com.ibm.icu.text
Class ArabicShaping

java.lang.Object
  com.ibm.icu.text.ArabicShaping

public final class ArabicShaping
extends Object

Shape Arabic text on a character basis.

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.

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.

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.

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).

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.

Status:

Stable ICU 2.0.

Field Summary

static int	DIGIT_TYPE_AN Digit type option: Use Arabic-Indic digits (U+0660...U+0669).
static int	DIGIT_TYPE_AN_EXTENDED Digit type option: Use Eastern (Extended) Arabic-Indic digits (U+06f0...U+06f9).
static int	DIGIT_TYPE_MASK Bit mask for digit type options.
static int	DIGITS_AN2EN Digit shaping option: Replace Arabic-Indic digits by European digits (U+0030...U+0039).
static int	DIGITS_EN2AN Digit shaping option: Replace European digits (U+0030...U+0039) by Arabic-Indic digits.
static int	DIGITS_EN2AN_INIT_AL 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).
static int	DIGITS_EN2AN_INIT_LR 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).
static int	DIGITS_MASK Bit mask for digit shaping options.
static int	DIGITS_NOOP Digit shaping option: do not perform digit shaping.
static int	LAMALEF_AUTO Memory option: the result must have the same length as the source.
static int	LAMALEF_BEGIN Memory option: the result must have the same length as the source.
static int	LAMALEF_END Memory option: the result must have the same length as the source.
static int	LAMALEF_MASK Bit mask for LamAlef memory options.
static int	LAMALEF_NEAR Memory option: the result must have the same length as the source.
static int	LAMALEF_RESIZE Memory option: allow the result to have a different length than the source.
static int	LENGTH_FIXED_SPACES_AT_BEGINNING Memory option: the result must have the same length as the source.
static int	LENGTH_FIXED_SPACES_AT_END Memory option: the result must have the same length as the source.
static int	LENGTH_FIXED_SPACES_NEAR Memory option: the result must have the same length as the source.
static int	LENGTH_GROW_SHRINK Memory option: allow the result to have a different length than the source.
static int	LENGTH_MASK Bit mask for memory options.
static int	LETTERS_MASK Bit mask for letter shaping options.
static int	LETTERS_NOOP Letter shaping option: do not perform letter shaping.
static int	LETTERS_SHAPE Letter shaping option: replace normative letter characters in the U+0600 (Arabic) block, by shaped ones in the U+FE70 (Presentation Forms B) block.
static int	LETTERS_SHAPE_TASHKEEL_ISOLATED 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.
static int	LETTERS_UNSHAPE Letter shaping option: replace shaped letter characters in the U+FE70 (Presentation Forms B) block by normative ones in the U+0600 (Arabic) block.
static int	SEEN_MASK Bit mask for Seen memory options.
static int	SEEN_TWOCELL_NEAR Memory option: the result must have the same length as the source.
static int	SHAPE_TAIL_NEW_UNICODE If this option is used, shaping will use the new Unicode code point for TAIL (i.e. 0xFE73).
static int	SHAPE_TAIL_TYPE_MASK Bit mask for new Unicode Tail option
static int	SPACES_RELATIVE_TO_TEXT_BEGIN_END This option effects the meaning of BEGIN and END options. if this option is not used the default for BEGIN and END will be as following: The Default (for both Visual LTR, Visual RTL and Logical Text) 1.
static int	SPACES_RELATIVE_TO_TEXT_MASK Bit mask for swapping BEGIN and END for Visual LTR text
static int	TASHKEEL_BEGIN Memory option: the result must have the same length as the source.
static int	TASHKEEL_END Memory option: the result must have the same length as the source.
static int	TASHKEEL_MASK Bit mask for Tashkeel replacement with Space or Tatweel memory options.
static int	TASHKEEL_REPLACE_BY_TATWEEL Memory option: the result must have the same length as the source.
static int	TASHKEEL_RESIZE Memory option: allow the result to have a different length than the source.
static int	TEXT_DIRECTION_LOGICAL Direction indicator: the source is in logical (keyboard) order.
static int	TEXT_DIRECTION_MASK Bit mask for direction indicators.
static int	TEXT_DIRECTION_VISUAL_LTR Direction indicator: the source is in visual (display) order, that is, the leftmost displayed character is stored first.
static int	TEXT_DIRECTION_VISUAL_RTL Direction indicator:the source is in visual RTL order, the rightmost displayed character stored first.
static int	YEHHAMZA_MASK Bit mask for YehHamza memory options.
static int	YEHHAMZA_TWOCELL_NEAR Memory option: the result must have the same length as the source.

