ICU 70.1  70.1
normalizer2.h
Go to the documentation of this file.
1 // © 2016 and later: Unicode, Inc. and others.
2 // License & terms of use: http://www.unicode.org/copyright.html
3 /*
4 *******************************************************************************
5 *
6 * Copyright (C) 2009-2013, International Business Machines
7 * Corporation and others. All Rights Reserved.
8 *
9 *******************************************************************************
10 * file name: normalizer2.h
11 * encoding: UTF-8
12 * tab size: 8 (not used)
13 * indentation:4
14 *
15 * created on: 2009nov22
16 * created by: Markus W. Scherer
17 */
18 
19 #ifndef __NORMALIZER2_H__
20 #define __NORMALIZER2_H__
21 
27 #include "unicode/utypes.h"
28 
29 #if U_SHOW_CPLUSPLUS_API
30 
31 #if !UCONFIG_NO_NORMALIZATION
32 
33 #include "unicode/stringpiece.h"
34 #include "unicode/uniset.h"
35 #include "unicode/unistr.h"
36 #include "unicode/unorm2.h"
37 
38 U_NAMESPACE_BEGIN
39 
40 class ByteSink;
41 
86 public:
92 
104  static const Normalizer2 *
106 
118  static const Normalizer2 *
120 
132  static const Normalizer2 *
134 
146  static const Normalizer2 *
148 
160  static const Normalizer2 *
162 
184  static const Normalizer2 *
185  getInstance(const char *packageName,
186  const char *name,
187  UNormalization2Mode mode,
188  UErrorCode &errorCode);
189 
201  normalize(const UnicodeString &src, UErrorCode &errorCode) const {
202  UnicodeString result;
203  normalize(src, result, errorCode);
204  return result;
205  }
219  virtual UnicodeString &
221  UnicodeString &dest,
222  UErrorCode &errorCode) const = 0;
223 
246  virtual void
247  normalizeUTF8(uint32_t options, StringPiece src, ByteSink &sink,
248  Edits *edits, UErrorCode &errorCode) const;
249 
264  virtual UnicodeString &
266  const UnicodeString &second,
267  UErrorCode &errorCode) const = 0;
282  virtual UnicodeString &
284  const UnicodeString &second,
285  UErrorCode &errorCode) const = 0;
286 
300  virtual UBool
301  getDecomposition(UChar32 c, UnicodeString &decomposition) const = 0;
302 
327  virtual UBool
328  getRawDecomposition(UChar32 c, UnicodeString &decomposition) const;
329 
345  virtual UChar32
347 
356  virtual uint8_t
358 
373  virtual UBool
374  isNormalized(const UnicodeString &s, UErrorCode &errorCode) const = 0;
394  virtual UBool
396 
397 
414  quickCheck(const UnicodeString &s, UErrorCode &errorCode) const = 0;
415 
438  virtual int32_t
439  spanQuickCheckYes(const UnicodeString &s, UErrorCode &errorCode) const = 0;
440 
454  virtual UBool hasBoundaryBefore(UChar32 c) const = 0;
455 
470  virtual UBool hasBoundaryAfter(UChar32 c) const = 0;
471 
485  virtual UBool isInert(UChar32 c) const = 0;
486 };
487 
500 public:
511  FilteredNormalizer2(const Normalizer2 &n2, const UnicodeSet &filterSet) :
512  norm2(n2), set(filterSet) {}
513 
519 
533  virtual UnicodeString &
535  UnicodeString &dest,
536  UErrorCode &errorCode) const U_OVERRIDE;
537 
560  virtual void
561  normalizeUTF8(uint32_t options, StringPiece src, ByteSink &sink,
562  Edits *edits, UErrorCode &errorCode) const U_OVERRIDE;
563 
578  virtual UnicodeString &
580  const UnicodeString &second,
581  UErrorCode &errorCode) const U_OVERRIDE;
596  virtual UnicodeString &
598  const UnicodeString &second,
599  UErrorCode &errorCode) const U_OVERRIDE;
600 
612  virtual UBool
614 
626  virtual UBool
628 
639  virtual UChar32
641 
650  virtual uint8_t
652 
664  virtual UBool
665  isNormalized(const UnicodeString &s, UErrorCode &errorCode) const U_OVERRIDE;
685  virtual UBool
699  quickCheck(const UnicodeString &s, UErrorCode &errorCode) const U_OVERRIDE;
711  virtual int32_t
713 
723 
733 
741  virtual UBool isInert(UChar32 c) const U_OVERRIDE;
742 private:
743  UnicodeString &
744  normalize(const UnicodeString &src,
745  UnicodeString &dest,
746  USetSpanCondition spanCondition,
747  UErrorCode &errorCode) const;
748 
749  void
750  normalizeUTF8(uint32_t options, const char *src, int32_t length,
751  ByteSink &sink, Edits *edits,
752  USetSpanCondition spanCondition,
753  UErrorCode &errorCode) const;
754 
755  UnicodeString &
756  normalizeSecondAndAppend(UnicodeString &first,
757  const UnicodeString &second,
758  UBool doNormalize,
759  UErrorCode &errorCode) const;
760 
761  const Normalizer2 &norm2;
762  const UnicodeSet &set;
763 };
764 
765 U_NAMESPACE_END
766 
767 #endif // !UCONFIG_NO_NORMALIZATION
768 
769 #endif /* U_SHOW_CPLUSPLUS_API */
770 
771 #endif // __NORMALIZER2_H__
A ByteSink can be filled with bytes.
