michael@0: /*
michael@0: **********************************************************************
michael@0: *   Copyright (C) 2001-2008,2010 IBM and others. All rights reserved.
michael@0: **********************************************************************
michael@0: *   Date        Name        Description
michael@0: *  03/22/2000   helena      Creation.
michael@0: **********************************************************************
michael@0: */
michael@0: 
michael@0: #include "unicode/utypes.h"
michael@0: 
michael@0: #if !UCONFIG_NO_COLLATION && !UCONFIG_NO_BREAK_ITERATION
michael@0: 
michael@0: #include "unicode/brkiter.h"
michael@0: #include "unicode/schriter.h"
michael@0: #include "unicode/search.h"
michael@0: #include "usrchimp.h"
michael@0: #include "cmemory.h"
michael@0: 
michael@0: // public constructors and destructors -----------------------------------
michael@0: U_NAMESPACE_BEGIN
michael@0: 
michael@0: SearchIterator::SearchIterator(const SearchIterator &other)
michael@0:     : UObject(other)
michael@0: {   
michael@0:     m_breakiterator_            = other.m_breakiterator_;
michael@0:     m_text_                     = other.m_text_;
michael@0:     m_search_                   = (USearch *)uprv_malloc(sizeof(USearch));   
michael@0:     m_search_->breakIter        = other.m_search_->breakIter;
michael@0:     m_search_->isCanonicalMatch = other.m_search_->isCanonicalMatch;
michael@0:     m_search_->isOverlap        = other.m_search_->isOverlap;
michael@0:     m_search_->elementComparisonType = other.m_search_->elementComparisonType;
michael@0:     m_search_->matchedIndex     = other.m_search_->matchedIndex;
michael@0:     m_search_->matchedLength    = other.m_search_->matchedLength;
michael@0:     m_search_->text             = other.m_search_->text;
michael@0:     m_search_->textLength       = other.m_search_->textLength;
michael@0: }
michael@0: 
michael@0: SearchIterator::~SearchIterator()
michael@0: {
michael@0:     if (m_search_ != NULL) {
michael@0:         uprv_free(m_search_);
michael@0:     }
michael@0: }
michael@0: 
michael@0: // public get and set methods ----------------------------------------
michael@0: 
michael@0: void SearchIterator::setAttribute(USearchAttribute       attribute,
michael@0:                                   USearchAttributeValue  value,
michael@0:                                   UErrorCode            &status)
michael@0: {
michael@0:     if (U_SUCCESS(status)) {
michael@0:         switch (attribute)
michael@0:         {
michael@0:         case USEARCH_OVERLAP :
michael@0:             m_search_->isOverlap = (value == USEARCH_ON ? TRUE : FALSE);
michael@0:             break;
michael@0:         case USEARCH_CANONICAL_MATCH :
michael@0:             m_search_->isCanonicalMatch = (value == USEARCH_ON ? TRUE : FALSE);
michael@0:             break;
michael@0:         case USEARCH_ELEMENT_COMPARISON :
michael@0:             if (value == USEARCH_PATTERN_BASE_WEIGHT_IS_WILDCARD || value == USEARCH_ANY_BASE_WEIGHT_IS_WILDCARD) {
michael@0:                 m_search_->elementComparisonType = (int16_t)value;
michael@0:             } else {
michael@0:                 m_search_->elementComparisonType = 0;
michael@0:             }
michael@0:             break;
michael@0:         default:
michael@0:             status = U_ILLEGAL_ARGUMENT_ERROR;
michael@0:         }
michael@0:     }
michael@0:     if (value == USEARCH_ATTRIBUTE_VALUE_COUNT) {
michael@0:         status = U_ILLEGAL_ARGUMENT_ERROR;
michael@0:     }
michael@0: }
michael@0: 
michael@0: USearchAttributeValue SearchIterator::getAttribute(
michael@0:                                           USearchAttribute  attribute) const
michael@0: {
michael@0:     switch (attribute) {
michael@0:     case USEARCH_OVERLAP :
michael@0:         return (m_search_->isOverlap == TRUE ? USEARCH_ON : USEARCH_OFF);
michael@0:     case USEARCH_CANONICAL_MATCH :
