michael@0: /**
michael@0:  ************************************************************************************
michael@0:  * Copyright (C) 2006-2012, International Business Machines Corporation and others. *
michael@0:  * All Rights Reserved.                                                             *
michael@0:  ************************************************************************************
michael@0:  */
michael@0: 
michael@0: #ifndef BRKENG_H
michael@0: #define BRKENG_H
michael@0: 
michael@0: #include "unicode/utypes.h"
michael@0: #include "unicode/uobject.h"
michael@0: #include "unicode/utext.h"
michael@0: #include "unicode/uscript.h"
michael@0: 
michael@0: U_NAMESPACE_BEGIN
michael@0: 
michael@0: class UnicodeSet;
michael@0: class UStack;
michael@0: class DictionaryMatcher;
michael@0: 
michael@0: /*******************************************************************
michael@0:  * LanguageBreakEngine
michael@0:  */
michael@0: 
michael@0: /**
michael@0:  * <p>LanguageBreakEngines implement language-specific knowledge for
michael@0:  * finding text boundaries within a run of characters belonging to a
michael@0:  * specific set. The boundaries will be of a specific kind, e.g. word,
michael@0:  * line, etc.</p>
michael@0:  *
michael@0:  * <p>LanguageBreakEngines should normally be implemented so as to
michael@0:  * be shared between threads without locking.</p>
michael@0:  */
michael@0: class LanguageBreakEngine : public UMemory {
michael@0:  public:
michael@0: 
michael@0:   /**
michael@0:    * <p>Default constructor.</p>
michael@0:    *
michael@0:    */
michael@0:   LanguageBreakEngine();
michael@0: 
michael@0:   /**
michael@0:    * <p>Virtual destructor.</p>
michael@0:    */
michael@0:   virtual ~LanguageBreakEngine();
michael@0: 
michael@0:  /**
michael@0:   * <p>Indicate whether this engine handles a particular character for
michael@0:   * a particular kind of break.</p>
michael@0:   *
michael@0:   * @param c A character which begins a run that the engine might handle
michael@0:   * @param breakType The type of text break which the caller wants to determine
michael@0:   * @return TRUE if this engine handles the particular character and break
michael@0:   * type.
michael@0:   */
michael@0:   virtual UBool handles(UChar32 c, int32_t breakType) const = 0;
michael@0: 
michael@0:  /**
michael@0:   * <p>Find any breaks within a run in the supplied text.</p>
michael@0:   *
michael@0:   * @param text A UText representing the text. The
michael@0:   * iterator is left at the end of the run of characters which the engine
michael@0:   * is capable of handling.
michael@0:   * @param startPos The start of the run within the supplied text.
michael@0:   * @param endPos The end of the run within the supplied text.
michael@0:   * @param reverse Whether the caller is looking for breaks in a reverse
michael@0:   * direction.
michael@0:   * @param breakType The type of break desired, or -1.
michael@0:   * @param foundBreaks An allocated C array of the breaks found, if any
michael@0:   * @return The number of breaks found.