Constructor Summary

ArabicShaping(int options)
Construct ArabicShaping using the options flags.

Method Summary

boolean	equals(Object rhs)
int	hashCode()
void	shape(char[] source, int start, int length) Convert a range of text in place.
int	shape(char[] source, int sourceStart, int sourceLength, char[] dest, int destStart, int destSize) 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.
String	shape(String text) Convert a string, returning the new string.
String	toString()

Methods inherited from class java.lang.Object

clone, finalize, getClass, notify, notifyAll, wait, wait, wait

Field Detail

SEEN_TWOCELL_NEAR

public static final int SEEN_TWOCELL_NEAR

Memory option: the result must have the same length as the source. Shaping mode: The SEEN family character will expand into two characters using space near the SEEN family character(i.e. the space after the character). if there are no spaces found, ArabicShapingException will be thrown De-shaping mode: Any Seen character followed by Tail character will be replaced by one cell Seen and a space will replace the Tail. Affects: Seen options

See Also:

Constant Field Values

Status:

Stable ICU 4.2.

SEEN_MASK

public static final int SEEN_MASK

Bit mask for Seen memory options.

See Also:

Constant Field Values

Status:

Stable ICU 4.2.

YEHHAMZA_TWOCELL_NEAR

public static final int YEHHAMZA_TWOCELL_NEAR

Memory option: the result must have the same length as the source. Shaping mode: The YEHHAMZA character will expand into two characters using space near it (i.e. the space after the character) if there are no spaces found, ArabicShapingException will be thrown De-shaping mode: Any Yeh (final or isolated) character followed by Hamza character will be replaced by one cell YehHamza and space will replace the Hamza. Affects: YehHamza options

See Also:

Constant Field Values

Status:

Stable ICU 4.2.

YEHHAMZA_MASK

public static final int YEHHAMZA_MASK

Bit mask for YehHamza memory options.

See Also:

Constant Field Values

Status:

Stable ICU 4.2.

TASHKEEL_BEGIN

public static final int TASHKEEL_BEGIN

Memory option: the result must have the same length as the source. Shaping mode: Tashkeel characters will be replaced by spaces. Spaces will be placed at beginning of the buffer De-shaping mode: N/A Affects: Tashkeel options

See Also:

Constant Field Values

Status:

Stable ICU 4.2.

TASHKEEL_END

public static final int TASHKEEL_END

Memory option: the result must have the same length as the source. Shaping mode: Tashkeel characters will be replaced by spaces. Spaces will be placed at end of the buffer De-shaping mode: N/A Affects: Tashkeel options

See Also:

Constant Field Values

Status:

Stable ICU 4.2.

TASHKEEL_RESIZE

public static final int TASHKEEL_RESIZE

Memory option: allow the result to have a different length than the source. Shaping mode: Tashkeel characters will be removed, buffer length will shrink. De-shaping mode: N/A Affects: Tashkeel options

See Also:

Constant Field Values

Status:

Stable ICU 4.2.

