1.1 --- /dev/null Thu Jan 01 00:00:00 1970 +0000
1.2 +++ b/intl/icu/source/i18n/rbt_set.h Wed Dec 31 06:09:35 2014 +0100
1.3 @@ -0,0 +1,165 @@
1.4 +/*
1.5 +**********************************************************************
1.6 +* Copyright (C) 1999-2007, International Business Machines Corporation
1.7 +* and others. All Rights Reserved.
1.8 +**********************************************************************
1.9 +* Date Name Description
1.10 +* 11/17/99 aliu Creation.
1.11 +**********************************************************************
1.12 +*/
1.13 +#ifndef RBT_SET_H
1.14 +#define RBT_SET_H
1.15 +
1.16 +#include "unicode/utypes.h"
1.17 +
1.18 +#if !UCONFIG_NO_TRANSLITERATION
1.19 +
1.20 +#include "unicode/uobject.h"
1.21 +#include "unicode/utrans.h"
1.22 +#include "uvector.h"
1.23 +
1.24 +U_NAMESPACE_BEGIN
1.25 +
1.26 +class Replaceable;
1.27 +class TransliterationRule;
1.28 +class TransliterationRuleData;
1.29 +class UnicodeFilter;
1.30 +class UnicodeString;
1.31 +class UnicodeSet;
1.32 +
1.33 +/**
1.34 + * A set of rules for a <code>RuleBasedTransliterator</code>.
1.35 + * @author Alan Liu
1.36 + */
1.37 +class TransliterationRuleSet : public UMemory {
1.38 + /**
1.39 + * Vector of rules, in the order added. This is used while the
1.40 + * rule set is getting built. After that, freeze() reorders and
1.41 + * indexes the rules into rules[]. Any given rule is stored once
1.42 + * in ruleVector, and one or more times in rules[]. ruleVector
1.43 + * owns and deletes the rules.
1.44 + */
1.45 + UVector* ruleVector;
1.46 +
1.47 + /**
1.48 + * Sorted and indexed table of rules. This is created by freeze()
1.49 + * from the rules in ruleVector. It contains alias pointers to
1.50 + * the rules in ruleVector. It is zero before freeze() is called
1.51 + * and non-zero thereafter.
1.52 + */
1.53 + TransliterationRule** rules;
1.54 +
1.55 + /**
1.56 + * Index table. For text having a first character c, compute x = c&0xFF.
1.57 + * Now use rules[index[x]..index[x+1]-1]. This index table is created by
1.58 + * freeze(). Before freeze() is called it contains garbage.
1.59 + */
1.60 + int32_t index[257];
1.61 +
1.62 + /**
1.63 + * Length of the longest preceding context
1.64 + */
1.65 + int32_t maxContextLength;
1.66 +
1.67 +public:
1.68 +
1.69 + /**
1.70 + * Construct a new empty rule set.
1.71 + * @param status Output parameter filled in with success or failure status.
1.72 + */
1.73 + TransliterationRuleSet(UErrorCode& status);
1.74 +
1.75 + /**
1.76 + * Copy constructor.
1.77 + */
1.78 + TransliterationRuleSet(const TransliterationRuleSet&);
1.79 +
1.80 + /**
1.81 + * Destructor.
1.82 + */
1.83 + virtual ~TransliterationRuleSet();
1.84 +
1.85 + /**
1.86 + * Change the data object that this rule belongs to. Used
1.87 + * internally by the TransliterationRuleData copy constructor.
1.88 + * @param data the new data value to be set.
1.89 + */
1.90 + void setData(const TransliterationRuleData* data);
1.91 +
1.92 + /**
1.93 + * Return the maximum context length.
1.94 + * @return the length of the longest preceding context.
1.95 + */
1.96 + virtual int32_t getMaximumContextLength(void) const;
1.97 +
1.98 + /**
1.99 + * Add a rule to this set. Rules are added in order, and order is
1.100 + * significant. The last call to this method must be followed by
1.101 + * a call to <code>freeze()</code> before the rule set is used.
1.102 + * This method must <em>not</em> be called after freeze() has been
1.103 + * called.
1.104 + *
1.105 + * @param adoptedRule the rule to add
1.106 + */
1.107 + virtual void addRule(TransliterationRule* adoptedRule,
1.108 + UErrorCode& status);
1.109 +
1.110 + /**
1.111 + * Check this for masked rules and index it to optimize performance.
1.112 + * The sequence of operations is: (1) add rules to a set using
1.113 + * <code>addRule()</code>; (2) freeze the set using
1.114 + * <code>freeze()</code>; (3) use the rule set. If
1.115 + * <code>addRule()</code> is called after calling this method, it
1.116 + * invalidates this object, and this method must be called again.
1.117 + * That is, <code>freeze()</code> may be called multiple times,
1.118 + * although for optimal performance it shouldn't be.
1.119 + * @param parseError A pointer to UParseError to receive information about errors
1.120 + * occurred.
1.121 + * @param status Output parameter filled in with success or failure status.
1.122 + */
1.123 + virtual void freeze(UParseError& parseError, UErrorCode& status);
1.124 +
1.125 + /**
1.126 + * Transliterate the given text with the given UTransPosition
1.127 + * indices. Return TRUE if the transliteration should continue
1.128 + * or FALSE if it should halt (because of a U_PARTIAL_MATCH match).
1.129 + * Note that FALSE is only ever returned if isIncremental is TRUE.
1.130 + * @param text the text to be transliterated
1.131 + * @param index the position indices, which will be updated
1.132 + * @param isIncremental if TRUE, assume new text may be inserted
1.133 + * at index.limit, and return FALSE if thre is a partial match.
1.134 + * @return TRUE unless a U_PARTIAL_MATCH has been obtained,
1.135 + * indicating that transliteration should stop until more text
1.136 + * arrives.
1.137 + */
1.138 + UBool transliterate(Replaceable& text,
1.139 + UTransPosition& index,
1.140 + UBool isIncremental);
1.141 +
1.142 + /**
1.143 + * Create rule strings that represents this rule set.
1.144 + * @param result string to receive the rule strings. Current
1.145 + * contents will be deleted.
1.146 + * @param escapeUnprintable True, will escape the unprintable characters
1.147 + * @return A reference to 'result'.
1.148 + */
1.149 + virtual UnicodeString& toRules(UnicodeString& result,
1.150 + UBool escapeUnprintable) const;
1.151 +
1.152 + /**
1.153 + * Return the set of all characters that may be modified
1.154 + * (getTarget=false) or emitted (getTarget=true) by this set.
1.155 + */
1.156 + UnicodeSet& getSourceTargetSet(UnicodeSet& result,
1.157 + UBool getTarget) const;
1.158 +
1.159 +private:
1.160 +
1.161 + TransliterationRuleSet &operator=(const TransliterationRuleSet &other); // forbid copying of this class
1.162 +};
1.163 +
1.164 +U_NAMESPACE_END
1.165 +
1.166 +#endif /* #if !UCONFIG_NO_TRANSLITERATION */
1.167 +
1.168 +#endif
