arabicshaping.java

一个java操作pdf文件的开发包,很好用的.

/*
*******************************************************************************
*   Copyright (C) 2001, International Business Machines
*   Corporation and others.  All Rights Reserved.
*******************************************************************************
*/

package com.lowagie.text.pdf;

//import com.ibm.icu.lang.*;

/**
 * Shape Arabic text on a character basis.
 *
 * <p>ArabicShaping performs basic operations for "shaping" Arabic text. It is most
 * useful for use with legacy data formats and legacy display technology
 * (simple terminals). All operations are performed on Unicode characters.</p>
 *
 * <p>Text-based shaping means that some character code points in the text are
 * replaced by others depending on the context. It transforms one kind of text
 * into another. In comparison, modern displays for Arabic text select
 * appropriate, context-dependent font glyphs for each text element, which means
 * that they transform text into a glyph vector.</p>
 *
 * <p>Text transformations are necessary when modern display technology is not
 * available or when text needs to be transformed to or from legacy formats that
 * use "shaped" characters. Since the Arabic script is cursive, connecting
 * adjacent letters to each other, computers select images for each letter based
 * on the surrounding letters. This usually results in four images per Arabic
 * letter: initial, middle, final, and isolated forms. In Unicode, on the other
 * hand, letters are normally stored abstract, and a display system is expected
 * to select the necessary glyphs. (This makes searching and other text
 * processing easier because the same letter has only one code.) It is possible
 * to mimic this with text transformations because there are characters in
 * Unicode that are rendered as letters with a specific shape
 * (or cursive connectivity). They were included for interoperability with
 * legacy systems and codepages, and for unsophisticated display systems.</p>
 *
 * <p>A second kind of text transformations is supported for Arabic digits:
 * For compatibility with legacy codepages that only include European digits,
 * it is possible to replace one set of digits by another, changing the
 * character code points. These operations can be performed for either
 * Arabic-Indic Digits (U+0660...U+0669) or Eastern (Extended) Arabic-Indic
 * digits (U+06f0...U+06f9).</p>
 *
 * <p>Some replacements may result in more or fewer characters (code points).
 * By default, this means that the destination buffer may receive text with a
 * length different from the source length. Some legacy systems rely on the
 * length of the text to be constant. They expect extra spaces to be added
 * or consumed either next to the affected character or at the end of the
 * text.</p>
 */
public final class ArabicShaping {
    private final int options;
    private boolean isLogical; // convenience

    /**
     * Convert a range of text in the source array, putting the result 
     * into a range of text in the destination array, and return the number
     * of characters written.
     *
     * @param source An array containing the input text
     * @param sourceStart The start of the range of text to convert
     * @param sourceLength The length of the range of text to convert
     * @param dest The destination array that will receive the result.
     *   It may be <code>NULL</code> only if  <code>destSize</code> is 0.  
     * @param destStart The start of the range of the destination buffer to use.
     * @param destSize The size (capacity) of the destination buffer.
     *   If <code>destSize</code> is 0, then no output is produced,
     *   but the necessary buffer size is returned ("preflighting").  This
     *   does not validate the text against the options, for example, 
     *   if letters are being unshaped, and spaces are being consumed
     *   following lamalef, this will not detect a lamalef without a 
     *   corresponding space.  An error will be thrown when the actual
     *   conversion is attempted.
     * @return The number of chars written to the destination buffer.
     *   If an error occurs, then no output was written, or it may be
     *   incomplete.
     * @if the text cannot be converted according to the options.
     */
    public int shape(char[] source, int sourceStart, int sourceLength,
                     char[] dest, int destStart, int destSize) {
        if (source == null) {
            throw new IllegalArgumentException("source can not be null");
        }
        if (sourceStart < 0 || sourceLength < 0 || sourceStart + sourceLength > source.length) {
            throw new IllegalArgumentException("bad source start (" + sourceStart +
                                               ") or length (" + sourceLength +
                                               ") for buffer of length " + source.length);
        }
        if (dest == null && destSize != 0) {
            throw new IllegalArgumentException("null dest requires destSize == 0");
        }
        if ((destSize != 0) &&
            (destStart < 0 || destSize < 0 || destStart + destSize > dest.length)) {
            throw new IllegalArgumentException("bad dest start (" + destStart + 
                                               ") or size (" + destSize + 
                                               ") for buffer of length " + dest.length);
        }

        return internalShape(source, sourceStart, sourceLength, dest, destStart, destSize);
    }

