Class BidiOrder
This implementation is not optimized for performance. It is intended as a reference implementation that closely follows the specification of the Bidirectional Algorithm in The Unicode Standard version 3.0.
Input:
There are two levels of input to the algorithm, since clients may prefer
to supply some information from out-of-band sources rather than relying on
the default behavior.
- unicode type array
- unicode type array, with externally supplied base line direction
Output:
Output is separated into several stages as well, to better enable clients
to evaluate various aspects of implementation conformance.
- levels array over entire paragraph
- reordering array over entire paragraph
- levels array over line
- reordering array over line
As the algorithm is defined to operate on a single paragraph at a time, this implementation is written to handle single paragraphs. Thus rule P1 is presumed by this implementation-- the data provided to the implementation is assumed to be a single paragraph, and either contains no 'B' codes, or a single 'B' code at the end of the input. 'B' is allowed as input to illustrate how the algorithm assigns it a level.
Also note that rules L3 and L4 depend on the rendering engine that uses the result of the bidi algorithm. This implementation assumes that the rendering engine expects combining marks in visual order (e.g. to the left of their base character in RTL runs) and that it adjust the glyphs used to render mirrored characters that are in RTL runs so that they render appropriately.
-
Field Summary
FieldsModifier and TypeFieldDescriptionstatic final byte
Right-to-Left Arabicprivate static final byte
Arabic Numberprivate static final byte
Paragraph Separatorprivate static char[]
private static final byte
Boundary Neutralprivate static final byte
Common Number Separatorprivate byte[]
static final byte
European Numberprivate static final byte
European Number Separatorprivate static final byte
European Number Terminatorprivate final byte[]
static final byte
Left-to-rightprivate static final byte
Left-to-Right Embeddingprivate static final byte
Left-to-Right Overrideprivate static final byte
Non-Spacing Markprivate static final byte
Other Neutralsprivate byte
private static final byte
Pop Directional Formatstatic final byte
Right-to-Leftprivate byte[]
private byte[]
private static final byte
Right-to-Left Embeddingprivate static final byte
Right-to-Left Overrideprivate static final byte[]
private static final byte
Segment Separatorprivate int
private static final byte
Maximum bidi type value.private static final byte
Minimum bidi type value.private static final byte
Whitespace -
Constructor Summary
ConstructorsConstructorDescriptionBidiOrder
(char[] text, int offset, int length, byte paragraphEmbeddingLevel) -
Method Summary
Modifier and TypeMethodDescriptionprivate static int[]
computeMultilineReordering
(byte[] levels, int[] linebreaks) Return multiline reordering array for a given level array.private static int[]
computeReordering
(byte[] levels) Return reordering array for a given level array.private void
Process embedding format codes.private void
1) determining the paragraph level.private int
findRunLimit
(int index, int limit, byte[] validSet) Return the limit of the run starting at index that includes only resultTypes in validSet.private int
findRunStart
(int index, byte[] validSet) Return the start of the run including index that includes only resultTypes in validSet.byte
Return the base level of the paragraph.(package private) static final byte
getDirection
(char c) byte[]
private byte[]
getLevels
(int[] linebreaks) Return levels array breaking lines at offsets in linebreaks.private static boolean
isWhitespace
(byte biditype) Return true if the type is considered a whitespace type for the line break rules.private static byte[]
processEmbeddings
(byte[] resultTypes, byte paragraphEmbeddingLevel) 2) determining explicit levels Rules X1 - X8 The interaction of these rules makes handling them a bit complex.private int
reinsertExplicitCodes
(int textLength) Reinsert levels information for explicit codes.private int
Rules X9.private void
resolveImplicitLevels
(int start, int limit, byte level, byte sor, byte eor) 7) resolving implicit embedding levels Rules I1, I2.private void
resolveNeutralTypes
(int start, int limit, byte level, byte sor, byte eor) 6) resolving neutral types Rules N1-N2.private void
resolveWeakTypes
(int start, int limit, byte level, byte sor, byte eor) 3) resolving weak types Rules W1-W7.private void
The algorithm.private void
setLevels
(int start, int limit, byte newLevel) Set resultLevels from start up to (but not including) limit to newLevel.private void
setTypes
(int start, int limit, byte newType) Set resultTypes from start up to (but not including) limit to newType.private static byte
typeForLevel
(int level) Return the strong type (L or R) corresponding to the level.private static void
validateLineBreaks
(int[] linebreaks, int textLength) Throw exception if line breaks array is invalid.private static void
validateParagraphEmbeddingLevel
(byte paragraphEmbeddingLevel) Throw exception if paragraph embedding level is invalid.private static void
validateTypes
(byte[] types) Throw exception if type array is invalid.