michael@0:         return (m_search_->isCanonicalMatch == TRUE ? USEARCH_ON : 
michael@0:                                                                 USEARCH_OFF);
michael@0:     case USEARCH_ELEMENT_COMPARISON :
michael@0:         {
michael@0:             int16_t value = m_search_->elementComparisonType;
michael@0:             if (value == USEARCH_PATTERN_BASE_WEIGHT_IS_WILDCARD || value == USEARCH_ANY_BASE_WEIGHT_IS_WILDCARD) {
michael@0:                 return (USearchAttributeValue)value;
michael@0:             } else {
michael@0:                 return USEARCH_STANDARD_ELEMENT_COMPARISON;
michael@0:             }
michael@0:         }
michael@0:     default :
michael@0:         return USEARCH_DEFAULT;
michael@0:     }
michael@0: }
michael@0:     
michael@0: int32_t SearchIterator::getMatchedStart() const
michael@0: {
michael@0:     return m_search_->matchedIndex;
michael@0: }
michael@0: 
michael@0: int32_t SearchIterator::getMatchedLength() const
michael@0: {
michael@0:     return m_search_->matchedLength;
michael@0: }
michael@0:     
michael@0: void SearchIterator::getMatchedText(UnicodeString &result) const
michael@0: {
michael@0:     int32_t matchedindex  = m_search_->matchedIndex;
michael@0:     int32_t     matchedlength = m_search_->matchedLength;
michael@0:     if (matchedindex != USEARCH_DONE && matchedlength != 0) {
michael@0:         result.setTo(m_search_->text + matchedindex, matchedlength); 
michael@0:     }
michael@0:     else {
michael@0:         result.remove();
michael@0:     }
michael@0: }
michael@0:     
michael@0: void SearchIterator::setBreakIterator(BreakIterator *breakiter, 
michael@0:                                       UErrorCode &status)
michael@0: {
michael@0:     if (U_SUCCESS(status)) {
michael@0: #if 0
michael@0:         m_search_->breakIter = NULL;
michael@0:         // the c++ breakiterator may not make use of ubreakiterator.
michael@0:         // so we'll have to keep track of it ourselves.
michael@0: #else
michael@0:         // Well, gee... the Constructors that take a BreakIterator
michael@0:         // all cast the BreakIterator to a UBreakIterator and
michael@0:         // pass it to the corresponding usearch_openFromXXX
michael@0:         // routine, so there's no reason not to do this.
michael@0:         //
michael@0:         // Besides, a UBreakIterator is a BreakIterator, so
michael@0:         // any subclass of BreakIterator should work fine here...
michael@0:         m_search_->breakIter = (UBreakIterator *) breakiter;
michael@0: #endif
michael@0:         
michael@0:         m_breakiterator_ = breakiter;
michael@0:     }
michael@0: }
michael@0:     
michael@0: const BreakIterator * SearchIterator::getBreakIterator(void) const
michael@0: {
michael@0:     return m_breakiterator_;
michael@0: }
michael@0: 
michael@0: void SearchIterator::setText(const UnicodeString &text, UErrorCode &status)
michael@0: {
michael@0:     if (U_SUCCESS(status)) {
michael@0:         if (text.length() == 0) {
michael@0:             status = U_ILLEGAL_ARGUMENT_ERROR;
michael@0:         }
michael@0:         else {
michael@0:             m_text_        = text;
michael@0:             m_search_->text = m_text_.getBuffer();
michael@0:             m_search_->textLength = m_text_.length();
michael@0:         }
michael@0:     }
michael@0: }
michael@0: 
michael@0: void SearchIterator::setText(CharacterIterator &text, UErrorCode &status)
michael@0: {
michael@0:     if (U_SUCCESS(status)) {
michael@0:         text.getText(m_text_);
michael@0:         setText(m_text_, status);
michael@0:     }
michael@0: }
michael@0:     
michael@0: const UnicodeString & SearchIterator::getText(void) const
michael@0: {
michael@0:     return m_text_;
michael@0: }
michael@0: 
michael@0: // operator overloading ----------------------------------------------
michael@0: 
michael@0: UBool SearchIterator::operator==(const SearchIterator &that) const
michael@0: {