    /**
     * Convert a range of text in place.  This may only be used if the Length option
     * does not grow or shrink the text.
     *
     * @param source An array containing the input text
     * @param start The start of the range of text to convert
     * @param length The length of the range of text to convert
     * @if the text cannot be converted according to the options.
     */
    public void shape(char[] source, int start, int length) {
        if ((options & LENGTH_MASK) == LENGTH_GROW_SHRINK) {
            throw new RuntimeException("Cannot shape in place with length option grow/shrink.");
        }
        shape(source, start, length, source, start, length);
    }

    /**
     * Convert a string, returning the new string.
     *
     * @param source The string to convert.
     * @return The converted string.
     * @if the string cannot be converted according to the options.
     */
    public String shape(String text) {
        char[] src = text.toCharArray();
        char[] dest = src;
        if (((options & LENGTH_MASK) == LENGTH_GROW_SHRINK) &&
            ((options & LETTERS_MASK) == LETTERS_UNSHAPE)) {

            dest = new char[src.length * 2]; // max
        }
        int len = shape(src, 0, src.length, dest, 0, dest.length);

        return new String(dest, 0, len);
    }

    /**
     * Construct ArabicShaping using the options flags.
     * The flags are as follows:<br>
     * 'LENGTH' flags control whether the text can change size, and if not,
     * how to maintain the size of the text when LamAlef ligatures are 
     * formed or broken.<br>
     * 'TEXT_DIRECTION' flags control whether the text is read and written
     * in visual order or in logical order.<br>
     * 'LETTERS_SHAPE' flags control whether conversion is to or from
     * presentation forms.<br>
     * 'DIGITS' flags control whether digits are shaped, and whether from
     * European to Arabic-Indic or vice-versa.<br>
     * 'DIGIT_TYPE' flags control whether standard or extended Arabic-Indic
     * digits are used when performing digit conversion.
     */
    public ArabicShaping(int options) {
        this.options = options;
        if ((options & DIGITS_MASK) > 0x80) {
            throw new IllegalArgumentException("bad DIGITS options");
        }
        isLogical = (options & TEXT_DIRECTION_MASK) == TEXT_DIRECTION_LOGICAL;
    }

    /**
     * Memory option: allow the result to have a different length than the source.
     */
    public static final int LENGTH_GROW_SHRINK = 0;

    /**
     * Memory option: the result must have the same length as the source.
     * If more room is necessary, then try to consume spaces next to modified characters.
     */
    public static final int LENGTH_FIXED_SPACES_NEAR = 1;

    /**
     * Memory option: the result must have the same length as the source.
     * If more room is necessary, then try to consume spaces at the end of the text.
     */
    public static final int LENGTH_FIXED_SPACES_AT_END = 2;

    /**
     * Memory option: the result must have the same length as the source.
     * If more room is necessary, then try to consume spaces at the beginning of the text.
     */
    public static final int LENGTH_FIXED_SPACES_AT_BEGINNING = 3;

    /** 
     * Bit mask for memory options. 
     */
    public static final int LENGTH_MASK = 3;


    /** 
     * Direction indicator: the source is in logical (keyboard) order. 
     */
    public static final int TEXT_DIRECTION_LOGICAL = 0;

    /** 
     * Direction indicator: the source is in visual (display) order, that is,
     * the leftmost displayed character is stored first.
     */
    public static final int TEXT_DIRECTION_VISUAL_LTR = 4;

