1 /* 2 * Copyright (C) 2011 The Android Open Source Project 3 * 4 * Licensed under the Apache License, Version 2.0 (the "License"); you may not 5 * use this file except in compliance with the License. You may obtain a copy of 6 * the License at 7 * 8 * http://www.apache.org/licenses/LICENSE-2.0 9 * 10 * Unless required by applicable law or agreed to in writing, software 11 * distributed under the License is distributed on an "AS IS" BASIS, WITHOUT 12 * WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the 13 * License for the specific language governing permissions and limitations under 14 * the License. 15 */ 16 17 package com.android.inputmethod.latin.makedict; 18 19 import java.util.ArrayList; 20 import java.util.Arrays; 21 import java.util.Collections; 22 import java.util.HashMap; 23 import java.util.Iterator; 24 import java.util.LinkedList; 25 26 /** 27 * A dictionary that can fusion heads and tails of words for more compression. 28 */ 29 public class FusionDictionary implements Iterable<Word> { 30 31 private static final boolean DBG = MakedictLog.DBG; 32 33 /** 34 * A node of the dictionary, containing several CharGroups. 35 * 36 * A node is but an ordered array of CharGroups, which essentially contain all the 37 * real information. 38 * This class also contains fields to cache size and address, to help with binary 39 * generation. 40 */ 41 public static class Node { 42 ArrayList<CharGroup> mData; 43 // To help with binary generation 44 int mCachedSize; 45 int mCachedAddress; 46 public Node() { 47 mData = new ArrayList<CharGroup>(); 48 mCachedSize = Integer.MIN_VALUE; 49 mCachedAddress = Integer.MIN_VALUE; 50 } 51 public Node(ArrayList<CharGroup> data) { 52 mData = data; 53 mCachedSize = Integer.MIN_VALUE; 54 mCachedAddress = Integer.MIN_VALUE; 55 } 56 } 57 58 /** 59 * A string with a frequency. 60 * 61 * This represents an "attribute", that is either a bigram or a shortcut. 62 */ 63 public static class WeightedString { 64 final String mWord; 65 int mFrequency; 66 public WeightedString(String word, int frequency) { 67 mWord = word; 68 mFrequency = frequency; 69 } 70 71 @Override 72 public int hashCode() { 73 return Arrays.hashCode(new Object[] { mWord, mFrequency }); 74 } 75 76 @Override 77 public boolean equals(Object o) { 78 if (o == this) return true; 79 if (!(o instanceof WeightedString)) return false; 80 WeightedString w = (WeightedString)o; 81 return mWord.equals(w.mWord) && mFrequency == w.mFrequency; 82 } 83 } 84 85 /** 86 * A group of characters, with a frequency, shortcut targets, bigrams, and children. 87 * 88 * This is the central class of the in-memory representation. A CharGroup is what can 89 * be seen as a traditional "trie node", except it can hold several characters at the 90 * same time. A CharGroup essentially represents one or several characters in the middle 91 * of the trie trie; as such, it can be a terminal, and it can have children. 92 * In this in-memory representation, whether the CharGroup is a terminal or not is represented 93 * in the frequency, where NOT_A_TERMINAL (= -1) means this is not a terminal and any other 94 * value is the frequency of this terminal. A terminal may have non-null shortcuts and/or 95 * bigrams, but a non-terminal may not. Moreover, children, if present, are null. 96 */ 97 public static class CharGroup { 98 public static final int NOT_A_TERMINAL = -1; 99 final int mChars[]; 100 ArrayList<WeightedString> mShortcutTargets; 101 ArrayList<WeightedString> mBigrams; 102 int mFrequency; // NOT_A_TERMINAL == mFrequency indicates this is not a terminal. 