ICU 60.2  60.2
rbbi.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 * Copyright (C) 1999-2016 International Business Machines Corporation *
6 * and others. All rights reserved. *
7 ***************************************************************************
8 
9 **********************************************************************
10 * Date Name Description
11 * 10/22/99 alan Creation.
12 * 11/11/99 rgillam Complete port from Java.
13 **********************************************************************
14 */
15 
16 #ifndef RBBI_H
17 #define RBBI_H
18 
19 #include "unicode/utypes.h"
20 
26 #if !UCONFIG_NO_BREAK_ITERATION
27 
28 #include "unicode/brkiter.h"
29 #include "unicode/udata.h"
30 #include "unicode/parseerr.h"
31 #include "unicode/schriter.h"
32 #include "unicode/uchriter.h"
33 
35 
37 class LanguageBreakEngine;
38 struct RBBIDataHeader;
39 class RBBIDataWrapper;
40 class UnhandledEngine;
41 class UStack;
42 
55 
56 private:
61  UText *fText;
62 
68  CharacterIterator *fCharIter;
69 
75  StringCharacterIterator *fSCharIter;
76 
82  UCharCharacterIterator *fDCharIter;
83 
88  RBBIDataWrapper *fData;
89 
101  int32_t fPosition;
102 
106  int32_t fRuleStatusIndex;
107 
111  UBool fDone;
112 
116  public: // TODO: debug, return to private.
117  class BreakCache;
118  BreakCache *fBreakCache;
119  private:
125  uint32_t fDictionaryCharCount;
126 
131  class DictionaryCache;
132  DictionaryCache *fDictionaryCache;
133 
141  UStack *fLanguageBreakEngines;
142 
150  UnhandledEngine *fUnhandledBreakEngine;
151 
157  int32_t fBreakType;
158 
159  //=======================================================================
160  // constructors
161  //=======================================================================
162 
173  RuleBasedBreakIterator(RBBIDataHeader* data, UErrorCode &status);
174 
176  friend class RBBIRuleBuilder;
178  friend class BreakIterator;
179 
180 public:
181 
187 
195 
205  UParseError &parseError,
206  UErrorCode &status);
207 
231  RuleBasedBreakIterator(const uint8_t *compiledRules,
232  uint32_t ruleLength,
233  UErrorCode &status);
234 
248 
253  virtual ~RuleBasedBreakIterator();
254 
263 
272  virtual UBool operator==(const BreakIterator& that) const;
273 
281  UBool operator!=(const BreakIterator& that) const;
282 
293  virtual BreakIterator* clone() const;
294 
300  virtual int32_t hashCode(void) const;
301 
307  virtual const UnicodeString& getRules(void) const;
308 
309  //=======================================================================
310  // BreakIterator overrides
311  //=======================================================================
312 
338  virtual CharacterIterator& getText(void) const;
339 
340 
355  virtual UText *getUText(UText *fillIn, UErrorCode &status) const;
356 
364  virtual void adoptText(CharacterIterator* newText);
365 
377  virtual void setText(const UnicodeString& newText);
378 
392  virtual void setText(UText *text, UErrorCode &status);
393 
399  virtual int32_t first(void);
400 
406  virtual int32_t last(void);
407 
418  virtual int32_t next(int32_t n);
419 
425  virtual int32_t next(void);
426 
432  virtual int32_t previous(void);
433 
441  virtual int32_t following(int32_t offset);
442 
450  virtual int32_t preceding(int32_t offset);
451 
460  virtual UBool isBoundary(int32_t offset);
461 
470  virtual int32_t current(void) const;
471 
472 
506  virtual int32_t getRuleStatus() const;
507 
531  virtual int32_t getRuleStatusVec(int32_t *fillInVec, int32_t capacity, UErrorCode &status);
532 
544  virtual UClassID getDynamicClassID(void) const;
545 
557  static UClassID U_EXPORT2 getStaticClassID(void);
558 
585  virtual BreakIterator * createBufferClone(void *stackBuffer,
586  int32_t &BufferSize,
587  UErrorCode &status);
588 
589 
607  virtual const uint8_t *getBinaryRules(uint32_t &length);
608 
634  virtual RuleBasedBreakIterator &refreshInputText(UText *input, UErrorCode &status);
635 
636 
637 private:
638  //=======================================================================
639  // implementation
640  //=======================================================================
646  void reset(void);
647 
652  void setBreakType(int32_t type);
653 
658  void init(UErrorCode &status);
659 
668  int32_t handlePrevious(int32_t fromPosition);
669 
682  int32_t handleNext();
683 
684 
691  const LanguageBreakEngine *getLanguageBreakEngine(UChar32 c);
692 
693  public:
694 #ifndef U_HIDE_INTERNAL_API
695 
699  void dumpCache();
700 #endif /* U_HIDE_INTERNAL_API */
701 };
702 
703 //------------------------------------------------------------------------------
704 //
705 // Inline Functions Definitions ...
