1 // Copyright (C) 2016 and later: Unicode, Inc. and others.
2 // License & terms of use: http://www.unicode.org/copyright.html
4 ********************************************************************************
5 * Copyright (C) 1997-2015, International Business Machines
6 * Corporation and others. All Rights Reserved.
7 ********************************************************************************
13 #include "unicode/utypes.h"
14 #include "unicode/brkiter.h"
16 #if !UCONFIG_NO_BREAK_ITERATION && !UCONFIG_NO_FILTERED_BREAK_ITERATION
22 * \brief C++ API: FilteredBreakIteratorBuilder
26 * The BreakIteratorFilter is used to modify the behavior of a BreakIterator
27 * by constructing a new BreakIterator which suppresses certain segment boundaries.
28 * See http://www.unicode.org/reports/tr35/tr35-general.html#Segmentation_Exceptions .
29 * For example, a typical English Sentence Break Iterator would break on the space
30 * in the string "Mr. Smith" (resulting in two segments),
31 * but with "Mr." as an exception, a filtered break iterator
32 * would consider the string "Mr. Smith" to be a single segment.
36 class U_COMMON_API FilteredBreakIteratorBuilder : public UObject {
42 virtual ~FilteredBreakIteratorBuilder();
45 * Construct a FilteredBreakIteratorBuilder based on rules in a locale.
46 * The rules are taken from CLDR exception data for the locale,
47 * see http://www.unicode.org/reports/tr35/tr35-general.html#Segmentation_Exceptions
48 * This is the equivalent of calling createInstance(UErrorCode&)
49 * and then repeatedly calling addNoBreakAfter(...) with the contents
50 * of the CLDR exception data.
51 * @param where the locale.
52 * @param status The error code.
53 * @return the new builder
56 static FilteredBreakIteratorBuilder *createInstance(const Locale& where, UErrorCode& status);
59 * Construct an empty FilteredBreakIteratorBuilder.
60 * In this state, it will not suppress any segment boundaries.
61 * @param status The error code.
62 * @return the new builder
65 static FilteredBreakIteratorBuilder *createInstance(UErrorCode &status);
68 * Suppress a certain string from being the end of a segment.
69 * For example, suppressing "Mr.", then segments ending in "Mr." will not be returned
71 * @param string the string to suppress, such as "Mr."
72 * @param status error code
73 * @return returns TRUE if the string was not present and now added,
74 * FALSE if the call was a no-op because the string was already being suppressed.
77 virtual UBool suppressBreakAfter(const UnicodeString& string, UErrorCode& status) = 0;
80 * Stop suppressing a certain string from being the end of the segment.
81 * This function does not create any new segment boundaries, but only serves to un-do
82 * the effect of earlier calls to suppressBreakAfter, or to un-do the effect of
83 * locale data which may be suppressing certain strings.
84 * @param exception the exception to remove
85 * @param status error code
86 * @return returns TRUE if the string was present and now removed,
87 * FALSE if the call was a no-op because the string was not being suppressed.
90 virtual UBool unsuppressBreakAfter(const UnicodeString& string, UErrorCode& status) = 0;
93 * Wrap (adopt) an existing break iterator in a new filtered instance.
94 * The resulting BreakIterator is owned by the caller.
95 * The BreakIteratorFilter may be destroyed before the BreakIterator is destroyed.
96 * Note that the adoptBreakIterator is adopted by the new BreakIterator
97 * and should no longer be used by the caller.
98 * The FilteredBreakIteratorBuilder may be reused.
99 * @param adoptBreakIterator the break iterator to adopt
100 * @param status error code
101 * @return the new BreakIterator, owned by the caller.
104 virtual BreakIterator *build(BreakIterator* adoptBreakIterator, UErrorCode& status) = 0;
111 FilteredBreakIteratorBuilder();
117 #endif // #if !UCONFIG_NO_BREAK_ITERATION && !UCONFIG_NO_FILTERED_BREAK_ITERATION
119 #endif // #ifndef FILTEREDBRK_H