103 Node mChildren; 104 // The two following members to help with binary generation 105 int mCachedSize; 106 int mCachedAddress; 107 108 public CharGroup(final int[] chars, final ArrayList<WeightedString> shortcutTargets, 109 final ArrayList<WeightedString> bigrams, final int frequency) { 110 mChars = chars; 111 mFrequency = frequency; 112 mShortcutTargets = shortcutTargets; 113 mBigrams = bigrams; 114 mChildren = null; 115 } 116 117 public CharGroup(final int[] chars, final ArrayList<WeightedString> shortcutTargets, 118 final ArrayList<WeightedString> bigrams, final int frequency, final Node children) { 119 mChars = chars; 120 mFrequency = frequency; 121 mShortcutTargets = shortcutTargets; 122 mBigrams = bigrams; 123 mChildren = children; 124 } 125 126 public void addChild(CharGroup n) { 127 if (null == mChildren) { 128 mChildren = new Node(); 129 } 130 mChildren.mData.add(n); 131 } 132 133 public boolean isTerminal() { 134 return NOT_A_TERMINAL != mFrequency; 135 } 136 137 public boolean hasSeveralChars() { 138 assert(mChars.length > 0); 139 return 1 < mChars.length; 140 } 141 142 /** 143 * Adds a word to the bigram list. Updates the frequency if the word already 144 * exists. 145 */ 146 public void addBigram(final String word, final int frequency) { 147 if (mBigrams == null) { 148 mBigrams = new ArrayList<WeightedString>(); 149 } 150 WeightedString bigram = getBigram(word); 151 if (bigram != null) { 152 bigram.mFrequency = frequency; 153 } else { 154 bigram = new WeightedString(word, frequency); 155 mBigrams.add(bigram); 156 } 157 } 158 159 /** 160 * Gets the shortcut target for the given word. Returns null if the word is not in the 161 * shortcut list. 162 */ 163 public WeightedString getShortcut(final String word) { 164 // TODO: Don't do a linear search 165 if (mShortcutTargets != null) { 166 final int size = mShortcutTargets.size(); 167 for (int i = 0; i < size; ++i) { 168 WeightedString shortcut = mShortcutTargets.get(i); 169 if (shortcut.mWord.equals(word)) { 170 return shortcut; 171 } 172 } 173 } 174 return null; 175 } 176 177 /** 178 * Gets the bigram for the given word. 179 * Returns null if the word is not in the bigrams list. 180 */ 181 public WeightedString getBigram(final String word) { 182 // TODO: Don't do a linear search 183 if (mBigrams != null) { 184 final int size = mBigrams.size(); 185 for (int i = 0; i < size; ++i) { 186 WeightedString bigram = mBigrams.get(i); 187 if (bigram.mWord.equals(word)) { 188 return bigram; 189 } 190 } 191 } 192 return null; 193 } 194 195 /** 196 * Updates the CharGroup with the given properties. Adds the shortcut and bigram lists to 197 * the existing ones if any. Note: unigram, bigram, and shortcut frequencies are only 198 * updated if they are higher than the existing ones. 199 */ 200 public void update(int frequency, ArrayList<WeightedString> shortcutTargets, 201 ArrayList<WeightedString> bigrams) { 202 if (frequency > mFrequency) { 203 mFrequency = frequency; 204 } 205 if (shortcutTargets != null) { 206 if (mShortcutTargets == null) { 207 mShortcutTargets = shortcutTargets; 208 } else { 209 final int size = shortcutTargets.size(); 210 for (int i = 0; i < size; ++i) { 211 final WeightedString shortcut = shortcutTargets.get(i); 212 final WeightedString existingShortcut = getShortcut(shortcut.mWord); 213 if (existingShortcut == null) { 214 mShortcutTargets.add(shortcut); 215 } else if (existingShortcut.mFrequency < shortcut.mFrequency) { 216 existingShortcut.mFrequency = shortcut.mFrequency; 217 } 218 } 219 } 220 } 221 if (bigrams != null) { 222 if (mBigrams == null) { 223 mBigrams = bigrams; 224 } else { 225 final int size = bigrams.size(); 226 for (int i = 0; i < size; ++i) { 227 final WeightedString bigram = bigrams.get(i); 228 final WeightedString existingBigram = getBigram(bigram.mWord); 229 if (existingBigram == null) { 230 mBigrams.add(bigram); 231 } else if (existingBigram.mFrequency < bigram.mFrequency) { 232 existingBigram.mFrequency = bigram.mFrequency; 233 } 234 } 235 } 236 } 237 } 238 } 239 240 /** 241 * Options global to the dictionary. 242 * 243 * There are no options at the moment, so this class is empty. 