Definition: bytestream.h:53
Records lengths of string edits but not replacement text.
Definition: edits.h:80
Normalization filtered by a UnicodeSet.
Definition: normalizer2.h:499
virtual UBool isInert(UChar32 c) const U_OVERRIDE
Tests if the character is normalization-inert.
virtual UBool getRawDecomposition(UChar32 c, UnicodeString &decomposition) const U_OVERRIDE
Gets the raw decomposition mapping of c.
~FilteredNormalizer2()
Destructor.
virtual void normalizeUTF8(uint32_t options, StringPiece src, ByteSink &sink, Edits *edits, UErrorCode &errorCode) const U_OVERRIDE
Normalizes a UTF-8 string and optionally records how source substrings relate to changed and unchange...
virtual UBool isNormalizedUTF8(StringPiece s, UErrorCode &errorCode) const U_OVERRIDE
Tests if the UTF-8 string is normalized.
virtual UBool hasBoundaryBefore(UChar32 c) const U_OVERRIDE
Tests if the character always has a normalization boundary before it, regardless of context.
virtual UnicodeString & normalizeSecondAndAppend(UnicodeString &first, const UnicodeString &second, UErrorCode &errorCode) const U_OVERRIDE
Appends the normalized form of the second string to the first string (merging them at the boundary) a...
virtual UnicodeString & normalize(const UnicodeString &src, UnicodeString &dest, UErrorCode &errorCode) const U_OVERRIDE
Writes the normalized form of the source string to the destination string (replacing its contents) an...
virtual UnicodeString & append(UnicodeString &first, const UnicodeString &second, UErrorCode &errorCode) const U_OVERRIDE
Appends the second string to the first string (merging them at the boundary) and returns the first st...
virtual UBool getDecomposition(UChar32 c, UnicodeString &decomposition) const U_OVERRIDE
Gets the decomposition mapping of c.
virtual int32_t spanQuickCheckYes(const UnicodeString &s, UErrorCode &errorCode) const U_OVERRIDE
Returns the end of the normalized substring of the input string.
virtual uint8_t getCombiningClass(UChar32 c) const U_OVERRIDE
Gets the combining class of c.
virtual UChar32 composePair(UChar32 a, UChar32 b) const U_OVERRIDE
Performs pairwise composition of a & b and returns the composite if there is one.
virtual UBool hasBoundaryAfter(UChar32 c) const U_OVERRIDE
Tests if the character always has a normalization boundary after it, regardless of context.
virtual UNormalizationCheckResult quickCheck(const UnicodeString &s, UErrorCode &errorCode) const U_OVERRIDE
Tests if the string is normalized.
virtual UBool isNormalized(const UnicodeString &s, UErrorCode &errorCode) const U_OVERRIDE
Tests if the string is normalized.
FilteredNormalizer2(const Normalizer2 &n2, const UnicodeSet &filterSet)
Constructs a filtered normalizer wrapping any Normalizer2 instance and a filter set.
Definition: normalizer2.h:511
Unicode normalization functionality for standard Unicode normalization or for using custom mapping ta...
Definition: normalizer2.h:85
virtual UBool getDecomposition(UChar32 c, UnicodeString &decomposition) const =0
Gets the decomposition mapping of c.
virtual int32_t spanQuickCheckYes(const UnicodeString &s, UErrorCode &errorCode) const =0
Returns the end of the normalized substring of the input string.
virtual UBool isNormalized(const UnicodeString &s, UErrorCode &errorCode) const =0
Tests if the string is normalized.
virtual UBool hasBoundaryBefore(UChar32 c) const =0
Tests if the character always has a normalization boundary before it, regardless of context.
static const Normalizer2 * getNFKCInstance(UErrorCode &errorCode)
Returns a Normalizer2 instance for Unicode NFKC normalization.
