/**
 * Builds the normalization tables. This is a separate class so that it
 * can be unloaded once not needed.<br>
 * Copyright (C) 1998-2007 International Business Machines Corporation and
 * Unicode, Inc. All Rights Reserved.<br>
 * The Unicode Consortium makes no expressed or implied warranty of any
 * kind, and assumes no liability for errors or omissions.
 * No liability is assumed for incidental and consequential damages
 * in connection with or arising out of the use of the information here.
 * @author Mark Davis
 * Updates for supplementary code points:
 * Vladimir Weinstein & Markus Scherer
 */

package com.ibm.icu.dev.test.normalizer;

import java.io.BufferedReader;
import java.util.BitSet;

import com.ibm.icu.dev.test.TestUtil;
import com.ibm.icu.dev.test.UTF16Util;

class NormalizerBuilder {
    //private static final String copyright = "Copyright (C) 1998-2003 International Business Machines Corporation and Unicode, Inc.";

    /**
     * Testing flags
     */

    private static final boolean DEBUG = false;
    //private static final boolean GENERATING = false;

    /**
     * Constants for the data file version to use.
     */
    /*static final boolean NEW_VERSION = true;
    private static final String DIR = "D:\\UnicodeData\\" + (NEW_VERSION ? "WorkingGroups\\" : "");

    static final String UNIDATA_VERSION = NEW_VERSION ? "3.0.0d12" : "2.1.9";
    static final String EXCLUSIONS_VERSION = NEW_VERSION ? "1d4" : "1";

    public static final String UNICODE_DATA = DIR + "UnicodeData-" + UNIDATA_VERSION + ".txt";
    public static final String COMPOSITION_EXCLUSIONS = DIR + "CompositionExclusions-" + EXCLUSIONS_VERSION +".txt";
    */

    /**
     * Called exactly once by NormalizerData to build the static data
     */

    static NormalizerData build(boolean fullData) {
        try {
            IntHashtable canonicalClass = new IntHashtable(0);
            IntStringHashtable decompose = new IntStringHashtable(null);
            LongHashtable compose = new LongHashtable(NormalizerData.NOT_COMPOSITE);
            BitSet isCompatibility = new BitSet();
            BitSet isExcluded = new BitSet();
            if (fullData) {
                //System.out.println("Building Normalizer Data from file.");
                readExclusionList(isExcluded);
                //System.out.println(isExcluded.get(0x00C0));
                buildDecompositionTables(canonicalClass, decompose, compose,
                  isCompatibility, isExcluded);
            } else {    // for use in Applets
                //System.out.println("Building abridged data.");
                setMinimalDecomp(canonicalClass, decompose, compose,
                  isCompatibility, isExcluded);
            }
            return new NormalizerData(canonicalClass, decompose, compose,
                  isCompatibility, isExcluded);
        } catch (java.io.IOException e) {
            System.err.println("Can't load data file." + e + ", " + e.getMessage());
            return null;
        }
    }

// =============================================================
// Building Decomposition Tables
// =============================================================

    /**
     * Reads exclusion list and stores the data
     */
    private static void readExclusionList(BitSet isExcluded) throws java.io.IOException {
        if (DEBUG) System.out.println("Reading Exclusions");

        BufferedReader in = TestUtil.getDataReader("unicode/CompositionExclusions.txt");

        while (true) {
            // read a line, discarding comments and blank lines

            String line = in.readLine();
            if (line == null) break;
            int comment = line.indexOf('#');                    // strip comments
            if (comment != -1) line = line.substring(0,comment);
            if (line.length() == 0) continue;                   // ignore blanks
            if(line.indexOf(' ') != -1) {
                line = line.substring(0, line.indexOf(' '));
            }
            // store -1 in the excluded table for each character hit

            int value = Integer.parseInt(line,16);
            isExcluded.set(value);
            //System.out.println("Excluding " + hex(value));
        }
        in.close();
        if (DEBUG) System.out.println("Done reading Exclusions");
    }

    /**
     * Builds a decomposition table from a UnicodeData file
     */
    private static void buildDecompositionTables(
      IntHashtable canonicalClass, IntStringHashtable decompose,
      LongHashtable compose, BitSet isCompatibility, BitSet isExcluded)
      throws java.io.IOException {
        if (DEBUG) System.out.println("Reading Unicode Character Database");
        //BufferedReader in = new BufferedReader(new FileReader(UNICODE_DATA), 64*1024);
        BufferedReader in = null;
        try {
            in = TestUtil.getDataReader("unicode/UnicodeData.txt");
        } catch (Exception e) {
            System.err.println("Failed to read UnicodeData.txt");
            System.exit(1);
        }

        int value;
        long pair;
        int counter = 0;
        while (true) {

            // read a line, discarding comments and blank lines

            String line = in.readLine();
            if (line == null) break;
            int comment = line.indexOf('#');                    // strip comments
            if (comment != -1) line = line.substring(0,comment);
            if (line.length() == 0) continue;
            if (DEBUG) {
                counter++;
                if ((counter & 0xFF) == 0) System.out.println("At: " + line);
            }