244 */ 245 public static class DictionaryOptions { 246 public final boolean mGermanUmlautProcessing; 247 public final boolean mFrenchLigatureProcessing; 248 public final HashMap<String, String> mAttributes; 249 public DictionaryOptions(final HashMap<String, String> attributes, 250 final boolean germanUmlautProcessing, final boolean frenchLigatureProcessing) { 251 mAttributes = attributes; 252 mGermanUmlautProcessing = germanUmlautProcessing; 253 mFrenchLigatureProcessing = frenchLigatureProcessing; 254 } 255 } 256 257 public final DictionaryOptions mOptions; 258 public final Node mRoot; 259 260 public FusionDictionary(final Node root, final DictionaryOptions options) { 261 mRoot = root; 262 mOptions = options; 263 } 264 265 public void addOptionAttribute(final String key, final String value) { 266 mOptions.mAttributes.put(key, value); 267 } 268 269 /** 270 * Helper method to convert a String to an int array. 271 */ 272 static private int[] getCodePoints(final String word) { 273 // TODO: this is a copy-paste of the contents of StringUtils.toCodePointArray, 274 // which is not visible from the makedict package. Factor this code. 275 final char[] characters = word.toCharArray(); 276 final int length = characters.length; 277 final int[] codePoints = new int[Character.codePointCount(characters, 0, length)]; 278 int codePoint = Character.codePointAt(characters, 0); 279 int dsti = 0; 280 for (int srci = Character.charCount(codePoint); 281 srci < length; srci += Character.charCount(codePoint), ++dsti) { 282 codePoints[dsti] = codePoint; 283 codePoint = Character.codePointAt(characters, srci); 284 } 285 codePoints[dsti] = codePoint; 286 return codePoints; 287 } 288 289 /** 290 * Helper method to add a word as a string. 291 * 292 * This method adds a word to the dictionary with the given frequency. Optional 293 * lists of bigrams and shortcuts can be passed here. For each word inside, 294 * they will be added to the dictionary as necessary. 295 * 296 * @param word the word to add. 297 * @param frequency the frequency of the word, in the range [0..255]. 298 * @param shortcutTargets a list of shortcut targets for this word, or null. 299 */ 300 public void add(final String word, final int frequency, 301 final ArrayList<WeightedString> shortcutTargets) { 302 add(getCodePoints(word), frequency, shortcutTargets); 303 } 304 305 /** 306 * Sanity check for a node. 307 * 308 * This method checks that all CharGroups in a node are ordered as expected. 309 * If they are, nothing happens. If they aren't, an exception is thrown. 310 */ 311 private void checkStack(Node node) { 312 ArrayList<CharGroup> stack = node.mData; 313 int lastValue = -1; 314 for (int i = 0; i < stack.size(); ++i) { 315 int currentValue = stack.get(i).mChars[0]; 316 if (currentValue <= lastValue) 317 throw new RuntimeException("Invalid stack"); 318 else 319 lastValue = currentValue; 320 } 321 } 322 323 /** 324 * Helper method to add a new bigram to the dictionary. 325 * 326 * @param word1 the previous word of the context 327 * @param word2 the next word of the context 328 * @param frequency the bigram frequency 329 */ 330 public void setBigram(final String word1, final String word2, final int frequency) { 331 CharGroup charGroup = findWordInTree(mRoot, word1); 332 if (charGroup != null) { 333 final CharGroup charGroup2 = findWordInTree(mRoot, word2); 334 if (charGroup2 == null) { 335 add(getCodePoints(word2), 0, null); 336 } 337 charGroup.addBigram(word2, frequency); 338 } else { 339 throw new RuntimeException("First word of bigram not found"); 340 } 341 } 342 343 /** 344 * Add a word to this dictionary. 345 * 346 * The shortcuts, if any, have to be in the dictionary already. If they aren't, 347 * an exception is thrown. 348 * 349 * @param word the word, as an int array. 350 * @param frequency the frequency of the word, in the range [0..255]. 351 * @param shortcutTargets an optional list of shortcut targets for this word (null if none). 