michael@0:     if (this == &that) {
michael@0:         return TRUE;
michael@0:     }
michael@0:     return (m_breakiterator_            == that.m_breakiterator_ &&
michael@0:             m_search_->isCanonicalMatch == that.m_search_->isCanonicalMatch &&
michael@0:             m_search_->isOverlap        == that.m_search_->isOverlap &&
michael@0:             m_search_->elementComparisonType == that.m_search_->elementComparisonType &&
michael@0:             m_search_->matchedIndex     == that.m_search_->matchedIndex &&
michael@0:             m_search_->matchedLength    == that.m_search_->matchedLength &&
michael@0:             m_search_->textLength       == that.m_search_->textLength &&
michael@0:             getOffset() == that.getOffset() &&
michael@0:             (uprv_memcmp(m_search_->text, that.m_search_->text, 
michael@0:                               m_search_->textLength * sizeof(UChar)) == 0));
michael@0: }
michael@0: 
michael@0: // public methods ----------------------------------------------------
michael@0: 
michael@0: int32_t SearchIterator::first(UErrorCode &status)
michael@0: {
michael@0:     if (U_FAILURE(status)) {
michael@0:         return USEARCH_DONE;
michael@0:     }
michael@0:     setOffset(0, status);
michael@0:     return handleNext(0, status);
michael@0: }
michael@0: 
michael@0: int32_t SearchIterator::following(int32_t position, 
michael@0:                                       UErrorCode &status)
michael@0: {
michael@0:     if (U_FAILURE(status)) {
michael@0:         return USEARCH_DONE;
michael@0:     }
michael@0:     setOffset(position, status);
michael@0:     return handleNext(position, status);
michael@0: }
michael@0:     
michael@0: int32_t SearchIterator::last(UErrorCode &status)
michael@0: {
michael@0:     if (U_FAILURE(status)) {
michael@0:         return USEARCH_DONE;
michael@0:     }
michael@0:     setOffset(m_search_->textLength, status);
michael@0:     return handlePrev(m_search_->textLength, status);
michael@0: }
michael@0: 
michael@0: int32_t SearchIterator::preceding(int32_t position, 
michael@0:                                       UErrorCode &status)
michael@0: {
michael@0:     if (U_FAILURE(status)) {
michael@0:         return USEARCH_DONE;
michael@0:     }
michael@0:     setOffset(position, status);
michael@0:     return handlePrev(position, status);
michael@0: }
michael@0: 
michael@0: int32_t SearchIterator::next(UErrorCode &status)
michael@0: {
michael@0:     if (U_SUCCESS(status)) {
michael@0:         int32_t offset = getOffset();
michael@0:         int32_t matchindex  = m_search_->matchedIndex;
michael@0:         int32_t     matchlength = m_search_->matchedLength;
michael@0:         m_search_->reset = FALSE;
michael@0:         if (m_search_->isForwardSearching == TRUE) {
michael@0:             int32_t textlength = m_search_->textLength;
michael@0:             if (offset == textlength || matchindex == textlength || 
michael@0:                 (matchindex != USEARCH_DONE && 
michael@0:                 matchindex + matchlength >= textlength)) {
michael@0:                 // not enough characters to match
michael@0:                 setMatchNotFound();
michael@0:                 return USEARCH_DONE; 
michael@0:             }
michael@0:         }
michael@0:         else {
michael@0:             // switching direction. 
michael@0:             // if matchedIndex == USEARCH_DONE, it means that either a 
michael@0:             // setOffset has been called or that previous ran off the text
michael@0:             // string. the iterator would have been set to offset 0 if a 
michael@0:             // match is not found.
michael@0:             m_search_->isForwardSearching = TRUE;
michael@0:             if (m_search_->matchedIndex != USEARCH_DONE) {
michael@0:                 // there's no need to set the collation element iterator
michael@0:                 // the next call to next will set the offset.