            // find the values of the particular fields that we need
            // Sample line: 00C0;LATIN ...A GRAVE;Lu;0;L;0041 0300;;;;N;LATIN ... GRAVE;;;00E0;

            int start = 0;
            int end = line.indexOf(';'); // code
            value = Integer.parseInt(line.substring(start,end),16);
            if (true && value == '\u00c0') {
                //System.out.println("debug: " + line);
            }
            end = line.indexOf(';',start=end+1); // name
            /*String name = line.substring(start,end);*/
            end = line.indexOf(';',start=end+1); // general category
            end = line.indexOf(';',start=end+1); // canonical class

            // check consistency: canonical classes must be from 0 to 255

            int cc = Integer.parseInt(line.substring(start,end));
            if (cc != (cc & 0xFF)) System.err.println("Bad canonical class at: " + line);
            canonicalClass.put(value,cc);
            end = line.indexOf(';',start=end+1); // BIDI
            end = line.indexOf(';',start=end+1); // decomp

            // decomp requires more processing.
            // store whether it is canonical or compatibility.
            // store the decomp in one table, and the reverse mapping (from pairs) in another

            if (start != end) {
                String segment = line.substring(start, end);
                boolean compat = segment.charAt(0) == '<';
                if (compat) isCompatibility.set(value);
                String decomp = fromHex(segment);

                // a small snippet of code to generate the Applet data

                /*if (GENERATING) {
                    if (value < 0xFF) {
                        System.out.println(
                            "\"\\u" + hex((char)value) + "\", "
                            + "\"\\u" + hex(decomp, "\\u") + "\", "
                            + (compat ? "\"K\"," : "\"\",")
                            + "// " + name);
                    }
                }*/

                // check consistency: all canon decomps must be singles or pairs!
                int decompLen = UTF16Util.countCodePoint(decomp);
                if (decompLen < 1 || decompLen > 2 && !compat) {
                    System.err.println("Bad decomp at: " + line);
                }
                decompose.put(value, decomp);

                // only compositions are canonical pairs
                // skip if script exclusion

                if (!compat && !isExcluded.get(value)) {
                    int first = '\u0000';
                    int second = UTF16Util.nextCodePoint(decomp, 0);
                    if (decompLen > 1) {
                        first = second;
                        second = UTF16Util.nextCodePoint(decomp,
                            UTF16Util.codePointLength(first));
                    }

                    // store composition pair in single integer

                    pair = ((long)first << 32) | second;
                    if (DEBUG && value == '\u00C0') {
                        System.out.println("debug2: " + line);
                    }
                    compose.put(pair, value);
                } else if (DEBUG) {
                    System.out.println("Excluding: " + decomp);
                }
            }
        }
        in.close();
        if (DEBUG) System.out.println("Done reading Unicode Character Database");

        // add algorithmic Hangul decompositions
        // this is more compact if done at runtime, but for simplicity we
        // do it this way.

        if (DEBUG) System.out.println("Adding Hangul");

        for (int SIndex = 0; SIndex < SCount; ++SIndex) {
            int TIndex = SIndex % TCount;
            char first, second;
            if (TIndex != 0) { // triple
                first = (char)(SBase + SIndex - TIndex);
                second = (char)(TBase + TIndex);
            } else {
                first = (char)(LBase + SIndex / NCount);
                second = (char)(VBase + (SIndex % NCount) / TCount);
            }
            pair = ((long)first << 32) | second;
            value = SIndex + SBase;
            decompose.put(value, String.valueOf(first) + second);
            compose.put(pair, value);
        }
        if (DEBUG) System.out.println("Done adding Hangul");
    }

    /**
     * Hangul composition constants
     */
    static final int
        SBase = 0xAC00, LBase = 0x1100, VBase = 0x1161, TBase = 0x11A7,
        LCount = 19, VCount = 21, TCount = 28,
        NCount = VCount * TCount,   // 588
        SCount = LCount * NCount;   // 11172