352 */ 353 private void add(final int[] word, final int frequency, 354 final ArrayList<WeightedString> shortcutTargets) { 355 assert(frequency >= 0 && frequency <= 255); 356 Node currentNode = mRoot; 357 int charIndex = 0; 358 359 CharGroup currentGroup = null; 360 int differentCharIndex = 0; // Set by the loop to the index of the char that differs 361 int nodeIndex = findIndexOfChar(mRoot, word[charIndex]); 362 while (CHARACTER_NOT_FOUND != nodeIndex) { 363 currentGroup = currentNode.mData.get(nodeIndex); 364 differentCharIndex = compareArrays(currentGroup.mChars, word, charIndex); 365 if (ARRAYS_ARE_EQUAL != differentCharIndex 366 && differentCharIndex < currentGroup.mChars.length) break; 367 if (null == currentGroup.mChildren) break; 368 charIndex += currentGroup.mChars.length; 369 if (charIndex >= word.length) break; 370 currentNode = currentGroup.mChildren; 371 nodeIndex = findIndexOfChar(currentNode, word[charIndex]); 372 } 373 374 if (-1 == nodeIndex) { 375 // No node at this point to accept the word. Create one. 376 final int insertionIndex = findInsertionIndex(currentNode, word[charIndex]); 377 final CharGroup newGroup = new CharGroup( 378 Arrays.copyOfRange(word, charIndex, word.length), 379 shortcutTargets, null /* bigrams */, frequency); 380 currentNode.mData.add(insertionIndex, newGroup); 381 if (DBG) checkStack(currentNode); 382 } else { 383 // There is a word with a common prefix. 384 if (differentCharIndex == currentGroup.mChars.length) { 385 if (charIndex + differentCharIndex >= word.length) { 386 // The new word is a prefix of an existing word, but the node on which it 387 // should end already exists as is. Since the old CharNode was not a terminal, 388 // make it one by filling in its frequency and other attributes 389 currentGroup.update(frequency, shortcutTargets, null); 390 } else { 391 // The new word matches the full old word and extends past it. 392 // We only have to create a new node and add it to the end of this. 393 final CharGroup newNode = new CharGroup( 394 Arrays.copyOfRange(word, charIndex + differentCharIndex, word.length), 395 shortcutTargets, null /* bigrams */, frequency); 396 currentGroup.mChildren = new Node(); 397 currentGroup.mChildren.mData.add(newNode); 398 } 399 } else { 400 if (0 == differentCharIndex) { 401 // Exact same word. Update the frequency if higher. This will also add the 402 // new shortcuts to the existing shortcut list if it already exists. 403 currentGroup.update(frequency, shortcutTargets, null); 404 } else { 405 // Partial prefix match only. We have to replace the current node with a node 406 // containing the current prefix and create two new ones for the tails. 407 Node newChildren = new Node(); 408 final CharGroup newOldWord = new CharGroup( 409 Arrays.copyOfRange(currentGroup.mChars, differentCharIndex, 410 currentGroup.mChars.length), currentGroup.mShortcutTargets, 411 currentGroup.mBigrams, currentGroup.mFrequency, currentGroup.mChildren); 412 newChildren.mData.add(newOldWord); 413 414 final CharGroup newParent; 415 if (charIndex + differentCharIndex >= word.length) { 416 newParent = new CharGroup( 417 Arrays.copyOfRange(currentGroup.mChars, 0, differentCharIndex), 418 shortcutTargets, null /* bigrams */, frequency, newChildren); 419 } else { 420 newParent = new CharGroup( 421 Arrays.copyOfRange(currentGroup.mChars, 0, differentCharIndex), 422 null /* shortcutTargets */, null /* bigrams */, -1, newChildren); 423 final CharGroup newWord = new CharGroup(Arrays.copyOfRange(word, 424 charIndex + differentCharIndex, word.length), 425 shortcutTargets, null /* bigrams */, frequency); 426 final int addIndex = word[charIndex + differentCharIndex] 427 > currentGroup.mChars[differentCharIndex] ? 1 : 0; 428 newChildren.mData.add(addIndex, newWord); 429 } 430 currentNode.mData.set(nodeIndex, newParent); 431 } 432 if (DBG) checkStack(currentNode); 433 } 434 } 435 } 436 437 private static int ARRAYS_ARE_EQUAL = 0; 438 439 /** 440 * Custom comparison of two int arrays taken to contain character codes. 441 * 442 * This method compares the two arrays passed as an argument in a lexicographic way, 443 * with an offset in the dst string. 444 * This method does NOT test for the first character. It is taken to be equal. 