~Normalizer2()
Destructor.
virtual UChar32 composePair(UChar32 a, UChar32 b) const
Performs pairwise composition of a & b and returns the composite if there is one.
static const Normalizer2 * getNFCInstance(UErrorCode &errorCode)
Returns a Normalizer2 instance for Unicode NFC normalization.
virtual UnicodeString & normalize(const UnicodeString &src, UnicodeString &dest, UErrorCode &errorCode) const =0
Writes the normalized form of the source string to the destination string (replacing its contents) an...
static const Normalizer2 * getInstance(const char *packageName, const char *name, UNormalization2Mode mode, UErrorCode &errorCode)
Returns a Normalizer2 instance which uses the specified data file (packageName/name similar to ucnv_o...
virtual UnicodeString & append(UnicodeString &first, const UnicodeString &second, UErrorCode &errorCode) const =0
Appends the second string to the first string (merging them at the boundary) and returns the first st...
virtual UnicodeString & normalizeSecondAndAppend(UnicodeString &first, const UnicodeString &second, UErrorCode &errorCode) const =0
Appends the normalized form of the second string to the first string (merging them at the boundary) a...
UnicodeString normalize(const UnicodeString &src, UErrorCode &errorCode) const
Returns the normalized form of the source string.
Definition: normalizer2.h:201
virtual UNormalizationCheckResult quickCheck(const UnicodeString &s, UErrorCode &errorCode) const =0
Tests if the string is normalized.
virtual UBool hasBoundaryAfter(UChar32 c) const =0
Tests if the character always has a normalization boundary after it, regardless of context.
virtual uint8_t getCombiningClass(UChar32 c) const
Gets the combining class of c.
virtual UBool isInert(UChar32 c) const =0
Tests if the character is normalization-inert.
static const Normalizer2 * getNFKCCasefoldInstance(UErrorCode &errorCode)
Returns a Normalizer2 instance for Unicode NFKC_Casefold normalization.
virtual void normalizeUTF8(uint32_t options, StringPiece src, ByteSink &sink, Edits *edits, UErrorCode &errorCode) const
Normalizes a UTF-8 string and optionally records how source substrings relate to changed and unchange...
virtual UBool isNormalizedUTF8(StringPiece s, UErrorCode &errorCode) const
Tests if the UTF-8 string is normalized.
virtual UBool getRawDecomposition(UChar32 c, UnicodeString &decomposition) const
Gets the raw decomposition mapping of c.
static const Normalizer2 * getNFDInstance(UErrorCode &errorCode)
Returns a Normalizer2 instance for Unicode NFD normalization.
static const Normalizer2 * getNFKDInstance(UErrorCode &errorCode)
Returns a Normalizer2 instance for Unicode NFKD normalization.
A string-like object that points to a sized piece of memory.
Definition: stringpiece.h:60
UObject is the common ICU "boilerplate" class.
Definition: uobject.h:223
A mutable set of Unicode characters and multicharacter strings.
Definition: uniset.h:285
UnicodeString is a string class that stores Unicode characters directly and provides similar function...
Definition: unistr.h:296
C++ API: StringPiece: Read-only byte string wrapper class.
int32_t UChar32
Define UChar32 as a type for single Unicode code points.
Definition: umachine.h:467
int8_t UBool
The ICU boolean type, a signed-byte integer.
Definition: umachine.h:269
#define U_OVERRIDE
Defined to the C++11 "override" keyword if available.
Definition: umachine.h:130
C++ API: Unicode Set.
C++ API: Unicode String.
C API: New API for Unicode Normalization.
UNormalizationCheckResult
Result values for normalization quick check functions.
Definition: unorm2.h:97
UNormalization2Mode
Constants for normalization modes.
Definition: unorm2.h:48
USetSpanCondition
Argument values for whether span() and similar functions continue while the current character is cont...
Definition: uset.h:159
Basic definitions for ICU, for both C and C++ APIs.
UErrorCode
Standard ICU4C error code type, a substitute for exceptions.
Definition: utypes.h:415
#define U_COMMON_API
Set to export library symbols from inside the common library, and to import them from outside.
Definition: utypes.h:300