result[index][1];
|
---|
| 485 | index++;
|
---|
| 486 | }
|
---|
| 487 |
|
---|
| 488 | Integer value = coverageCounters.get(i);
|
---|
| 489 |
|
---|
| 490 | if (value != null) {
|
---|
| 491 | result[index][0] = i;
|
---|
| 492 | result[index][1] += value;
|
---|
| 493 | }
|
---|
| 494 | }
|
---|
| 495 |
|
---|
| 496 | return result;
|
---|
| 497 | }
|
---|
| 498 |
|
---|
| 499 | /**
|
---|
| 500 | *
|
---|
| 501 | */
|
---|
| 502 | private void dump(int noOfEqualTasks, int noOfAllTasks, String prefix) {
|
---|
| 503 | Console.print(prefix);
|
---|
| 504 | Console.print("\t: ");
|
---|
| 505 |
|
---|
| 506 | if (noOfAllTasks > 0) {
|
---|
| 507 | double value = ((double) (noOfEqualTasks * 100)) / noOfAllTasks;
|
---|
| 508 | Console.print(noOfEqualTasks + " (");
|
---|
| 509 | Console.print(new DecimalFormat("###.#").format(value));
|
---|
| 510 | Console.print("%)");
|
---|
| 511 | }
|
---|
| 512 | else {
|
---|
| 513 | Console.print("n.a.");
|
---|
| 514 | }
|
---|
| 515 | }
|
---|
| 516 |
|
---|
| 517 | /**
|
---|
| 518 | *
|
---|
| 519 | */
|
---|
| 520 | private int countEqualities(int lowerBorder1,
|
---|
| 521 | int higherBorder1,
|
---|
| 522 | int lowerBorder2,
|
---|
| 523 | int higherBorder2,
|
---|
| 524 | TaskEquality equality)
|
---|
| 525 | {
|
---|
| 526 | int counter = 0;
|
---|
| 527 |
|
---|
| 528 | for (int i = lowerBorder1; i <= higherBorder1; i++) {
|
---|
| 529 | Map<Integer, Map<TaskEquality, Integer>> counters1 = equalityCounters.get(i);
|
---|
| 530 |
|
---|
| 531 | if (counters1 != null) {
|
---|
| 532 | for (int j = lowerBorder2; j <= higherBorder2; j++) {
|
---|
| 533 | Map<TaskEquality, Integer> counters2 = counters1.get(j);
|
---|
| 534 |
|
---|
| 535 | if (counters2 != null) {
|
---|
| 536 | Integer counterObj = counters2.get(equality);
|
---|
| 537 |
|
---|
| 538 | if (counterObj != null) {
|
---|
| 539 | counter += counterObj;
|
---|
| 540 | }
|
---|
| 541 | }
|
---|
| 542 | }
|
---|
| 543 | }
|
---|
| 544 | }
|
---|
| 545 |
|
---|
| 546 | return counter;
|
---|
| 547 | }
|
---|
| 548 |
|
---|
| 549 | /**
|
---|
| 550 | *
|
---|
| 551 | */
|
---|
| 552 | private int countAllInBin(int lowerBorder1, int higherBorder1) {
|
---|
| 553 | int coverageCounter = 0;
|
---|
| 554 |
|
---|
| 555 | for (int i = lowerBorder1; i <= higherBorder1; i++) {
|
---|
| 556 | Integer value = coverageCounters.get(i);
|
---|
| 557 |
|
---|
| 558 | if (value != null) {
|
---|
| 559 | coverageCounter += value;
|
---|
| 560 | }
|
---|
| 561 | }
|
---|
| 562 |
|
---|
| 563 | return coverageCounter;
|
---|
| 564 | }
|
---|
| 565 | }
|
---|
| 566 |
|
---|
| 567 | /**
|
---|
| 568 | *
|
---|
| 569 | */
|
---|
| 570 | private static class ComparisonRunnable implements Runnable {
|
---|
| 571 |
|
---|
| 572 | /** */
|
---|
| 573 | private ITaskModel model1;
|
---|
| 574 |
|
---|
| 575 | /** */
|
---|
| 576 | private ITaskModel model2;
|
---|
| 577 |
|
---|
| 578 | /** */
|
---|
| 579 | private List<ISequence> tasks1;
|
---|
| 580 |
|
---|
| 581 | /** */
|
---|
| 582 | private List<ISequence> tasks2;
|
---|
| 583 |
|
---|
| 584 | /** */
|
---|
| 585 | private Map<Integer, Map<IEventTarget, List<ISequence>>> lengthIndex1;
|
---|
| 586 |
|
---|
| 587 | /** */
|
---|
| 588 | private Map<ISequence, Integer> lengthIndex2;