445 * I repeat: this method starts the comparison at 1 <> dstOffset + 1. 446 * The index where the strings differ is returned. ARRAYS_ARE_EQUAL = 0 is returned if the 447 * strings are equal. This works BECAUSE we don't look at the first character. 448 * 449 * @param src the left-hand side string of the comparison. 450 * @param dst the right-hand side string of the comparison. 451 * @param dstOffset the offset in the right-hand side string. 452 * @return the index at which the strings differ, or ARRAYS_ARE_EQUAL = 0 if they don't. 453 */ 454 private static int compareArrays(final int[] src, final int[] dst, int dstOffset) { 455 // We do NOT test the first char, because we come from a method that already 456 // tested it. 457 for (int i = 1; i < src.length; ++i) { 458 if (dstOffset + i >= dst.length) return i; 459 if (src[i] != dst[dstOffset + i]) return i; 460 } 461 if (dst.length > src.length) return src.length; 462 return ARRAYS_ARE_EQUAL; 463 } 464 465 /** 466 * Helper class that compares and sorts two chargroups according to their 467 * first element only. I repeat: ONLY the first element is considered, the rest 468 * is ignored. 469 * This comparator imposes orderings that are inconsistent with equals. 470 */ 471 static private class CharGroupComparator implements java.util.Comparator<CharGroup> { 472 @Override 473 public int compare(CharGroup c1, CharGroup c2) { 474 if (c1.mChars[0] == c2.mChars[0]) return 0; 475 return c1.mChars[0] < c2.mChars[0] ? -1 : 1; 476 } 477 } 478 final static private CharGroupComparator CHARGROUP_COMPARATOR = new CharGroupComparator(); 479 480 /** 481 * Finds the insertion index of a character within a node. 482 */ 483 private static int findInsertionIndex(final Node node, int character) { 484 final ArrayList<CharGroup> data = node.mData; 485 final CharGroup reference = new CharGroup(new int[] { character }, 486 null /* shortcutTargets */, null /* bigrams */, 0); 487 int result = Collections.binarySearch(data, reference, CHARGROUP_COMPARATOR); 488 return result >= 0 ? result : -result - 1; 489 } 490 491 private static int CHARACTER_NOT_FOUND = -1; 492 493 /** 494 * Find the index of a char in a node, if it exists. 495 * 496 * @param node the node to search in. 497 * @param character the character to search for. 498 * @return the position of the character if it's there, or CHARACTER_NOT_FOUND = -1 else. 499 */ 500 private static int findIndexOfChar(final Node node, int character) { 501 final int insertionIndex = findInsertionIndex(node, character); 502 if (node.mData.size() <= insertionIndex) return CHARACTER_NOT_FOUND; 503 return character == node.mData.get(insertionIndex).mChars[0] ? insertionIndex 504 : CHARACTER_NOT_FOUND; 505 } 506 507 /** 508 * Helper method to find a word in a given branch. 509 */ 510 public static CharGroup findWordInTree(Node node, final String s) { 511 int index = 0; 512 final StringBuilder checker = DBG ? new StringBuilder() : null; 513 514 CharGroup currentGroup; 515 do { 516 int indexOfGroup = findIndexOfChar(node, s.codePointAt(index)); 517 if (CHARACTER_NOT_FOUND == indexOfGroup) return null; 518 currentGroup = node.mData.get(indexOfGroup); 519 if (DBG) checker.append(new String(currentGroup.mChars, 0, currentGroup.mChars.length)); 520 index += currentGroup.mChars.length; 521 if (index < s.length()) { 522 node = currentGroup.mChildren; 523 } 524 } while (null != node && index < s.length()); 525 526 if (DBG && !s.equals(checker.toString())) return null; 527 return currentGroup; 528 } 529 530 /** 531 * Helper method to find out whether a word is in the dict or not. 532 */ 533 public boolean hasWord(final String s) { 534 if (null == s || "".equals(s)) { 535 throw new RuntimeException("Can't search for a null or empty string"); 536 } 537 return null != findWordInTree(mRoot, s); 538 } 539 540 /** 541 * Recursively count the number of character groups in a given branch of the trie. 542 * 543 * @param node the parent node. 544 * @return the number of char groups in all the branch under this node. 545 */ 546 public static int countCharGroups(final Node node) { 547 final int nodeSize = node.mData.size(); 548 int size = nodeSize; 549 for (int i = nodeSize - 1; i >= 0; --i) { 550 CharGroup group = node.mData.get(i); 551 if (null != group.mChildren) 552 size += countCharGroups(group.mChildren); 553 } 554 return size; 555 } 556 557 /** 558 * Recursively count the number of nodes in a given branch of the trie. 559 * 560 * @param node the node to count. 