michael@0:                 return matchindex;
michael@0:             }
michael@0:         }
michael@0: 
michael@0:         if (matchlength > 0) {
michael@0:             // if matchlength is 0 we are at the start of the iteration
michael@0:             if (m_search_->isOverlap) {
michael@0:                 offset ++;
michael@0:             }
michael@0:             else {
michael@0:                 offset += matchlength;
michael@0:             }
michael@0:         }
michael@0:         return handleNext(offset, status);
michael@0:     }
michael@0:     return USEARCH_DONE;
michael@0: }
michael@0: 
michael@0: int32_t SearchIterator::previous(UErrorCode &status)
michael@0: {
michael@0:     if (U_SUCCESS(status)) {
michael@0:         int32_t offset;
michael@0:         if (m_search_->reset) {
michael@0:             offset                       = m_search_->textLength;
michael@0:             m_search_->isForwardSearching = FALSE;
michael@0:             m_search_->reset              = FALSE;
michael@0:             setOffset(offset, status);
michael@0:         }
michael@0:         else {
michael@0:             offset = getOffset();
michael@0:         }
michael@0:         
michael@0:         int32_t matchindex = m_search_->matchedIndex;
michael@0:         if (m_search_->isForwardSearching == TRUE) {
michael@0:             // switching direction. 
michael@0:             // if matchedIndex == USEARCH_DONE, it means that either a 
michael@0:             // setOffset has been called or that next ran off the text
michael@0:             // string. the iterator would have been set to offset textLength if 
michael@0:             // a match is not found.
michael@0:             m_search_->isForwardSearching = FALSE;
michael@0:             if (matchindex != USEARCH_DONE) {
michael@0:                 return matchindex;
michael@0:             }
michael@0:         }
michael@0:         else {
michael@0:             if (offset == 0 || matchindex == 0) {
michael@0:                 // not enough characters to match
michael@0:                 setMatchNotFound();
michael@0:                 return USEARCH_DONE; 
michael@0:             }
michael@0:         }
michael@0: 
michael@0:         if (matchindex != USEARCH_DONE) {
michael@0:             if (m_search_->isOverlap) {
michael@0:                 matchindex += m_search_->matchedLength - 2;
michael@0:             }
michael@0: 
michael@0:             return handlePrev(matchindex, status); 
michael@0:         }
michael@0: 
michael@0:         return handlePrev(offset, status);
michael@0:     }
michael@0: 
michael@0:     return USEARCH_DONE;
michael@0: }
michael@0: 
michael@0: void SearchIterator::reset()
michael@0: {
michael@0:     UErrorCode status = U_ZERO_ERROR;
michael@0:     setMatchNotFound();
michael@0:     setOffset(0, status);
michael@0:     m_search_->isOverlap          = FALSE;
michael@0:     m_search_->isCanonicalMatch   = FALSE;
michael@0:     m_search_->elementComparisonType = 0;
michael@0:     m_search_->isForwardSearching = TRUE;
michael@0:     m_search_->reset              = TRUE;
michael@0: }
michael@0: 
michael@0: // protected constructors and destructors -----------------------------
michael@0: 
michael@0: SearchIterator::SearchIterator()
michael@0: {
michael@0:     m_search_                     = (USearch *)uprv_malloc(sizeof(USearch));
michael@0:     m_search_->breakIter          = NULL;
michael@0:     m_search_->isOverlap          = FALSE;
michael@0:     m_search_->isCanonicalMatch   = FALSE;
michael@0:     m_search_->elementComparisonType = 0;
michael@0:     m_search_->isForwardSearching = TRUE;
michael@0:     m_search_->reset              = TRUE;
michael@0:     m_search_->matchedIndex       = USEARCH_DONE;
michael@0:     m_search_->matchedLength      = 0;
michael@0:     m_search_->text               = NULL;
michael@0:     m_search_->textLength         = 0;
michael@0:     m_breakiterator_              = NULL;
michael@0: }
michael@0: 