706 //
707 //------------------------------------------------------------------------------
708 
710  return !operator==(that);
711 }
712 
714 
715 #endif /* #if !UCONFIG_NO_BREAK_ITERATION */
716 
717 #endif
C++ API: Break Iterator.
BreakIterator & operator=(const BreakIterator &other)
virtual int32_t next(void)=0
Advance the iterator to the boundary following the current boundary.
virtual UBool isBoundary(int32_t offset)=0
Return true if the specified position is a boundary position.
virtual void adoptText(CharacterIterator *it)=0
Change the text over which this operates.
void * UClassID
UClassID is used to identify classes without using the compiler's RTTI.
Definition: uobject.h:93
virtual CharacterIterator & getText(void) const =0
Return a CharacterIterator over the text being analyzed.
virtual UText * getUText(UText *fillIn, UErrorCode &status) const =0
Get a UText for the text being analyzed.
virtual int32_t first(void)=0
Sets the current iteration position to the beginning of the text, position zero.
virtual int32_t following(int32_t offset)=0
Advance the iterator to the first boundary following the specified offset.
#define U_NAMESPACE_BEGIN
This is used to begin a declaration of a public ICU C++ API.
Definition: uversion.h:131
Abstract class that defines an API for iteration on text objects.
Definition: chariter.h:358
C++ API: String Character Iterator.
A concrete subclass of CharacterIterator that iterates over the characters (code units or code points...
Definition: uchriter.h:35
A concrete subclass of CharacterIterator that iterates over the characters (code units or code points...
Definition: schriter.h:45
The BreakIterator class implements methods for finding the location of boundaries in text...
Definition: brkiter.h:102
virtual int32_t last(void)=0
Set the iterator position to the index immediately BEYOND the last character in the text being scanne...
virtual int32_t current(void) const =0
Return character index of the current iterator position within the text.
int32_t UChar32
Define UChar32 as a type for single Unicode code points.
Definition: umachine.h:396
virtual UBool operator==(const BreakIterator &) const =0
Return true if another object is semantically equal to this one.
virtual int32_t getRuleStatusVec(int32_t *fillInVec, int32_t capacity, UErrorCode &status)
For RuleBasedBreakIterators, get the status (tag) values from the break rule(s) that determined the m...
C API: Data loading interface.
struct UDataMemory UDataMemory
Forward declaration of the data memory type.
Definition: udata.h:158
virtual int32_t previous(void)=0
Set the iterator position to the boundary preceding the current boundary.
virtual int32_t getRuleStatus() const
For RuleBasedBreakIterators, return the status tag from the break rule that determined the most recen...
virtual UClassID getDynamicClassID(void) const =0
Return a polymorphic class ID for this object.
virtual void setText(const UnicodeString &text)=0
Change the text over which this operates.
virtual BreakIterator & refreshInputText(UText *input, UErrorCode &status)=0
Set the subject text string upon which the break iterator is operating without changing any other asp...
virtual BreakIterator * clone(void) const =0
Return a polymorphic copy of this object.
UBool operator!=(const BreakIterator &rhs) const
Returns the complement of the result of operator==.
Definition: brkiter.h:131
C++ API: char16_t Character Iterator.
#define U_NAMESPACE_END
This is used to end a declaration of a public ICU C++ API.
Definition: uversion.h:132
C API: Parse Error Information.
UErrorCode
Error code to replace exception handling, so that the code is compatible with all C++ compilers...
Definition: utypes.h:396
virtual BreakIterator * createBufferClone(void *stackBuffer, int32_t &BufferSize, UErrorCode &status)=0
Deprecated functionality.
UText struct.
Definition: utext.h:1345
A subclass of BreakIterator whose behavior is specified using a list of rules.
Definition: rbbi.h:54
virtual int32_t preceding(int32_t offset)=0
Set the iterator position to the first boundary preceding the specified offset.
A UParseError struct is used to returned detailed information about parsing errors.
Definition: parseerr.h:58
Basic definitions for ICU, for both C and C++ APIs.
#define U_COMMON_API
Set to export library symbols from inside the common library, and to import them from outside...
Definition: utypes.h:359
UnicodeString is a string class that stores Unicode characters directly and provides similar function...
Definition: unistr.h:286
UBool operator!=(const BreakIterator &that) const
Not-equal operator.
Definition: rbbi.h:709
int8_t UBool
The ICU boolean type.
Definition: umachine.h:236