    /** 
     * Bit mask for direction indicators. 
     */
    public static final int TEXT_DIRECTION_MASK = 4;


    /**
     * Letter shaping option: do not perform letter shaping. 
     */
    public static final int LETTERS_NOOP = 0;

    /** 
     * Letter shaping option: replace normative letter characters in the U+0600 (Arabic) block,
     * by shaped ones in the U+FE70 (Presentation Forms B) block. Performs Lam-Alef ligature
     * substitution.
     */
    public static final int LETTERS_SHAPE = 8;

    /** 
     * Letter shaping option: replace shaped letter characters in the U+FE70 (Presentation Forms B) block
     * by normative ones in the U+0600 (Arabic) block.  Converts Lam-Alef ligatures to pairs of Lam and
     * Alef characters, consuming spaces if required.
     */
    public static final int LETTERS_UNSHAPE = 0x10;

    /**
     * Letter shaping option: replace normative letter characters in the U+0600 (Arabic) block,
     * except for the TASHKEEL characters at U+064B...U+0652, by shaped ones in the U+Fe70
     * (Presentation Forms B) block.  The TASHKEEL characters will always be converted to
     * the isolated forms rather than to their correct shape.
     */
    public static final int LETTERS_SHAPE_TASHKEEL_ISOLATED = 0x18;

    /** 
     * Bit mask for letter shaping options. 
     */
    public static final int LETTERS_MASK = 0x18;


    /** 
     * Digit shaping option: do not perform digit shaping. 
     */
    public static final int DIGITS_NOOP = 0;

    /**
     * Digit shaping option: Replace European digits (U+0030...U+0039) by Arabic-Indic digits.
     */
    public static final int DIGITS_EN2AN = 0x20;

    /**
     * Digit shaping option: Replace Arabic-Indic digits by European digits (U+0030...U+0039).
     */
    public static final int DIGITS_AN2EN = 0x40;

    /**
     * Digit shaping option:
     * Replace European digits (U+0030...U+0039) by Arabic-Indic digits
     * if the most recent strongly directional character
     * is an Arabic letter (its Bidi direction value is RIGHT_TO_LEFT_ARABIC). 
     * The initial state at the start of the text is assumed to be not an Arabic,
     * letter, so European digits at the start of the text will not change.
     * Compare to DIGITS_ALEN2AN_INIT_AL.
     */
    public static final int DIGITS_EN2AN_INIT_LR = 0x60;

    /**
     * Digit shaping option:
     * Replace European digits (U+0030...U+0039) by Arabic-Indic digits
     * if the most recent strongly directional character
     * is an Arabic letter (its Bidi direction value is RIGHT_TO_LEFT_ARABIC). 
     * The initial state at the start of the text is assumed to be an Arabic,
     * letter, so European digits at the start of the text will change.
     * Compare to DIGITS_ALEN2AN_INT_LR.
     */
    public static final int DIGITS_EN2AN_INIT_AL = 0x80;

    /** Not a valid option value. */
    private static final int DIGITS_RESERVED = 0xa0;

    /** 
     * Bit mask for digit shaping options. 
     */
    public static final int DIGITS_MASK = 0xe0;

    /** 
     * Digit type option: Use Arabic-Indic digits (U+0660...U+0669). 
     */
    public static final int DIGIT_TYPE_AN = 0;

    /** 
     * Digit type option: Use Eastern (Extended) Arabic-Indic digits (U+06f0...U+06f9). 
     */
    public static final int DIGIT_TYPE_AN_EXTENDED = 0x100;

    /** 
     * Bit mask for digit type options. 
     */
    public static final int DIGIT_TYPE_MASK = 0x0100; // 0x3f00?