    /**
     * For use in an applet: just load a minimal set of data.
     */
    private static void setMinimalDecomp(IntHashtable canonicalClass, IntStringHashtable decompose,
      LongHashtable compose, BitSet isCompatibility, BitSet isExcluded) {
        String[] decomposeData = {
            "\u005E", "\u0020\u0302", "K",
            "\u005F", "\u0020\u0332", "K",
            "\u0060", "\u0020\u0300", "K",
            "\u00A0", "\u0020", "K",
            "\u00A8", "\u0020\u0308", "K",
            "\u00AA", "\u0061", "K",
            "\u00AF", "\u0020\u0304", "K",
            "\u00B2", "\u0032", "K",
            "\u00B3", "\u0033", "K",
            "\u00B4", "\u0020\u0301", "K",
            "\u00B5", "\u03BC", "K",
            "\u00B8", "\u0020\u0327", "K",
            "\u00B9", "\u0031", "K",
            "\u00BA", "\u006F", "K",
            "\u00BC", "\u0031\u2044\u0034", "K",
            "\u00BD", "\u0031\u2044\u0032", "K",
            "\u00BE", "\u0033\u2044\u0034", "K",
            "\u00C0", "\u0041\u0300", "",
            "\u00C1", "\u0041\u0301", "",
            "\u00C2", "\u0041\u0302", "",
            "\u00C3", "\u0041\u0303", "",
            "\u00C4", "\u0041\u0308", "",
            "\u00C5", "\u0041\u030A", "",
            "\u00C7", "\u0043\u0327", "",
            "\u00C8", "\u0045\u0300", "",
            "\u00C9", "\u0045\u0301", "",
            "\u00CA", "\u0045\u0302", "",
            "\u00CB", "\u0045\u0308", "",
            "\u00CC", "\u0049\u0300", "",
            "\u00CD", "\u0049\u0301", "",
            "\u00CE", "\u0049\u0302", "",
            "\u00CF", "\u0049\u0308", "",
            "\u00D1", "\u004E\u0303", "",
            "\u00D2", "\u004F\u0300", "",
            "\u00D3", "\u004F\u0301", "",
            "\u00D4", "\u004F\u0302", "",
            "\u00D5", "\u004F\u0303", "",
            "\u00D6", "\u004F\u0308", "",
            "\u00D9", "\u0055\u0300", "",
            "\u00DA", "\u0055\u0301", "",
            "\u00DB", "\u0055\u0302", "",
            "\u00DC", "\u0055\u0308", "",
            "\u00DD", "\u0059\u0301", "",
            "\u00E0", "\u0061\u0300", "",
            "\u00E1", "\u0061\u0301", "",
            "\u00E2", "\u0061\u0302", "",
            "\u00E3", "\u0061\u0303", "",
            "\u00E4", "\u0061\u0308", "",
            "\u00E5", "\u0061\u030A", "",
            "\u00E7", "\u0063\u0327", "",
            "\u00E8", "\u0065\u0300", "",
            "\u00E9", "\u0065\u0301", "",
            "\u00EA", "\u0065\u0302", "",
            "\u00EB", "\u0065\u0308", "",
            "\u00EC", "\u0069\u0300", "",
            "\u00ED", "\u0069\u0301", "",
            "\u00EE", "\u0069\u0302", "",
            "\u00EF", "\u0069\u0308", "",
            "\u00F1", "\u006E\u0303", "",
            "\u00F2", "\u006F\u0300", "",
            "\u00F3", "\u006F\u0301", "",
            "\u00F4", "\u006F\u0302", "",
            "\u00F5", "\u006F\u0303", "",
            "\u00F6", "\u006F\u0308", "",
            "\u00F9", "\u0075\u0300", "",
            "\u00FA", "\u0075\u0301", "",
            "\u00FB", "\u0075\u0302", "",
            "\u00FC", "\u0075\u0308", "",
            "\u00FD", "\u0079\u0301", "",
// EXTRAS, outside of Latin 1
            "\u1EA4", "\u00C2\u0301", "",
            "\u1EA5", "\u00E2\u0301", "",
            "\u1EA6", "\u00C2\u0300", "",
            "\u1EA7", "\u00E2\u0300", "",
        };