-
Field Details
-
initialTypes
private final byte[] initialTypes -
embeddings
private byte[] embeddings -
paragraphEmbeddingLevel
private byte paragraphEmbeddingLevel -
textLength
private int textLength -
resultTypes
private byte[] resultTypes -
resultLevels
private byte[] resultLevels -
L
public static final byte LLeft-to-right- See Also:
-
LRE
private static final byte LRELeft-to-Right Embedding- See Also:
-
LRO
private static final byte LROLeft-to-Right Override- See Also:
-
R
public static final byte RRight-to-Left- See Also:
-
AL
public static final byte ALRight-to-Left Arabic- See Also:
-
RLE
private static final byte RLERight-to-Left Embedding- See Also:
-
RLO
private static final byte RLORight-to-Left Override- See Also:
-
PDF
private static final byte PDFPop Directional Format- See Also:
-
EN
public static final byte ENEuropean Number- See Also:
-
ES
private static final byte ESEuropean Number Separator- See Also:
-
ET
private static final byte ETEuropean Number Terminator- See Also:
-
AN
private static final byte ANArabic Number- See Also:
-
CS
private static final byte CSCommon Number Separator- See Also:
-
NSM
private static final byte NSMNon-Spacing Mark- See Also:
-
BN
private static final byte BNBoundary Neutral- See Also:
-
B
private static final byte BParagraph Separator- See Also:
-
S
private static final byte SSegment Separator- See Also:
-
WS
private static final byte WSWhitespace- See Also:
-
ON
private static final byte ONOther Neutrals- See Also:
-
TYPE_MIN
private static final byte TYPE_MINMinimum bidi type value.- See Also:
-
TYPE_MAX
private static final byte TYPE_MAXMaximum bidi type value.- See Also:
-
rtypes
private static final byte[] rtypes -
baseTypes
private static char[] baseTypes
-
-
Constructor Details
-
BidiOrder
BidiOrder(char[] text, int offset, int length, byte paragraphEmbeddingLevel)
-
-
Method Details
-
getDirection
static final byte getDirection(char c) -
runAlgorithm
private void runAlgorithm()The algorithm. Does not include line-based processing (Rules L1, L2). These are applied later in the line-based phase of the algorithm. -
determineParagraphEmbeddingLevel
private void determineParagraphEmbeddingLevel()1) determining the paragraph level.Rules P2, P3.
At the end of this function, the member variable paragraphEmbeddingLevel is set to either 0 or 1.
-
determineExplicitEmbeddingLevels
private void determineExplicitEmbeddingLevels()Process embedding format codes.Calls processEmbeddings to generate an embedding array from the explicit format codes. The embedding overrides in the array are then applied to the result types, and the result levels are initialized.
- See Also:
-
removeExplicitCodes
private int removeExplicitCodes()Rules X9. Remove explicit codes so that they may be ignored during the remainder of the main portion of the algorithm. The length of the resulting text is returned.- Returns:
- the length of the data excluding explicit codes and BN.
-
reinsertExplicitCodes
private int reinsertExplicitCodes(int textLength) Reinsert levels information for explicit codes. This is for ease of relating the level information to the original input data. Note that the levels assigned to these codes are arbitrary, they're chosen so as to avoid breaking level runs.- Parameters:
textLength
- the length of the data after compression- Returns:
- the length of the data (original length of types array supplied to constructor)
-
processEmbeddings
private static byte[] processEmbeddings(byte[] resultTypes, byte paragraphEmbeddingLevel) 2) determining explicit levels Rules X1 - X8 The interaction of these rules makes handling them a bit complex. This examines resultTypes but does not modify it. It returns embedding and override information in the result array. The low 7 bits are the level, the high bit is set if the level is an override, and clear if it is an embedding. -
resolveWeakTypes
private void resolveWeakTypes(int start, int limit, byte level, byte sor, byte eor) 3) resolving weak types Rules W1-W7. Note that some weak types (EN, AN) remain after this processing is complete. -
resolveNeutralTypes
private void resolveNeutralTypes(int start, int limit, byte level, byte sor, byte eor) 6) resolving neutral types Rules N1-N2. -
resolveImplicitLevels
private void resolveImplicitLevels(int start, int limit, byte level, byte sor, byte eor) 7) resolving implicit embedding levels Rules I1, I2. -
getLevels
public byte[] getLevels() -
getLevels
private byte[] getLevels(int[] linebreaks) Return levels array breaking lines at offsets in linebreaks.
Rule L1.The returned levels array contains the resolved level for each bidi code passed to the constructor.
The linebreaks array must include at least one value. The values must be in strictly increasing order (no duplicates) between 1 and the length of the text, inclusive. The last value must be the length of the text.
- Parameters:
linebreaks
- the offsets at which to break the paragraph- Returns:
- the resolved levels of the text
-
computeMultilineReordering
private static int[] computeMultilineReordering(byte[] levels, int[] linebreaks) Return multiline reordering array for a given level array. Reordering does not occur across a line break. -
computeReordering
private static int[] computeReordering(byte[] levels) Return reordering array for a given level array. This reorders a single line. The reordering is a visual to logical map. For example, the leftmost char is string.charAt(order[0]). Rule L2. -
getBaseLevel
public byte getBaseLevel()Return the base level of the paragraph. -
isWhitespace
private static boolean isWhitespace(byte biditype) Return true if the type is considered a whitespace type for the line break rules. -
typeForLevel
private static byte typeForLevel(int level) Return the strong type (L or R) corresponding to the level. -
findRunLimit
private int findRunLimit(int index, int limit, byte[] validSet) Return the limit of the run starting at index that includes only resultTypes in validSet. This checks the value at index, and will return index if that value is not in validSet. -
findRunStart
private int findRunStart(int index, byte[] validSet) Return the start of the run including index that includes only resultTypes in validSet. This assumes the value at index is valid, and does not check it. -
setTypes
private void setTypes(int start, int limit, byte newType) Set resultTypes from start up to (but not including) limit to newType. -
setLevels
private void setLevels(int start, int limit, byte newLevel) Set resultLevels from start up to (but not including) limit to newLevel. -
validateTypes
private static void validateTypes(byte[] types) Throw exception if type array is invalid. -
validateParagraphEmbeddingLevel
private static void validateParagraphEmbeddingLevel(byte paragraphEmbeddingLevel) Throw exception if paragraph embedding level is invalid. Special allowance for -1 so that default processing can still be performed when using this API. -
validateLineBreaks
private static void validateLineBreaks(int[] linebreaks, int textLength) Throw exception if line breaks array is invalid.
-