TASHKEEL_REPLACE_BY_TATWEEL

public static final int TASHKEEL_REPLACE_BY_TATWEEL

Memory option: the result must have the same length as the source. Shaping mode: Tashkeel characters will be replaced by Tatweel if it is connected to adjacent characters (i.e. shaped on Tatweel) or replaced by space if it is not connected. De-shaping mode: N/A Affects: YehHamza options

See Also:

Constant Field Values

Status:

Stable ICU 4.2.

TASHKEEL_MASK

public static final int TASHKEEL_MASK

Bit mask for Tashkeel replacement with Space or Tatweel memory options.

See Also:

Constant Field Values

Status:

Stable ICU 4.2.

SPACES_RELATIVE_TO_TEXT_BEGIN_END

public static final int SPACES_RELATIVE_TO_TEXT_BEGIN_END

This option effects the meaning of BEGIN and END options. if this option is not used the default for BEGIN and END will be as following: The Default (for both Visual LTR, Visual RTL and Logical Text) 1. BEGIN always refers to the start address of physical memory. 2. END always refers to the end address of physical memory. If this option is used it will swap the meaning of BEGIN and END only for Visual LTR text. The affect on BEGIN and END Memory Options will be as following: A. BEGIN For Visual LTR text: This will be the beginning (right side) of the visual text (corresponding to the physical memory address end, same as END in default behavior) B. BEGIN For Logical text: Same as BEGIN in default behavior. C. END For Visual LTR text: This will be the end (left side) of the visual text. (corresponding to the physical memory address beginning, same as BEGIN in default behavior) D. END For Logical text: Same as END in default behavior. Affects: All LamAlef BEGIN, END and AUTO options.

See Also:

Constant Field Values

Status:

Stable ICU 4.2.

SPACES_RELATIVE_TO_TEXT_MASK

public static final int SPACES_RELATIVE_TO_TEXT_MASK

Bit mask for swapping BEGIN and END for Visual LTR text

See Also:

Constant Field Values

Status:

Stable ICU 4.2.

SHAPE_TAIL_NEW_UNICODE

public static final int SHAPE_TAIL_NEW_UNICODE

If this option is used, shaping will use the new Unicode code point for TAIL (i.e. 0xFE73). If this option is not specified (Default), old unofficial Unicode TAIL code point is used (i.e. 0x200B) De-shaping will not use this option as it will always search for both the new Unicode code point for the TAIL (i.e. 0xFE73) or the old unofficial Unicode TAIL code point (i.e. 0x200B) and de-shape the Seen-Family letter accordingly. Shaping Mode: Only shaping. De-shaping Mode: N/A. Affects: All Seen options

See Also:

Constant Field Values

Status:

Stable ICU 4.2.

SHAPE_TAIL_TYPE_MASK

public static final int SHAPE_TAIL_TYPE_MASK

Bit mask for new Unicode Tail option

See Also:

Constant Field Values

Status:

Stable ICU 4.2.

LENGTH_GROW_SHRINK

public static final int LENGTH_GROW_SHRINK

Memory option: allow the result to have a different length than the source.

See Also:

Constant Field Values

Status:

Stable ICU 2.0.

LAMALEF_RESIZE

public static final int LAMALEF_RESIZE

Memory option: allow the result to have a different length than the source. Affects: LamAlef options This option is an alias to LENGTH_GROW_SHRINK

See Also:

Constant Field Values

Status:

Stable ICU 4.2.

LENGTH_FIXED_SPACES_NEAR

public static final int LENGTH_FIXED_SPACES_NEAR

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.

See Also:

Constant Field Values

Status:

Stable ICU 2.0.

LAMALEF_NEAR

public static final int LAMALEF_NEAR

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. Affects: LamAlef options This option is an alias to LENGTH_FIXED_SPACES_NEAR

See Also:

Constant Field Values

Status:

Stable ICU 4.2.

LENGTH_FIXED_SPACES_AT_END

public static final int LENGTH_FIXED_SPACES_AT_END

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.

See Also:

Constant Field Values

Status:

Stable ICU 2.0.