561 * @return the number of nodes in this branch. 562 */ 563 public static int countNodes(final Node node) { 564 int size = 1; 565 for (int i = node.mData.size() - 1; i >= 0; --i) { 566 CharGroup group = node.mData.get(i); 567 if (null != group.mChildren) 568 size += countNodes(group.mChildren); 569 } 570 return size; 571 } 572 573 // Recursively find out whether there are any bigrams. 574 // This can be pretty expensive especially if there aren't any (we return as soon 575 // as we find one, so it's much cheaper if there are bigrams) 576 private static boolean hasBigramsInternal(final Node node) { 577 if (null == node) return false; 578 for (int i = node.mData.size() - 1; i >= 0; --i) { 579 CharGroup group = node.mData.get(i); 580 if (null != group.mBigrams) return true; 581 if (hasBigramsInternal(group.mChildren)) return true; 582 } 583 return false; 584 } 585 586 /** 587 * Finds out whether there are any bigrams in this dictionary. 588 * 589 * @return true if there is any bigram, false otherwise. 590 */ 591 // TODO: this is expensive especially for large dictionaries without any bigram. 592 // The up side is, this is always accurate and correct and uses no memory. We should 593 // find a more efficient way of doing this, without compromising too much on memory 594 // and ease of use. 595 public boolean hasBigrams() { 596 return hasBigramsInternal(mRoot); 597 } 598 599 // Historically, the tails of the words were going to be merged to save space. 600 // However, that would prevent the code to search for a specific address in log(n) 601 // time so this was abandoned. 602 // The code is still of interest as it does add some compression to any dictionary 603 // that has no need for attributes. Implementations that does not read attributes should be 604 // able to read a dictionary with merged tails. 605 // Also, the following code does support frequencies, as in, it will only merges 606 // tails that share the same frequency. Though it would result in the above loss of 607 // performance while searching by address, it is still technically possible to merge 608 // tails that contain attributes, but this code does not take that into account - it does 609 // not compare attributes and will merge terminals with different attributes regardless. 610 public void mergeTails() { 611 MakedictLog.i("Do not merge tails"); 612 return; 613 614 // MakedictLog.i("Merging nodes. Number of nodes : " + countNodes(root)); 615 // MakedictLog.i("Number of groups : " + countCharGroups(root)); 616 // 617 // final HashMap<String, ArrayList<Node>> repository = 618 // new HashMap<String, ArrayList<Node>>(); 619 // mergeTailsInner(repository, root); 620 // 621 // MakedictLog.i("Number of different pseudohashes : " + repository.size()); 622 // int size = 0; 623 // for (ArrayList<Node> a : repository.values()) { 624 // size += a.size(); 625 // } 626 // MakedictLog.i("Number of nodes after merge : " + (1 + size)); 627 // MakedictLog.i("Recursively seen nodes : " + countNodes(root)); 628 } 629 630 // The following methods are used by the deactivated mergeTails() 631 // private static boolean isEqual(Node a, Node b) { 632 // if (null == a && null == b) return true; 633 // if (null == a || null == b) return false; 634 // if (a.data.size() != b.data.size()) return false; 635 // final int size = a.data.size(); 636 // for (int i = size - 1; i >= 0; --i) { 637 // CharGroup aGroup = a.data.get(i); 