1.1 --- /dev/null Thu Jan 01 00:00:00 1970 +0000
1.2 +++ b/intl/icu/source/common/brkeng.h Wed Dec 31 06:09:35 2014 +0100
1.3 @@ -0,0 +1,289 @@
1.4 +/**
1.5 + ************************************************************************************
1.6 + * Copyright (C) 2006-2012, International Business Machines Corporation and others. *
1.7 + * All Rights Reserved. *
1.8 + ************************************************************************************
1.9 + */
1.10 +
1.11 +#ifndef BRKENG_H
1.12 +#define BRKENG_H
1.13 +
1.14 +#include "unicode/utypes.h"
1.15 +#include "unicode/uobject.h"
1.16 +#include "unicode/utext.h"
1.17 +#include "unicode/uscript.h"
1.18 +
1.19 +U_NAMESPACE_BEGIN
1.20 +
1.21 +class UnicodeSet;
1.22 +class UStack;
1.23 +class DictionaryMatcher;
1.24 +
1.25 +/*******************************************************************
1.26 + * LanguageBreakEngine
1.27 + */
1.28 +
1.29 +/**
1.30 + * <p>LanguageBreakEngines implement language-specific knowledge for
1.31 + * finding text boundaries within a run of characters belonging to a
1.32 + * specific set. The boundaries will be of a specific kind, e.g. word,
1.33 + * line, etc.</p>
1.34 + *
1.35 + * <p>LanguageBreakEngines should normally be implemented so as to
1.36 + * be shared between threads without locking.</p>
1.37 + */
1.38 +class LanguageBreakEngine : public UMemory {
1.39 + public:
1.40 +
1.41 + /**
1.42 + * <p>Default constructor.</p>
1.43 + *
1.44 + */
1.45 + LanguageBreakEngine();
1.46 +
1.47 + /**
1.48 + * <p>Virtual destructor.</p>
1.49 + */
1.50 + virtual ~LanguageBreakEngine();
1.51 +
1.52 + /**
1.53 + * <p>Indicate whether this engine handles a particular character for
1.54 + * a particular kind of break.</p>
1.55 + *
1.56 + * @param c A character which begins a run that the engine might handle
1.57 + * @param breakType The type of text break which the caller wants to determine
1.58 + * @return TRUE if this engine handles the particular character and break
1.59 + * type.
1.60 + */
1.61 + virtual UBool handles(UChar32 c, int32_t breakType) const = 0;
1.62 +
1.63 + /**
1.64 + * <p>Find any breaks within a run in the supplied text.</p>
1.65 + *
1.66 + * @param text A UText representing the text. The
1.67 + * iterator is left at the end of the run of characters which the engine
1.68 + * is capable of handling.
1.69 + * @param startPos The start of the run within the supplied text.
1.70 + * @param endPos The end of the run within the supplied text.
1.71 + * @param reverse Whether the caller is looking for breaks in a reverse
1.72 + * direction.
1.73 + * @param breakType The type of break desired, or -1.
1.74 + * @param foundBreaks An allocated C array of the breaks found, if any
1.75 + * @return The number of breaks found.