LAMALEF_END

public static final int LAMALEF_END

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. Affects: LamAlef options This option is an alias to LENGTH_FIXED_SPACES_AT_END

See Also:

Constant Field Values

Status:

Stable ICU 4.2.

LENGTH_FIXED_SPACES_AT_BEGINNING

public static final int LENGTH_FIXED_SPACES_AT_BEGINNING

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.

See Also:

Constant Field Values

Status:

Stable ICU 2.0.

LAMALEF_BEGIN

public static final int LAMALEF_BEGIN

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. Affects: LamAlef options This option is an alias to LENGTH_FIXED_SPACES_AT_BEGINNING

See Also:

Constant Field Values

Status:

Stable ICU 4.2.

LAMALEF_AUTO

public static final int LAMALEF_AUTO

Memory option: the result must have the same length as the source. Shaping Mode: For each LAMALEF character found, expand LAMALEF using space at end. If there is no space at end, use spaces at beginning of the buffer. If there is no space at beginning of the buffer, use spaces at the near (i.e. the space after the LAMALEF character). Deshaping Mode: Perform the same function as the flag equals LAMALEF_END. Affects: LamAlef options

See Also:

Constant Field Values

Status:

Stable ICU 4.2.

LENGTH_MASK

public static final int LENGTH_MASK

Bit mask for memory options.

See Also:

Constant Field Values

Status:

Stable ICU 2.0.

LAMALEF_MASK

public static final int LAMALEF_MASK

Bit mask for LamAlef memory options.

See Also:

Constant Field Values

Status:

Stable ICU 4.2.

TEXT_DIRECTION_LOGICAL

public static final int TEXT_DIRECTION_LOGICAL

Direction indicator: the source is in logical (keyboard) order.

See Also:

Constant Field Values

Status:

Stable ICU 2.0.

TEXT_DIRECTION_VISUAL_RTL

public static final int TEXT_DIRECTION_VISUAL_RTL

Direction indicator:the source is in visual RTL order, the rightmost displayed character stored first. This option is an alias to U_SHAPE_TEXT_DIRECTION_LOGICAL

See Also:

Constant Field Values

Status:

Stable ICU 4.2.

TEXT_DIRECTION_VISUAL_LTR

public static final int TEXT_DIRECTION_VISUAL_LTR

Direction indicator: the source is in visual (display) order, that is, the leftmost displayed character is stored first.

See Also:

Constant Field Values

Status:

Stable ICU 2.0.

TEXT_DIRECTION_MASK

public static final int TEXT_DIRECTION_MASK

Bit mask for direction indicators.

See Also:

Constant Field Values

Status:

Stable ICU 2.0.

LETTERS_NOOP

public static final int LETTERS_NOOP

Letter shaping option: do not perform letter shaping.

See Also:

Constant Field Values

Status:

Stable ICU 2.0.

LETTERS_SHAPE

public static final int LETTERS_SHAPE

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.

See Also:

Constant Field Values

Status:

Stable ICU 2.0.

LETTERS_UNSHAPE

public static final int LETTERS_UNSHAPE

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.

See Also:

Constant Field Values

Status:

Stable ICU 2.0.

LETTERS_SHAPE_TASHKEEL_ISOLATED

public static final int LETTERS_SHAPE_TASHKEEL_ISOLATED

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.

See Also:

Constant Field Values

Status:

Stable ICU 2.0.

LETTERS_MASK

public static final int LETTERS_MASK

Bit mask for letter shaping options.

See Also:

Constant Field Values

Status:

Stable ICU 2.0.

DIGITS_NOOP

public static final int DIGITS_NOOP

Digit shaping option: do not perform digit shaping.

See Also:

Constant Field Values

Status:

Stable ICU 2.0.

DIGITS_EN2AN

public static final int DIGITS_EN2AN

Digit shaping option: Replace European digits (U+0030...U+0039) by Arabic-Indic digits.