    public boolean equals(Object rhs) {
        return rhs != null && 
            rhs.getClass() == ArabicShaping.class && 
            options == ((ArabicShaping)rhs).options;
    }

    public int hashCode() {
        return options;
    }

    public String toString() {
        StringBuffer buf = new StringBuffer(super.toString());
        buf.append('[');
        switch (options & LENGTH_MASK) {
        case LENGTH_GROW_SHRINK: buf.append("grow/shrink"); break;
        case LENGTH_FIXED_SPACES_NEAR: buf.append("spaces near"); break;
        case LENGTH_FIXED_SPACES_AT_END: buf.append("spaces at end"); break;
        case LENGTH_FIXED_SPACES_AT_BEGINNING: buf.append("spaces at beginning"); break;
        }
        switch (options & TEXT_DIRECTION_MASK) {
        case TEXT_DIRECTION_LOGICAL: buf.append(", logical"); break;
        case TEXT_DIRECTION_VISUAL_LTR: buf.append(", visual"); break;
        }
        switch (options & LETTERS_MASK) {
        case LETTERS_NOOP: buf.append(", no letter shaping"); break;
        case LETTERS_SHAPE: buf.append(", shape letters"); break;
        case LETTERS_SHAPE_TASHKEEL_ISOLATED: buf.append(", shape letters tashkeel isolated"); break;
        case LETTERS_UNSHAPE: buf.append(", unshape letters"); break;
        }
        switch (options & DIGITS_MASK) {
        case DIGITS_NOOP: buf.append(", no digit shaping"); break;
        case DIGITS_EN2AN: buf.append(", shape digits to AN"); break;
        case DIGITS_AN2EN: buf.append(", shape digits to EN"); break;
        case DIGITS_EN2AN_INIT_LR: buf.append(", shape digits to AN contextually: default EN"); break;
        case DIGITS_EN2AN_INIT_AL: buf.append(", shape digits to AN contextually: default AL"); break;
        }
        switch (options & DIGIT_TYPE_MASK) {
        case DIGIT_TYPE_AN: buf.append(", standard Arabic-Indic digits"); break;
        case DIGIT_TYPE_AN_EXTENDED: buf.append(", extended Arabic-Indic digits"); break;
        }
        buf.append("]");

        return buf.toString();
    }

    //
    // ported api
    //

    private static final int IRRELEVANT = 4;
    private static final int LAMTYPE = 16;
    private static final int ALEFTYPE = 32;

    private static final int LINKR = 1;
    private static final int LINKL = 2;
    private static final int LINK_MASK = 3;

    private static final int irrelevantPos[] = { 
        0x0, 0x2, 0x4, 0x6, 0x8, 0xA, 0xC, 0xE 
    };

    private static final char convertLamAlef[] =  {
        '\u0622', // FEF5 
        '\u0622', // FEF6
        '\u0623', // FEF7
        '\u0623', // FEF8
        '\u0625', // FEF9
        '\u0625', // FEFA
        '\u0627', // FEFB
        '\u0627'  // FEFC 
    };

    private static final char convertNormalizedLamAlef[] = {
        '\u0622', // 065C
        '\u0623', // 065D
        '\u0625', // 065E
        '\u0627', // 065F
    };

    private static final int[] araLink = {
        1           + 32 + 256 * 0x11,  /*0x0622*/
        1           + 32 + 256 * 0x13,  /*0x0623*/
        1                + 256 * 0x15,  /*0x0624*/
        1           + 32 + 256 * 0x17,  /*0x0625*/
        1 + 2            + 256 * 0x19,  /*0x0626*/
        1           + 32 + 256 * 0x1D,  /*0x0627*/
        1 + 2            + 256 * 0x1F,  /*0x0628*/
        1                + 256 * 0x23,  /*0x0629*/
        1 + 2            + 256 * 0x25,  /*0x062A*/
        1 + 2            + 256 * 0x29,  /*0x062B*/
        1 + 2            + 256 * 0x2D,  /*0x062C*/
        1 + 2            + 256 * 0x31,  /*0x062D*/
        1 + 2            + 256 * 0x35,  /*0x062E*/