1.76 + */
1.77 + virtual int32_t findBreaks( UText *text,
1.78 + int32_t startPos,
1.79 + int32_t endPos,
1.80 + UBool reverse,
1.81 + int32_t breakType,
1.82 + UStack &foundBreaks ) const = 0;
1.83 +
1.84 +};
1.85 +
1.86 +/*******************************************************************
1.87 + * LanguageBreakFactory
1.88 + */
1.89 +
1.90 +/**
1.91 + * <p>LanguageBreakFactorys find and return a LanguageBreakEngine
1.92 + * that can determine breaks for characters in a specific set, if
1.93 + * such an object can be found.</p>
1.94 + *
1.95 + * <p>If a LanguageBreakFactory is to be shared between threads,
1.96 + * appropriate synchronization must be used; there is none internal
1.97 + * to the factory.</p>
1.98 + *
1.99 + * <p>A LanguageBreakEngine returned by a LanguageBreakFactory can
1.100 + * normally be shared between threads without synchronization, unless
1.101 + * the specific subclass of LanguageBreakFactory indicates otherwise.</p>
1.102 + *
1.103 + * <p>A LanguageBreakFactory is responsible for deleting any LanguageBreakEngine
1.104 + * it returns when it itself is deleted, unless the specific subclass of
1.105 + * LanguageBreakFactory indicates otherwise. Naturally, the factory should
1.106 + * not be deleted until the LanguageBreakEngines it has returned are no
1.107 + * longer needed.</p>
1.108 + */
1.109 +class LanguageBreakFactory : public UMemory {
1.110 + public:
1.111 +
1.112 + /**
1.113 + * <p>Default constructor.</p>
1.114 + *
1.115 + */
1.116 + LanguageBreakFactory();
1.117 +
1.118 + /**
1.119 + * <p>Virtual destructor.</p>
1.120 + */
1.121 + virtual ~LanguageBreakFactory();
1.122 +
1.123 + /**
1.124 + * <p>Find and return a LanguageBreakEngine that can find the desired
1.125 + * kind of break for the set of characters to which the supplied
1.126 + * character belongs. It is up to the set of available engines to
1.127 + * determine what the sets of characters are.</p>
1.128 + *
1.129 + * @param c A character that begins a run for which a LanguageBreakEngine is
1.130 + * sought.
1.131 + * @param breakType The kind of text break for which a LanguageBreakEngine is
1.132 + * sought.
1.133 + * @return A LanguageBreakEngine with the desired characteristics, or 0.
1.134 + */
1.135 + virtual const LanguageBreakEngine *getEngineFor(UChar32 c, int32_t breakType) = 0;
1.136 +
1.137 +};
1.138 +
1.139 +/*******************************************************************
1.140 + * UnhandledEngine
1.141 + */
1.142 +
1.143 +/**
1.144 + * <p>UnhandledEngine is a special subclass of LanguageBreakEngine that
1.145 + * handles characters that no other LanguageBreakEngine is available to
1.146 + * handle. It is told the character and the type of break; at its
1.147 + * discretion it may handle more than the specified character (e.g.,
1.148 + * the entire script to which that character belongs.</p>
1.149 + *
1.150 + * <p>UnhandledEngines may not be shared between threads without
1.151 + * external synchronization.</p>
1.152 + */
1.153 +
1.154 +class UnhandledEngine : public LanguageBreakEngine {
1.155 + private:
1.156 +
1.157 + /**
1.158 + * The sets of characters handled, for each break type
1.159 + * @internal
1.160 + */
1.161 +
1.162 + UnicodeSet *fHandled[4];
1.163 +
1.164 + public:
1.165 +
1.166 + /**
1.167 + * <p>Default constructor.</p>
1.168 + *
1.169 + */
1.170 + UnhandledEngine(UErrorCode &status);
1.171 +
1.172 + /**
1.173 + * <p>Virtual destructor.</p>
1.174 + */
1.175 + virtual ~UnhandledEngine();
1.176 +
1.177 + /**
1.178 + * <p>Indicate whether this engine handles a particular character for
1.179 + * a particular kind of break.</p>
1.180 + *
1.181 + * @param c A character which begins a run that the engine might handle
1.182 + * @param breakType The type of text break which the caller wants to determine
1.183 + * @return TRUE if this engine handles the particular character and break
1.184 + * type.
1.185 + */
1.186 + virtual UBool handles(UChar32 c, int32_t breakType) const;
1.187 +
1.188 + /**
1.189 + * <p>Find any breaks within a run in the supplied text.</p>
1.190 + *
1.191 + * @param text A UText representing the text (TODO: UText). The
1.192 + * iterator is left at the end of the run of characters which the engine
1.193 + * is capable of handling.
1.194 + * @param startPos The start of the run within the supplied text.
1.195 + * @param endPos The end of the run within the supplied text.