See Also:

Constant Field Values

Status:

Stable ICU 2.0.

DIGITS_AN2EN

public static final int DIGITS_AN2EN

Digit shaping option: Replace Arabic-Indic digits by European digits (U+0030...U+0039).

See Also:

Constant Field Values

Status:

Stable ICU 2.0.

DIGITS_EN2AN_INIT_LR

public static final int DIGITS_EN2AN_INIT_LR

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.

See Also:

Constant Field Values

Status:

Stable ICU 2.0.

DIGITS_EN2AN_INIT_AL

public static final int DIGITS_EN2AN_INIT_AL

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.

See Also:

Constant Field Values

Status:

Stable ICU 2.0.

DIGITS_MASK

public static final int DIGITS_MASK

Bit mask for digit shaping options.

See Also:

Constant Field Values

Status:

Stable ICU 2.0.

DIGIT_TYPE_AN

public static final int DIGIT_TYPE_AN

Digit type option: Use Arabic-Indic digits (U+0660...U+0669).

See Also:

Constant Field Values

Status:

Stable ICU 2.0.

DIGIT_TYPE_AN_EXTENDED

public static final int DIGIT_TYPE_AN_EXTENDED

Digit type option: Use Eastern (Extended) Arabic-Indic digits (U+06f0...U+06f9).

See Also:

Constant Field Values

Status:

Stable ICU 2.0.

DIGIT_TYPE_MASK

public static final int DIGIT_TYPE_MASK

Bit mask for digit type options.

See Also:

Constant Field Values

Status:

Stable ICU 2.0.

Constructor Detail

ArabicShaping

public ArabicShaping(int options)

Construct ArabicShaping using the options flags. The flags are as follows:
'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.
'TEXT_DIRECTION' flags control whether the text is read and written in visual order or in logical order.
'LETTERS_SHAPE' flags control whether conversion is to or from presentation forms.
'DIGITS' flags control whether digits are shaped, and whether from European to Arabic-Indic or vice-versa.
'DIGIT_TYPE' flags control whether standard or extended Arabic-Indic digits are used when performing digit conversion.

Status:

Stable ICU 2.0.

Method Detail

shape

public int shape(char[] source,
                 int sourceStart,
                 int sourceLength,
                 char[] dest,
                 int destStart,
                 int destSize)
          throws ArabicShapingException

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.

Parameters:

source - An array containing the input text

sourceStart - The start of the range of text to convert

sourceLength - The length of the range of text to convert

dest - The destination array that will receive the result. It may be NULL only if destSize is 0.

destStart - The start of the range of the destination buffer to use.

destSize - The size (capacity) of the destination buffer. If destSize 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.

Returns:

The number of chars written to the destination buffer. If an error occurs, then no output was written, or it may be incomplete.

Throws:

ArabicShapingException - if the text cannot be converted according to the options.

Status:

Stable ICU 2.0.

shape

public void shape(char[] source,
                  int start,
                  int length)
           throws ArabicShapingException

Convert a range of text in place. This may only be used if the Length option does not grow or shrink the text.

Parameters:

source - An array containing the input text

start - The start of the range of text to convert

length - The length of the range of text to convert

Throws:

ArabicShapingException - if the text cannot be converted according to the options.

Status:

Stable ICU 2.0.

shape

public String shape(String text)
             throws ArabicShapingException

Convert a string, returning the new string.

Parameters:

text - the string to convert

Returns:

the converted string

Throws:

ArabicShapingException - if the string cannot be converted according to the options.

Status:

Stable ICU 2.0.

equals

public boolean equals(Object rhs)

Overrides:

equals in class Object

Status:

Stable ICU 2.0.

hashCode

public int hashCode()

Overrides:

hashCode in class Object

Status:

Stable ICU 2.0.

toString

public String toString()

Overrides:

toString in class Object

Status:

Stable ICU 2.0.