1 // © 2016 and later: Unicode, Inc. and others.
2 // License & terms of use: http://www.unicode.org/copyright.html
3 /*
4 *******************************************************************************
5 *   Copyright (C) 2010-2016, International Business Machines
6 *   Corporation and others.  All Rights Reserved.
7 *******************************************************************************
8 *   file name:  ucharstriebuilder.h
9 *   encoding:   UTF-8
10 *   tab size:   8 (not used)
11 *   indentation:4
12 *
13 *   created on: 2010nov14
14 *   created by: Markus W. Scherer
15 */
16 
17 #ifndef __UCHARSTRIEBUILDER_H__
18 #define __UCHARSTRIEBUILDER_H__
19 
20 #include "unicode/utypes.h"
21 
22 #if U_SHOW_CPLUSPLUS_API
23 
24 #include "unicode/stringtriebuilder.h"
25 #include "unicode/ucharstrie.h"
26 #include "unicode/unistr.h"
27 
28 /**
29  * \file
30  * \brief C++ API: Builder for icu::UCharsTrie
31  */
32 
33 U_NAMESPACE_BEGIN
34 
35 class UCharsTrieElement;
36 
37 /**
38  * Builder class for UCharsTrie.
39  *
40  * This class is not intended for public subclassing.
41  * @stable ICU 4.8
42  */
43 class U_COMMON_API UCharsTrieBuilder : public StringTrieBuilder {
44 public:
45     /**
46      * Constructs an empty builder.
47      * @param errorCode Standard ICU error code.
48      * @stable ICU 4.8
49      */
50     UCharsTrieBuilder(UErrorCode &errorCode);
51 
52     /**
53      * Destructor.
54      * @stable ICU 4.8
55      */
56     virtual ~UCharsTrieBuilder();
57 
58     /**
59      * Adds a (string, value) pair.
60      * The string must be unique.
61      * The string contents will be copied; the builder does not keep
62      * a reference to the input UnicodeString or its buffer.
63      * @param s The input string.
64      * @param value The value associated with this string.
65      * @param errorCode Standard ICU error code. Its input value must
66      *                  pass the U_SUCCESS() test, or else the function returns
67      *                  immediately. Check for U_FAILURE() on output or use with
68      *                  function chaining. (See User Guide for details.)
69      * @return *this
70      * @stable ICU 4.8
71      */
72     UCharsTrieBuilder &add(const UnicodeString &s, int32_t value, UErrorCode &errorCode);
73 
74     /**
75      * Builds a UCharsTrie for the add()ed data.
76      * Once built, no further data can be add()ed until clear() is called.
77      *
78      * A UCharsTrie cannot be empty. At least one (string, value) pair
79      * must have been add()ed.
80      *
81      * This method passes ownership of the builder's internal result array to the new trie object.
82      * Another call to any build() variant will re-serialize the trie.
83      * After clear() has been called, a new array will be used as well.
84      * @param buildOption Build option, see UStringTrieBuildOption.
85      * @param errorCode Standard ICU error code. Its input value must
86      *                  pass the U_SUCCESS() test, or else the function returns
87      *                  immediately. Check for U_FAILURE() on output or use with
88      *                  function chaining. (See User Guide for details.)
89      * @return A new UCharsTrie for the add()ed data.
90      * @stable ICU 4.8
91      */
92     UCharsTrie *build(UStringTrieBuildOption buildOption, UErrorCode &errorCode);
93 
94     /**
95      * Builds a UCharsTrie for the add()ed data and char16_t-serializes it.
96      * Once built, no further data can be add()ed until clear() is called.
97      *
98      * A UCharsTrie cannot be empty. At least one (string, value) pair
99      * must have been add()ed.
100      *
101      * Multiple calls to buildUnicodeString() set the UnicodeStrings to the
102      * builder's same char16_t array, without rebuilding.
103      * If buildUnicodeString() is called after build(), the trie will be
104      * re-serialized into a new array.
105      * If build() is called after buildUnicodeString(), the trie object will become
106      * the owner of the previously returned array.
107      * After clear() has been called, a new array will be used as well.
108      * @param buildOption Build option, see UStringTrieBuildOption.