michael@0:   */
michael@0:   virtual int32_t findBreaks( UText *text,
michael@0:                               int32_t startPos,
michael@0:                               int32_t endPos,
michael@0:                               UBool reverse,
michael@0:                               int32_t breakType,
michael@0:                               UStack &foundBreaks ) const = 0;
michael@0: 
michael@0: };
michael@0: 
michael@0: /*******************************************************************
michael@0:  * LanguageBreakFactory
michael@0:  */
michael@0: 
michael@0: /**
michael@0:  * <p>LanguageBreakFactorys find and return a LanguageBreakEngine
michael@0:  * that can determine breaks for characters in a specific set, if
michael@0:  * such an object can be found.</p>
michael@0:  *
michael@0:  * <p>If a LanguageBreakFactory is to be shared between threads,
michael@0:  * appropriate synchronization must be used; there is none internal
michael@0:  * to the factory.</p>
michael@0:  *
michael@0:  * <p>A LanguageBreakEngine returned by a LanguageBreakFactory can
michael@0:  * normally be shared between threads without synchronization, unless
michael@0:  * the specific subclass of LanguageBreakFactory indicates otherwise.</p>
michael@0:  *
michael@0:  * <p>A LanguageBreakFactory is responsible for deleting any LanguageBreakEngine
michael@0:  * it returns when it itself is deleted, unless the specific subclass of
michael@0:  * LanguageBreakFactory indicates otherwise. Naturally, the factory should
michael@0:  * not be deleted until the LanguageBreakEngines it has returned are no
michael@0:  * longer needed.</p>
michael@0:  */
michael@0: class LanguageBreakFactory : public UMemory {
michael@0:  public:
michael@0: 
michael@0:   /**
michael@0:    * <p>Default constructor.</p>
michael@0:    *
michael@0:    */
michael@0:   LanguageBreakFactory();
michael@0: 
michael@0:   /**
michael@0:    * <p>Virtual destructor.</p>
michael@0:    */
michael@0:   virtual ~LanguageBreakFactory();
michael@0: 
michael@0:  /**
michael@0:   * <p>Find and return a LanguageBreakEngine that can find the desired
michael@0:   * kind of break for the set of characters to which the supplied
michael@0:   * character belongs. It is up to the set of available engines to
michael@0:   * determine what the sets of characters are.</p>
michael@0:   *
michael@0:   * @param c A character that begins a run for which a LanguageBreakEngine is
michael@0:   * sought.
michael@0:   * @param breakType The kind of text break for which a LanguageBreakEngine is
michael@0:   * sought.
michael@0:   * @return A LanguageBreakEngine with the desired characteristics, or 0.
michael@0:   */
michael@0:   virtual const LanguageBreakEngine *getEngineFor(UChar32 c, int32_t breakType) = 0;
michael@0: 
michael@0: };
michael@0: 
michael@0: /*******************************************************************
michael@0:  * UnhandledEngine
michael@0:  */
michael@0: 
michael@0: /**
michael@0:  * <p>UnhandledEngine is a special subclass of LanguageBreakEngine that
michael@0:  * handles characters that no other LanguageBreakEngine is available to
michael@0:  * handle. It is told the character and the type of break; at its
michael@0:  * discretion it may handle more than the specified character (e.g.,
michael@0:  * the entire script to which that character belongs.</p>
michael@0:  *
michael@0:  * <p>UnhandledEngines may not be shared between threads without
michael@0:  * external synchronization.</p>
michael@0:  */
michael@0: 
michael@0: class UnhandledEngine : public LanguageBreakEngine {
michael@0:  private:
michael@0: 
michael@0:     /**
michael@0:      * The sets of characters handled, for each break type
michael@0:      * @internal
michael@0:      */
michael@0: 
michael@0:   UnicodeSet    *fHandled[4];
michael@0: 
michael@0:  public:
michael@0: 
michael@0:   /**
michael@0:    * <p>Default constructor.</p>
michael@0:    *
michael@0:    */
michael@0:   UnhandledEngine(UErrorCode &status);
michael@0: 
michael@0:   /**
michael@0:    * <p>Virtual destructor.</p>
michael@0:    */
michael@0:   virtual ~UnhandledEngine();
michael@0: 
michael@0:  /**
michael@0:   * <p>Indicate whether this engine handles a particular character for
michael@0:   * a particular kind of break.</p>
michael@0:   *
michael@0:   * @param c A character which begins a run that the engine might handle
michael@0:   * @param breakType The type of text break which the caller wants to determine
michael@0:   * @return TRUE if this engine handles the particular character and break
michael@0:   * type.
michael@0:   */
michael@0:   virtual UBool handles(UChar32 c, int32_t breakType) const;
michael@0: 
michael@0:  /**
michael@0:   * <p>Find any breaks within a run in the supplied text.</p>
michael@0:   *
michael@0:   * @param text A UText representing the text (TODO: UText). The
michael@0:   * iterator is left at the end of the run of characters which the engine
michael@0:   * is capable of handling.
michael@0:   * @param startPos The start of the run within the supplied text.
michael@0:   * @param endPos The end of the run within the supplied text.
michael@0:   * @param reverse Whether the caller is looking for breaks in a reverse
michael@0:   * direction.
