/* ******************************************************************************* * Copyright (C) 1996-2010, International Business Machines Corporation and * * others. All Rights Reserved. * ******************************************************************************* */ package com.ibm.icu.text; import java.io.DataInputStream; import java.io.FileInputStream; import java.io.FileNotFoundException; import java.io.FileOutputStream; import java.io.IOException; import java.io.InputStream; import java.io.OutputStreamWriter; import java.io.PrintWriter; import java.io.UnsupportedEncodingException; import com.ibm.icu.util.CompactByteArray; /** * This is the class that represents the list of known words used by * DictionaryBasedBreakIterator. The conceptual data structure used * here is a trie: there is a node hanging off the root node for every * letter that can start a word. Each of these nodes has a node hanging * off of it for every letter that can be the second letter of a word * if this node is the first letter, and so on. The trie is represented * as a two-dimensional array that can be treated as a table of state * transitions. Indexes are used to compress this array, taking * advantage of the fact that this array will always be very sparse. */ class BreakDictionary { //================================================================================= // testing and debugging //================================================================================= ///CLOVER:OFF //The main method looks like it was useful once but now seems worthless. It is not used by any method or class. public static void main(String args[]) throws FileNotFoundException, UnsupportedEncodingException, IOException { String filename = args[0]; BreakDictionary dictionary = new BreakDictionary(new FileInputStream(filename)); PrintWriter out = null; if(args.length >= 2) { out = new PrintWriter(new OutputStreamWriter(new FileOutputStream(args[1]), "UnicodeLittle")); } dictionary.printWordList("", 0, out); if (out != null) { out.close(); } } ///CLOVER:ON ///CLOVER:OFF /* public */ void printWordList(String partialWord, int state, PrintWriter out) throws IOException { if (state == 0xFFFF) { System.out.println(partialWord); if (out != null) { out.println(partialWord); } } else { for (int i = 0; i < numCols; i++) { int newState = (at(state, i)) & 0xFFFF; if (newState != 0) { char newChar = reverseColumnMap[i]; String newPartialWord = partialWord; if (newChar != 0) { newPartialWord += newChar; } printWordList(newPartialWord, newState, out); } } } } ///CLOVER:ON /** * A map used to go from column numbers to characters. Used only * for debugging right now. */ private char[] reverseColumnMap = null; //================================================================================= // data members //================================================================================= /** * Maps from characters to column numbers. The main use of this is to * avoid making room in the array for empty columns. */ private CompactByteArray columnMap = null; /** * The number of actual columns in the table */ private int numCols; /* * Columns are organized into groups of 32. This says how many * column groups. (We could calculate this, but we store the * value to avoid having to repeatedly calculate it.) */ //private int numColGroups; /** * The actual compressed state table. Each conceptual row represents * a state, and the cells in it contain the row numbers of the states * to transition to for each possible letter. 0 is used to indicate * an illegal combination of letters (i.e., the error state). The * table is compressed by eliminating all the unpopulated (i.e., zero) * cells. Multiple conceptual rows can then be doubled up in a single * physical row by sliding them up and possibly shifting them to one * side or the other so the populated cells don't collide. Indexes * are used to identify unpopulated cells and to locate populated cells. */ private short[] table = null; /** * This index maps logical row numbers to physical row numbers */ private short[] rowIndex = null; /** * A bitmap is used to tell which cells in the comceptual table are * populated. This array contains all the unique bit combinations * in that bitmap. If the table is more than 32 columns wide, * successive entries in this array are used for a single row. */ private int[] rowIndexFlags = null; /** * This index maps from a logical row number into the bitmap table above. * (This keeps us from storing duplicate bitmap combinations.) Since there * are a lot of rows with only one populated cell, instead of wasting space * in the bitmap table, we just store a negative number in this index for * rows with one populated cell. The absolute value of that number is * the column number of the populated cell. */ private short[] rowIndexFlagsIndex = null; /** * For each logical row, this index contains a constant that is added to * the logical column number to get the physical column number */ private byte[] rowIndexShifts = null; //================================================================================= // deserialization //================================================================================= /* public */ BreakDictionary(InputStream dictionaryStream) throws IOException { readDictionaryFile(new DataInputStream(dictionaryStream)); } /* public */ void readDictionaryFile(DataInputStream in) throws IOException { int l; // read in the version number (right now we just ignore it) in.readInt(); // read in the column map (this is serialized in its internal form: // an index array followed by a data array) l = in.readInt(); char[] temp = new char[l]; for (int i = 0; i < temp.length; i++) temp[i] = (char)in.readShort(); l = in.readInt(); byte[] temp2 = new byte[l]; for (int i = 0; i < temp2.length; i++) temp2[i] = in.readByte(); columnMap = new CompactByteArray(temp, temp2); // read in numCols and numColGroups numCols = in.readInt(); /*numColGroups = */in.readInt(); // read in the row-number index l = in.readInt(); rowIndex = new short[l]; for (int i = 0; i < rowIndex.length; i++) rowIndex[i] = in.readShort(); // load in the populated-cells bitmap: index first, then bitmap list l = in.readInt(); rowIndexFlagsIndex = new short[l]; for (int i = 0; i < rowIndexFlagsIndex.length; i++) rowIndexFlagsIndex[i] = in.readShort(); l = in.readInt(); rowIndexFlags = new int[l]; for (int i = 0; i < rowIndexFlags.length; i++) rowIndexFlags[i] = in.readInt(); // load in the row-shift index l = in.readInt(); rowIndexShifts = new byte[l]; for (int i = 0; i < rowIndexShifts.length; i++) rowIndexShifts[i] = in.readByte(); // finally, load in the actual state table l = in.readInt(); table = new short[l]; for (int i = 0; i < table.length; i++) table[i] = in.readShort(); // this data structure is only necessary for testing and debugging purposes reverseColumnMap = new char[numCols]; for (char c = 0; c < 0xffff; c++) { int col = columnMap.elementAt(c); if (col != 0) { reverseColumnMap[col] = c; } } // close the stream in.close(); } //================================================================================= // access to the words //================================================================================= /** * Uses the column map to map the character to a column number, then * passes the row and column number to the other version of at() * @param row The current state * @param ch The character whose column we're interested in * @return The new state to transition to */ /* public */ final short at(int row, char ch) { int col = columnMap.elementAt(ch); return at(row, col); } /** * Returns the value in the cell with the specified (logical) row and * column numbers. In DictionaryBasedBreakIterator, the row number is * a state number, the column number is an input, and the return value * is the row number of the new state to transition to. (0 is the * "error" state, and -1 is the "end of word" state in a dictionary) * @param row The row number of the current state * @param col The column number of the input character (0 means "not a * dictionary character") * @return The row number of the new state to transition to */ /* public */ final short at(int row, int col) { if (cellIsPopulated(row, col)) { // we map from logical to physical row number by looking up the // mapping in rowIndex; we map from logical column number to // physical column number by looking up a shift value for this // logical row and offsetting the logical column number by // the shift amount. Then we can use internalAt() to actually // get the value out of the table. return internalAt(rowIndex[row], col + rowIndexShifts[row]); } else { return 0; } } /** * Given (logical) row and column numbers, returns true if the * cell in that position is populated */ private final boolean cellIsPopulated(int row, int col) { // look up the entry in the bitmap index for the specified row. // If it's a negative number, it's the column number of the only // populated cell in the row if (rowIndexFlagsIndex[row] < 0) { return col == -rowIndexFlagsIndex[row]; } // if it's a positive number, it's the offset of an entry in the bitmap // list. If the table is more than 32 columns wide, the bitmap is stored // successive entries in the bitmap list, so we have to divide the column // number by 32 and offset the number we got out of the index by the result. // Once we have the appropriate piece of the bitmap, test the appropriate // bit and return the result. else { int flags = rowIndexFlags[rowIndexFlagsIndex[row] + (col >> 5)]; return (flags & (1 << (col & 0x1f))) != 0; } } /** * Implementation of at() when we know the specified cell is populated. * @param row The PHYSICAL row number of the cell * @param col The PHYSICAL column number of the cell * @return The value stored in the cell */ private final short internalAt(int row, int col) { // the table is a one-dimensional array, so this just does the math necessary // to treat it as a two-dimensional array (we don't just use a two-dimensional // array because two-dimensional arrays are inefficient in Java) return table[row * numCols + col]; } }