109      * @param result A UnicodeString which will be set to the char16_t-serialized
110      *               UCharsTrie for the add()ed data.
111      * @param errorCode Standard ICU error code. Its input value must
112      *                  pass the U_SUCCESS() test, or else the function returns
113      *                  immediately. Check for U_FAILURE() on output or use with
114      *                  function chaining. (See User Guide for details.)
115      * @return result
116      * @stable ICU 4.8
117      */
118     UnicodeString &buildUnicodeString(UStringTrieBuildOption buildOption, UnicodeString &result,
119                                       UErrorCode &errorCode);
120 
121     /**
122      * Removes all (string, value) pairs.
123      * New data can then be add()ed and a new trie can be built.
124      * @return *this
125      * @stable ICU 4.8
126      */
clear()127     UCharsTrieBuilder &clear() {
128         strings.remove();
129         elementsLength=0;
130         ucharsLength=0;
131         return *this;
132     }
133 
134 private:
135     UCharsTrieBuilder(const UCharsTrieBuilder &other);  // no copy constructor
136     UCharsTrieBuilder &operator=(const UCharsTrieBuilder &other);  // no assignment operator
137 
138     void buildUChars(UStringTrieBuildOption buildOption, UErrorCode &errorCode);
139 
140     virtual int32_t getElementStringLength(int32_t i) const;
141     virtual char16_t getElementUnit(int32_t i, int32_t unitIndex) const;
142     virtual int32_t getElementValue(int32_t i) const;
143 
144     virtual int32_t getLimitOfLinearMatch(int32_t first, int32_t last, int32_t unitIndex) const;
145 
146     virtual int32_t countElementUnits(int32_t start, int32_t limit, int32_t unitIndex) const;
147     virtual int32_t skipElementsBySomeUnits(int32_t i, int32_t unitIndex, int32_t count) const;
148     virtual int32_t indexOfElementWithNextUnit(int32_t i, int32_t unitIndex, char16_t unit) const;
149 
matchNodesCanHaveValues()150     virtual UBool matchNodesCanHaveValues() const { return TRUE; }
151 
getMaxBranchLinearSubNodeLength()152     virtual int32_t getMaxBranchLinearSubNodeLength() const { return UCharsTrie::kMaxBranchLinearSubNodeLength; }
getMinLinearMatch()153     virtual int32_t getMinLinearMatch() const { return UCharsTrie::kMinLinearMatch; }
getMaxLinearMatchLength()154     virtual int32_t getMaxLinearMatchLength() const { return UCharsTrie::kMaxLinearMatchLength; }
155 
156     class UCTLinearMatchNode : public LinearMatchNode {
157     public:
158         UCTLinearMatchNode(const char16_t *units, int32_t len, Node *nextNode);
159         virtual UBool operator==(const Node &other) const;
160         virtual void write(StringTrieBuilder &builder);
161     private:
162         const char16_t *s;
163     };
164 
165     virtual Node *createLinearMatchNode(int32_t i, int32_t unitIndex, int32_t length,
166                                         Node *nextNode) const;
167 
168     UBool ensureCapacity(int32_t length);
169     virtual int32_t write(int32_t unit);
170     int32_t write(const char16_t *s, int32_t length);
171     virtual int32_t writeElementUnits(int32_t i, int32_t unitIndex, int32_t length);
172     virtual int32_t writeValueAndFinal(int32_t i, UBool isFinal);
173     virtual int32_t writeValueAndType(UBool hasValue, int32_t value, int32_t node);
174     virtual int32_t writeDeltaTo(int32_t jumpTarget);
175 
176     UnicodeString strings;
177     UCharsTrieElement *elements;
178     int32_t elementsCapacity;
179     int32_t elementsLength;
180 
181     // char16_t serialization of the trie.
182     // Grows from the back: ucharsLength measures from the end of the buffer!
183     char16_t *uchars;
184     int32_t ucharsCapacity;
185     int32_t ucharsLength;
186 };
187 
188 U_NAMESPACE_END
189 
190 #endif /* U_SHOW_CPLUSPLUS_API */
191 
192 #endif  // __UCHARSTRIEBUILDER_H__
193