michael@0:   * @param breakType The type of break desired, or -1.
michael@0:   * @param foundBreaks An allocated C array of the breaks found, if any
michael@0:   * @return The number of breaks found.
michael@0:   */
michael@0:   virtual int32_t findBreaks( UText *text,
michael@0:                               int32_t startPos,
michael@0:                               int32_t endPos,
michael@0:                               UBool reverse,
michael@0:                               int32_t breakType,
michael@0:                               UStack &foundBreaks ) const;
michael@0: 
michael@0:  /**
michael@0:   * <p>Tell the engine to handle a particular character and break type.</p>
michael@0:   *
michael@0:   * @param c A character which the engine should handle
michael@0:   * @param breakType The type of text break for which the engine should handle c
michael@0:   */
michael@0:   virtual void handleCharacter(UChar32 c, int32_t breakType);
michael@0: 
michael@0: };
michael@0: 
michael@0: /*******************************************************************
michael@0:  * ICULanguageBreakFactory
michael@0:  */
michael@0: 
michael@0: /**
michael@0:  * <p>ICULanguageBreakFactory is the default LanguageBreakFactory for
michael@0:  * ICU. It creates dictionary-based LanguageBreakEngines from dictionary
michael@0:  * data in the ICU data file.</p>
michael@0:  */
michael@0: class ICULanguageBreakFactory : public LanguageBreakFactory {
michael@0:  private:
michael@0: 
michael@0:     /**
michael@0:      * The stack of break engines created by this factory
michael@0:      * @internal
michael@0:      */
michael@0: 
michael@0:   UStack    *fEngines;
michael@0: 
michael@0:  public:
michael@0: 
michael@0:   /**
michael@0:    * <p>Standard constructor.</p>
michael@0:    *
michael@0:    */
michael@0:   ICULanguageBreakFactory(UErrorCode &status);
michael@0: 
michael@0:   /**
michael@0:    * <p>Virtual destructor.</p>
michael@0:    */
michael@0:   virtual ~ICULanguageBreakFactory();
michael@0: 
michael@0:  /**
michael@0:   * <p>Find and return a LanguageBreakEngine that can find the desired
michael@0:   * kind of break for the set of characters to which the supplied
michael@0:   * character belongs. It is up to the set of available engines to
michael@0:   * determine what the sets of characters are.</p>
michael@0:   *
michael@0:   * @param c A character that begins a run for which a LanguageBreakEngine is
michael@0:   * sought.
michael@0:   * @param breakType The kind of text break for which a LanguageBreakEngine is
michael@0:   * sought.
michael@0:   * @return A LanguageBreakEngine with the desired characteristics, or 0.
michael@0:   */
michael@0:   virtual const LanguageBreakEngine *getEngineFor(UChar32 c, int32_t breakType);
michael@0: 
michael@0: protected:
michael@0:  /**
michael@0:   * <p>Create a LanguageBreakEngine for the set of characters to which
michael@0:   * the supplied character belongs, for the specified break type.</p>
michael@0:   *
michael@0:   * @param c A character that begins a run for which a LanguageBreakEngine is
michael@0:   * sought.
michael@0:   * @param breakType The kind of text break for which a LanguageBreakEngine is
michael@0:   * sought.
michael@0:   * @return A LanguageBreakEngine with the desired characteristics, or 0.
michael@0:   */
michael@0:   virtual const LanguageBreakEngine *loadEngineFor(UChar32 c, int32_t breakType);
michael@0: 
michael@0:   /**
michael@0:    * <p>Create a DictionaryMatcher for the specified script and break type.</p>
michael@0:    * @param script An ISO 15924 script code that identifies the dictionary to be
michael@0:    * created.
michael@0:    * @param breakType The kind of text break for which a dictionary is 
michael@0:    * sought.
michael@0:    * @return A DictionaryMatcher with the desired characteristics, or NULL.
michael@0:    */
michael@0:   virtual DictionaryMatcher *loadDictionaryMatcherFor(UScriptCode script, int32_t breakType);
michael@0: };
michael@0: 
michael@0: U_NAMESPACE_END
michael@0: 
michael@0:     /* BRKENG_H */
michael@0: #endif