|
---|
| 589 |
|
---|
| 590 | /** */
|
---|
| 591 | private Map<ISequence, IEventTarget> firstTargetIndex;
|
---|
| 592 |
|
---|
| 593 | /** */
|
---|
| 594 | private Object semaphore;
|
---|
| 595 |
|
---|
| 596 | /** */
|
---|
| 597 | private SimilarityStatistics statistics = new SimilarityStatistics();
|
---|
| 598 |
|
---|
| 599 | /**
|
---|
| 600 | *
|
---|
| 601 | */
|
---|
| 602 | private ComparisonRunnable(ITaskModel model1,
|
---|
| 603 | List<ISequence> tasks1,
|
---|
| 604 | ITaskModel model2,
|
---|
| 605 | List<ISequence> tasks2,
|
---|
| 606 | Map<Integer, Map<IEventTarget, List<ISequence>>> lengthIndex1,
|
---|
| 607 | Map<ISequence, Integer> lengthIndex2,
|
---|
| 608 | Map<ISequence, IEventTarget> firstTargetIndex,
|
---|
| 609 | Object semaphore)
|
---|
| 610 | {
|
---|
| 611 | this.model1 = model1;
|
---|
| 612 | this.tasks1 = tasks1;
|
---|
| 613 | this.model2 = model2;
|
---|
| 614 | this.tasks2 = tasks2;
|
---|
| 615 | this.lengthIndex1 = lengthIndex1;
|
---|
| 616 | this.lengthIndex2 = lengthIndex2;
|
---|
| 617 | this.firstTargetIndex = firstTargetIndex;
|
---|
| 618 | this.semaphore = semaphore;
|
---|
| 619 | }
|
---|
| 620 |
|
---|
| 621 | /* (non-Javadoc)
|
---|
| 622 | * @see java.lang.Runnable#run()
|
---|
| 623 | */
|
---|
| 624 | @Override
|
---|
| 625 | public void run() {
|
---|
| 626 | TaskEqualityRuleManager equalityRuleManager = TaskEqualityRuleManager.getInstance();
|
---|
| 627 | TaskComparator comparator = new TaskComparator(TaskEquality.SEMANTICALLY_EQUAL);
|
---|
| 628 | TaskEquality mostCommonEquality;
|
---|
| 629 | ITask mostSimilarTask;
|
---|
| 630 | TaskEquality equality;
|
---|
| 631 | SimilarTasks similarity;
|
---|
| 632 |
|
---|
| 633 | List<ISequence> tasksToCompareWith = new ArrayList<ISequence>();
|
---|
| 634 | StopWatch watch = new StopWatch();
|
---|
| 635 | watch.start("all comparisons ");
|
---|
| 636 | int count = 0;
|
---|
| 637 | for (ISequence task : tasks1) {
|
---|
| 638 | int length = lengthIndex2.get(task);
|
---|
| 639 | IEventTarget firstTarget = firstTargetIndex.get(task);
|
---|
| 640 | tasksToCompareWith.clear();
|
---|
| 641 |
|
---|
| 642 | // add tasks with same length
|
---|
| 643 | Map<IEventTarget, List<ISequence>> eventTaskMap = lengthIndex1.get(length);
|
---|
| 644 | List<ISequence> toCompare;
|
---|
| 645 |
|
---|
| 646 | if (eventTaskMap != null) {
|
---|
| 647 | toCompare = eventTaskMap.get(firstTarget);
|
---|
| 648 |
|
---|
| 649 | if (toCompare != null) {
|
---|
| 650 | tasksToCompareWith.addAll(toCompare);
|
---|
| 651 | }
|
---|
| 652 | }
|
---|
| 653 |
|
---|
| 654 | // add tasks containing selections
|
---|
| 655 | eventTaskMap = lengthIndex1.get(-1);
|
---|
| 656 |
|
---|
| 657 | if (eventTaskMap != null) {
|
---|
| 658 | toCompare = eventTaskMap.get(firstTarget);
|
---|
| 659 |
|
---|
| 660 | if (toCompare != null) {
|
---|
| 661 | tasksToCompareWith.addAll(toCompare);
|
---|
| 662 | }
|
---|
| 663 | }
|
---|
| 664 |
|
---|
| 665 | mostCommonEquality = null;
|
---|
| 666 | mostSimilarTask = null;
|
---|
| 667 |
|
---|
| 668 | for (ITask taskToCompare : tasksToCompareWith) {
|
---|
| 669 | count++;
|
---|
| 670 | watch.start("normal comparison");
|
---|
| 671 | equality = equalityRuleManager.compare(task, taskToCompare);