1.196 + * @param reverse Whether the caller is looking for breaks in a reverse
1.197 + * direction.
1.198 + * @param breakType The type of break desired, or -1.
1.199 + * @param foundBreaks An allocated C array of the breaks found, if any
1.200 + * @return The number of breaks found.
1.201 + */
1.202 + virtual int32_t findBreaks( UText *text,
1.203 + int32_t startPos,
1.204 + int32_t endPos,
1.205 + UBool reverse,
1.206 + int32_t breakType,
1.207 + UStack &foundBreaks ) const;
1.208 +
1.209 + /**
1.210 + * <p>Tell the engine to handle a particular character and break type.</p>
1.211 + *
1.212 + * @param c A character which the engine should handle
1.213 + * @param breakType The type of text break for which the engine should handle c
1.214 + */
1.215 + virtual void handleCharacter(UChar32 c, int32_t breakType);
1.216 +
1.217 +};
1.218 +
1.219 +/*******************************************************************
1.220 + * ICULanguageBreakFactory
1.221 + */
1.222 +
1.223 +/**
1.224 + * <p>ICULanguageBreakFactory is the default LanguageBreakFactory for
1.225 + * ICU. It creates dictionary-based LanguageBreakEngines from dictionary
1.226 + * data in the ICU data file.</p>
1.227 + */
1.228 +class ICULanguageBreakFactory : public LanguageBreakFactory {
1.229 + private:
1.230 +
1.231 + /**
1.232 + * The stack of break engines created by this factory
1.233 + * @internal
1.234 + */
1.235 +
1.236 + UStack *fEngines;
1.237 +
1.238 + public:
1.239 +
1.240 + /**
1.241 + * <p>Standard constructor.</p>
1.242 + *
1.243 + */
1.244 + ICULanguageBreakFactory(UErrorCode &status);
1.245 +
1.246 + /**
1.247 + * <p>Virtual destructor.</p>
1.248 + */
1.249 + virtual ~ICULanguageBreakFactory();
1.250 +
1.251 + /**
1.252 + * <p>Find and return a LanguageBreakEngine that can find the desired
1.253 + * kind of break for the set of characters to which the supplied
1.254 + * character belongs. It is up to the set of available engines to
1.255 + * determine what the sets of characters are.</p>
1.256 + *
1.257 + * @param c A character that begins a run for which a LanguageBreakEngine is
1.258 + * sought.
1.259 + * @param breakType The kind of text break for which a LanguageBreakEngine is
1.260 + * sought.
1.261 + * @return A LanguageBreakEngine with the desired characteristics, or 0.
1.262 + */
1.263 + virtual const LanguageBreakEngine *getEngineFor(UChar32 c, int32_t breakType);
1.264 +
1.265 +protected:
1.266 + /**
1.267 + * <p>Create a LanguageBreakEngine for the set of characters to which
1.268 + * the supplied character belongs, for the specified break type.</p>
1.269 + *
1.270 + * @param c A character that begins a run for which a LanguageBreakEngine is
1.271 + * sought.
1.272 + * @param breakType The kind of text break for which a LanguageBreakEngine is
1.273 + * sought.
1.274 + * @return A LanguageBreakEngine with the desired characteristics, or 0.
1.275 + */
1.276 + virtual const LanguageBreakEngine *loadEngineFor(UChar32 c, int32_t breakType);
1.277 +
1.278 + /**
1.279 + * <p>Create a DictionaryMatcher for the specified script and break type.</p>
1.280 + * @param script An ISO 15924 script code that identifies the dictionary to be
1.281 + * created.
1.282 + * @param breakType The kind of text break for which a dictionary is
1.283 + * sought.
1.284 + * @return A DictionaryMatcher with the desired characteristics, or NULL.
1.285 + */
1.286 + virtual DictionaryMatcher *loadDictionaryMatcherFor(UScriptCode script, int32_t breakType);
1.287 +};
1.288 +
1.289 +U_NAMESPACE_END
1.290 +
1.291 + /* BRKENG_H */
1.292 +#endif