        int[] classData = {
            0x0300, 230,
            0x0301, 230,
            0x0302, 230,
            0x0303, 230,
            0x0304, 230,
            0x0305, 230,
            0x0306, 230,
            0x0307, 230,
            0x0308, 230,
            0x0309, 230,
            0x030A, 230,
            0x030B, 230,
            0x030C, 230,
            0x030D, 230,
            0x030E, 230,
            0x030F, 230,
            0x0310, 230,
            0x0311, 230,
            0x0312, 230,
            0x0313, 230,
            0x0314, 230,
            0x0315, 232,
            0x0316, 220,
            0x0317, 220,
            0x0318, 220,
            0x0319, 220,
            0x031A, 232,
            0x031B, 216,
            0x031C, 220,
            0x031D, 220,
            0x031E, 220,
            0x031F, 220,
            0x0320, 220,
            0x0321, 202,
            0x0322, 202,
            0x0323, 220,
            0x0324, 220,
            0x0325, 220,
            0x0326, 220,
            0x0327, 202,
            0x0328, 202,
            0x0329, 220,
            0x032A, 220,
            0x032B, 220,
            0x032C, 220,
            0x032D, 220,
            0x032E, 220,
            0x032F, 220,
            0x0330, 220,
            0x0331, 220,
            0x0332, 220,
            0x0333, 220,
            0x0334, 1,
            0x0335, 1,
            0x0336, 1,
            0x0337, 1,
            0x0338, 1,
            0x0339, 220,
            0x033A, 220,
            0x033B, 220,
            0x033C, 220,
            0x033D, 230,
            0x033E, 230,
            0x033F, 230,
            0x0340, 230,
            0x0341, 230,
            0x0342, 230,
            0x0343, 230,
            0x0344, 230,
            0x0345, 240,
            0x0360, 234,
            0x0361, 234
        };

        // build the same tables we would otherwise get from the
        // Unicode Character Database, just with limited data

        for (int i = 0; i < decomposeData.length; i+=3) {
            char value = decomposeData[i].charAt(0);
            String decomp = decomposeData[i+1];
            boolean compat = decomposeData[i+2].equals("K");
            if (compat) isCompatibility.set(value);
            decompose.put(value, decomp);
            if (!compat) {
                int first = '\u0000';
                int second = UTF16Util.nextCodePoint(decomp, 0);
                if (decomp.length() > 1) {
                    first = second;
                    second = UTF16Util.nextCodePoint(decomp,
                        UTF16Util.codePointLength(first));
                }
                long pair = (first << 16) | second;
                compose.put(pair, value);
            }
        }

        for (int i = 0; i < classData.length;) {
            canonicalClass.put(classData[i++], classData[i++]);
        }
    }

    /**
     * Utility: Parses a sequence of hex Unicode characters separated by spaces
     */
    static public String fromHex(String source) {
        StringBuffer result = new StringBuffer();
        for (int i = 0; i < source.length(); ++i) {
            char c = source.charAt(i);
            switch (c) {
              case ' ': break; // ignore
              case '0': case '1': case '2': case '3': case '4': case '5': case '6': case '7':
              case '8': case '9': case 'A': case 'B': case 'C': case 'D': case 'E': case 'F':
              case 'a': case 'b': case 'c': case 'd': case 'e': case 'f':
                int end = 0;
                int value = 0;
                try {
                    //System.out.println(source.substring(i, i + 4) + "************" + source);
                    end = source.indexOf(' ',i);
                    if (end < 0) {
                        end = source.length();
                    }
                    value = Integer.parseInt(source.substring(i, end),16);
                    UTF16Util.appendCodePoint(result, value);
                } catch (Exception e) {
                    System.out.println("i: " + i + ";end:" + end + "source:" + source);
                    //System.out.println(source.substring(i, i + 4) + "************" + source);
                    System.exit(1);
                }
                //i+= 3; // skip rest of number
                i = end;
                break;
              case '<': int j = source.indexOf('>',i); // skip <...>
                if (j > 0) {
                    i = j;
                    break;
                } // else fall through--error
              default:
                throw new IllegalArgumentException("Bad hex value in " + source);
            }
        }
        return result.toString();
    }

    /**
     * Utility: Supplies a zero-padded hex representation of an integer (without 0x)
     */
    static public String hex(int i) {
        String result = Long.toString(i & 0xFFFFFFFFL, 16).toUpperCase();
        return "00000000".substring(result.length(),8) + result;
    }

    /**
     * Utility: Supplies a zero-padded hex representation of a Unicode character (without 0x, \\u)
     */
    static public String hex(char i) {
        String result = Integer.toString(i, 16).toUpperCase();
        return "0000".substring(result.length(),4) + result;
    }

    /**
     * Utility: Supplies a zero-padded hex representation of a Unicode character (without 0x, \\u)
     */
    public static String hex(String s, String sep) {
        StringBuffer result = new StringBuffer();
        for (int i = 0; i < s.length(); ++i) {
            if (i != 0) result.append(sep);
            result.append(hex(s.charAt(i)));
        }
        return result.toString();
    }
}