|
---|
| 672 | watch.stop("normal comparison");
|
---|
| 673 |
|
---|
| 674 | if ((equality != TaskEquality.UNEQUAL) &&
|
---|
| 675 | ((mostCommonEquality == null) || (equality.isAtLeast(mostCommonEquality))))
|
---|
| 676 | {
|
---|
| 677 | // >>>>>>>>>>>>>>>>>>>>>> TEST IMPLEMENTATION >>>>>>>>>>>>>>>>>>>>>>>>>>>
|
---|
| 678 | // synchronized (System.out) {
|
---|
| 679 | // System.out.println("found equal tasks: " + equality);
|
---|
| 680 | // new TaskTreeEncoder().encode(task, System.out);
|
---|
| 681 | // new TaskTreeEncoder().encode(taskToCompare, System.out);
|
---|
| 682 | // }
|
---|
| 683 | // <<<<<<<<<<<<<<<<<<<<<< TEST IMPLEMENTATION <<<<<<<<<<<<<<<<<<<<<<<<<<<
|
---|
| 684 |
|
---|
| 685 | mostCommonEquality = equality;
|
---|
| 686 | mostSimilarTask = taskToCompare;
|
---|
| 687 |
|
---|
| 688 | if (mostCommonEquality.isAtLeast(TaskEquality.LEXICALLY_EQUAL)) {
|
---|
| 689 | // we found a lexically equal match, the most concrete that
|
---|
| 690 | // we can find. We can break up here
|
---|
| 691 | break;
|
---|
| 692 | }
|
---|
| 693 | }
|
---|
| 694 | }
|
---|
| 695 |
|
---|
| 696 | if (mostCommonEquality != null) {
|
---|
| 697 | statistics.store(task, model1, mostSimilarTask, model2, mostCommonEquality);
|
---|
| 698 | }
|
---|
| 699 | else {
|
---|
| 700 | int lowestDiffLevel = Integer.MAX_VALUE;
|
---|
| 701 | for (ITask taskToCompare : tasksToCompareWith) {
|
---|
| 702 | count++;
|
---|
| 703 | watch.start("similarity comparison");
|
---|
| 704 | similarity = SimilarTasks.compareTasks(task, taskToCompare, comparator);
|
---|
| 705 | watch.stop("similarity comparison");
|
---|
| 706 |
|
---|
| 707 | if (similarity.getDiffLevel() < lowestDiffLevel) {
|
---|
| 708 | lowestDiffLevel = similarity.getDiffLevel();
|
---|
| 709 | mostSimilarTask = taskToCompare;
|
---|
| 710 |
|
---|
| 711 | if (lowestDiffLevel <= 0) {
|
---|
| 712 | // >>>>>>>>>>>>>>>>>>>>>> TEST IMPLEMENTATION >>>>>>>>>>>>>>>>>>>>>
|
---|
| 713 | // synchronized (System.out) {
|
---|
| 714 | // System.out.println("found similar tasks");
|
---|
| 715 | // similarity.dump(System.out);
|
---|
| 716 | // }
|
---|
| 717 | // <<<<<<<<<<<<<<<<<<<<<< TEST IMPLEMENTATION <<<<<<<<<<<<<<<<<<<<<
|
---|
| 718 |
|
---|
| 719 | // we found a fully similar task. We will not find any more
|
---|
| 720 | // similar, hence we break
|
---|
| 721 | break;
|
---|
| 722 | }
|
---|
| 723 | }
|
---|
| 724 | }
|
---|
| 725 |
|
---|
| 726 | if (lowestDiffLevel <= 0) {
|
---|
| 727 | statistics.store(task, model1, mostSimilarTask, model2, null);
|
---|
| 728 | }
|
---|
| 729 | else {
|
---|
| 730 | statistics.store(task, model1);
|
---|
| 731 | }
|
---|
| 732 | }
|
---|
| 733 | }
|
---|
| 734 |
|
---|
| 735 | System.out.println("performed " + count + " comparisons");
|
---|
| 736 |
|
---|
| 737 | watch.stop("all comparisons ");
|
---|
| 738 | watch.dumpStatistics(System.out);
|
---|
| 739 |
|
---|
| 740 | synchronized (semaphore) {
|
---|
| 741 | semaphore.notify();
|
---|
| 742 | }
|
---|
| 743 | }
|
---|
| 744 |
|
---|
| 745 | /**
|
---|
| 746 | * @return the statistics
|
---|
| 747 | */
|
---|
| 748 | private SimilarityStatistics getStatistics() {
|
---|
| 749 | return statistics;
|
---|
| 750 | }
|
---|
| 751 |
|
---|
| 752 | }
|
---|
| 753 | }
|
---|