638 // CharGroup bGroup = b.data.get(i); 639 // if (aGroup.frequency != bGroup.frequency) return false; 640 // if (aGroup.alternates == null && bGroup.alternates != null) return false; 641 // if (aGroup.alternates != null && !aGroup.equals(bGroup.alternates)) return false; 642 // if (!Arrays.equals(aGroup.chars, bGroup.chars)) return false; 643 // if (!isEqual(aGroup.children, bGroup.children)) return false; 644 // } 645 // return true; 646 // } 647 648 // static private HashMap<String, ArrayList<Node>> mergeTailsInner( 649 // final HashMap<String, ArrayList<Node>> map, final Node node) { 650 // final ArrayList<CharGroup> branches = node.data; 651 // final int nodeSize = branches.size(); 652 // for (int i = 0; i < nodeSize; ++i) { 653 // CharGroup group = branches.get(i); 654 // if (null != group.children) { 655 // String pseudoHash = getPseudoHash(group.children); 656 // ArrayList<Node> similarList = map.get(pseudoHash); 657 // if (null == similarList) { 658 // similarList = new ArrayList<Node>(); 659 // map.put(pseudoHash, similarList); 660 // } 661 // boolean merged = false; 662 // for (Node similar : similarList) { 663 // if (isEqual(group.children, similar)) { 664 // group.children = similar; 665 // merged = true; 666 // break; 667 // } 668 // } 669 // if (!merged) { 670 // similarList.add(group.children); 671 // } 672 // mergeTailsInner(map, group.children); 673 // } 674 // } 675 // return map; 676 // } 677 678 // private static String getPseudoHash(final Node node) { 679 // StringBuilder s = new StringBuilder(); 680 // for (CharGroup g : node.data) { 681 // s.append(g.frequency); 682 // for (int ch : g.chars){ 683 // s.append(Character.toChars(ch)); 684 // } 685 // } 686 // return s.toString(); 687 // } 688 689 /** 690 * Iterator to walk through a dictionary. 691 * 692 * This is purely for convenience. 693 */ 694 public static class DictionaryIterator implements Iterator<Word> { 695 696 private static class Position { 697 public Iterator<CharGroup> pos; 698 public int length; 699 public Position(ArrayList<CharGroup> groups) { 700 pos = groups.iterator(); 701 length = 0; 702 } 703 } 704 final StringBuilder mCurrentString; 705 final LinkedList<Position> mPositions; 706 707 public DictionaryIterator(ArrayList<CharGroup> root) { 708 mCurrentString = new StringBuilder(); 709 mPositions = new LinkedList<Position>(); 710 final Position rootPos = new Position(root); 711 mPositions.add(rootPos); 712 } 713 714 @Override 715 public boolean hasNext() { 716 for (Position p : mPositions) { 717 if (p.pos.hasNext()) { 718 return true; 719 } 720 } 721 return false; 722 } 723 724 @Override 725 public Word next() { 726 Position currentPos = mPositions.getLast(); 727 mCurrentString.setLength(mCurrentString.length() - currentPos.length); 728 729 do { 730 if (currentPos.pos.hasNext()) { 731 final CharGroup currentGroup = currentPos.pos.next(); 732 currentPos.length = currentGroup.mChars.length; 733 for (int i : currentGroup.mChars) 734 mCurrentString.append(Character.toChars(i)); 735 if (null != currentGroup.mChildren) { 736 currentPos = new Position(currentGroup.mChildren.mData); 737 mPositions.addLast(currentPos); 738 } 739 if (currentGroup.mFrequency >= 0) 740 return new Word(mCurrentString.toString(), currentGroup.mFrequency, 741 currentGroup.mShortcutTargets, currentGroup.mBigrams); 742 } else { 743 mPositions.removeLast(); 744 currentPos = mPositions.getLast(); 745 mCurrentString.setLength(mCurrentString.length() - mPositions.getLast().length); 746 } 747 } while(true); 748 } 749 750 @Override 751 public void remove() { 752 throw new UnsupportedOperationException("Unsupported yet"); 753 } 754 755 } 756 757 /** 758 * Method to return an iterator. 759 * 760 * This method enables Java's enhanced for loop. With this you can have a FusionDictionary x 761 * and say : for (Word w : x) {} 762 */ 763 @Override 764 public Iterator<Word> iterator() { 765 return new DictionaryIterator(mRoot.mData); 766 } 767 } 768