com.ibm.lpex.bidi
Class BidiTransform

java.lang.Object
  com.ibm.lpex.bidi.BidiTransform

public class BidiTransform
extends Object

The BidiTransform class defines Text Layout transformations.

Layout transformations allow converting a given instance of BidiText (the source) into another instance (the destination) with possibly different Bidi flags while conserving the semantics of the text.

Bidi text is a combination of a sequence of characters and a set of Bidi flags. Bidi text is implemented by the BidiText class.

A BidiTransform instance contains fields which define what auxiliary outputs are required from the transformation, and fields to refer to these outputs. The work itself is done by the transform method of the BidiText class.

Boolean fields are used to specify options of the transform operation. Other fields are used to store auxiliary outputs of the transformation.

There are no specific constructors for this class. The default constructor creates a default BidiFlagSet and puts its reference in the "flags" instance member. Unless modified later, this qualifies a transformation to Implicit LTR swapped numerals-nominal unshaped text.

Multi-threading considerations: each thread must use its own instances of this class.

Field Summary

String	delimiters Option: delimiters for performing the Bidi transformation.
int[]	dstToSrcMap Output value: destination-to-source map from the last transform with dstToSrcMapRequired specified; if this option was not specified, the content of dstToSrcMap should be ignored.
boolean	dstToSrcMapRequired Option: create a destination to source mapping
BidiFlagSet	flags Option: Bidi flags of destination BidiText
boolean	impToImp Option: enable Implicit to Implicit transformations
boolean	insertMarkers Option: add Markers to destination text when needed for round trip.
ArabicOptionSet	options Option: shaping options for this transformation
byte[]	propertyMap Output value: property map from the last transform with propertyMapRequired specified; if this option was not specified, the content of propertyMap should be ignored.
boolean	propertyMapRequired Option: create a property map
boolean	removeMarkers Option: remove Markers from destination text after performing a Bidi transformation.
boolean	roundTrip Option: use "roundtrip" algorithm for reordering
int[]	srcToDstMap Output value: source-to-destination map from the last transform with srcToDstMapRequired specified; if this option was not specified, the content of srcToDstMap should be ignored.
boolean	srcToDstMapRequired Option: create a source to destination mapping
boolean	winCompatible Option: use "Windows compatible" algorithm for reordering.
boolean	wordBreak Option: consider white space to always follow base orientation

Constructor Summary

BidiTransform()

Method Summary

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

flags

public BidiFlagSet flags

Option: Bidi flags of destination BidiText

impToImp

public boolean impToImp

Option: enable Implicit to Implicit transformations

roundTrip

public boolean roundTrip

Option: use "roundtrip" algorithm for reordering

winCompatible

public boolean winCompatible

Option: use "Windows compatible" algorithm for reordering.

If this option is true, the reordering algorithm is modified to perform more closely like Windows. In particular, logical string "12ABC" in LTR orientation (where ABC represent Arabic or Hebrew letters) is reordered as "CBA12" instead of "12CBA". Also, logical string "abc 123 45" (where all digits represent Hindi numbers) is reordered as "abc 123 45" instead of "abc 45 123".

insertMarkers

public boolean insertMarkers

Option: add Markers to destination text when needed for round trip.

If this option is true, LRM and RLM markers may be inserted when transforming from visual to logical where needed to insure round trip.

This option is ignored if the removeMarkers option is true.

removeMarkers

public boolean removeMarkers

Option: remove Markers from destination text after performing a Bidi transformation.

If this option is true, LRM and RLM markers in the source text participate in the reordering, but they are removed from the destination text after performing the Bidi transformation.

delimiters

public String delimiters

Option: delimiters for performing the Bidi transformation.

This option is enabled if the string "delimiters" is not null and has a length of at least one character. This string specifies starting and ending delimiters such that the Bidi transformation is only performed on the data between starting and ending delimiters.

If the length of the string is only one character, the ending delimiter is assumed equal to the starting delimiter.

An even number of successive ending delimiters in the data to transform is considered part of the data and do not function as delimiters. In an odd number of successive ending delimiters, the last one is considered as a delimiter.

options

public ArabicOptionSet options

Option: shaping options for this transformation

wordBreak

public boolean wordBreak

Option: consider white space to always follow base orientation

srcToDstMapRequired

public boolean srcToDstMapRequired

Option: create a source to destination mapping

dstToSrcMapRequired

public boolean dstToSrcMapRequired

Option: create a destination to source mapping

propertyMapRequired

public boolean propertyMapRequired

Option: create a property map

Note: the map created is of no practical value if impToImp is true.

srcToDstMap

public int[] srcToDstMap

Output value: source-to-destination map from the last transform with srcToDstMapRequired specified; if this option was not specified, the content of srcToDstMap should be ignored.

If when starting a transformation this field refers to a large enough array of integers, this array will be re-used to put the new map. Otherwise a new array will be created.

This map has a number for each character processed in the source data by the last transform. This number is the index of where this character is moved in the character array of the destination BidiText. If the removeMarkers option was specified and LRM or RLM markers have been removed from the destination text, the corresponding elements of srcToDstMap will contain -1.

Note that the allocated array may have more elements than the number of characters processed in the source BidiText. In that case, the extra elements should be ignored. The number of relevant elements can be found in inpCount.

dstToSrcMap

public int[] dstToSrcMap

Output value: destination-to-source map from the last transform with dstToSrcMapRequired specified; if this option was not specified, the content of dstToSrcMap should be ignored.

If when starting a transformation this field refers to a large enough array of integers, this array will be re-used to put the new map. Otherwise a new array will be created.

This map has a number for each character in the "interesting" data of the destination BidiText. This number is the index of the source character from which the destination character originates. This index is relative to the beginning of the "interesting" data. If the offset of the source BidiText is not zero, index 0 does not indicate the first character of the data array, but the character at position "offset". If the insertMarkers option was specified and LRM or RLM markers have been added, the corresponding elements of dstToSrcMap will contain -1.

Note that the allocated array may have more elements than the number of characters in the "interesting" part of the destination BidiText. In that case, the extra elements should be ignored. The number of relevant elements can be found in outCount.

propertyMap

public byte[] propertyMap

Output value: property map from the last transform with propertyMapRequired specified; if this option was not specified, the content of propertyMap should be ignored.

If when starting a transformation this field refers to a large enough array of bytes, this array will be re-used to put the new map. Otherwise a new array will be created.

This map has a byte for each character processed in the source data by the last transform. The 6 lower bits of each property element is the Bidi level of the corresponding input character. The highest bit is a new-cell indicator for composed character environments: a value of 0 indicates a zero-length composing character element, and a value of 1 indicates an element that begins a new cell.

Note: the content of this map has no simple interpretation if impToImp is true.

Note also that the allocated array may have more elements than the number of characters processed in the source BidiText. In that case, the extra elements should be ignored. The number of relevant elements can be found in inpCount.

Constructor Detail

BidiTransform

public BidiTransform()