michael@0: SearchIterator::SearchIterator(const UnicodeString &text, 
michael@0:                                      BreakIterator *breakiter) :
michael@0:                                      m_breakiterator_(breakiter),
michael@0:                                      m_text_(text)
michael@0: {
michael@0:     m_search_                     = (USearch *)uprv_malloc(sizeof(USearch));
michael@0:     m_search_->breakIter          = NULL;
michael@0:     m_search_->isOverlap          = FALSE;
michael@0:     m_search_->isCanonicalMatch   = FALSE;
michael@0:     m_search_->elementComparisonType = 0;
michael@0:     m_search_->isForwardSearching = TRUE;
michael@0:     m_search_->reset              = TRUE;
michael@0:     m_search_->matchedIndex       = USEARCH_DONE;
michael@0:     m_search_->matchedLength      = 0;
michael@0:     m_search_->text               = m_text_.getBuffer();
michael@0:     m_search_->textLength         = text.length();
michael@0: }
michael@0: 
michael@0: SearchIterator::SearchIterator(CharacterIterator &text, 
michael@0:                                BreakIterator     *breakiter) :
michael@0:                                m_breakiterator_(breakiter)
michael@0: {
michael@0:     m_search_                     = (USearch *)uprv_malloc(sizeof(USearch));
michael@0:     m_search_->breakIter          = NULL;
michael@0:     m_search_->isOverlap          = FALSE;
michael@0:     m_search_->isCanonicalMatch   = FALSE;
michael@0:     m_search_->elementComparisonType = 0;
michael@0:     m_search_->isForwardSearching = TRUE;
michael@0:     m_search_->reset              = TRUE;
michael@0:     m_search_->matchedIndex       = USEARCH_DONE;
michael@0:     m_search_->matchedLength      = 0;
michael@0:     text.getText(m_text_);
michael@0:     m_search_->text               = m_text_.getBuffer();
michael@0:     m_search_->textLength         = m_text_.length();
michael@0:     m_breakiterator_             = breakiter;
michael@0: }
michael@0: 
michael@0: // protected methods ------------------------------------------------------
michael@0: 
michael@0: SearchIterator & SearchIterator::operator=(const SearchIterator &that)
michael@0: {
michael@0:     if (this != &that) {
michael@0:         m_breakiterator_            = that.m_breakiterator_;
michael@0:         m_text_                     = that.m_text_;
michael@0:         m_search_->breakIter        = that.m_search_->breakIter;
michael@0:         m_search_->isCanonicalMatch = that.m_search_->isCanonicalMatch;
michael@0:         m_search_->isOverlap        = that.m_search_->isOverlap;
michael@0:         m_search_->elementComparisonType = that.m_search_->elementComparisonType;
michael@0:         m_search_->matchedIndex     = that.m_search_->matchedIndex;
michael@0:         m_search_->matchedLength    = that.m_search_->matchedLength;
michael@0:         m_search_->text             = that.m_search_->text;
michael@0:         m_search_->textLength       = that.m_search_->textLength;
michael@0:     }
michael@0:     return *this;
michael@0: }
michael@0: 
michael@0: void SearchIterator::setMatchLength(int32_t length)
michael@0: {
michael@0:     m_search_->matchedLength = length;
michael@0: }
michael@0: 
michael@0: void SearchIterator::setMatchStart(int32_t position)
michael@0: {
michael@0:     m_search_->matchedIndex = position;
michael@0: }
michael@0: 
michael@0: void SearchIterator::setMatchNotFound() 
michael@0: {
michael@0:     setMatchStart(USEARCH_DONE);
michael@0:     setMatchLength(0);
michael@0:     UErrorCode status = U_ZERO_ERROR;
michael@0:     // by default no errors should be returned here since offsets are within 
michael@0:     // range.
michael@0:     if (m_search_->isForwardSearching) {
michael@0:         setOffset(m_search_->textLength, status);
michael@0:     }
michael@0:     else {
michael@0:         setOffset(0, status);
michael@0:     }
michael@0: }
michael@0: 
michael@0: 
michael@0: U_NAMESPACE_END
michael@0: 
michael@0: #endif /* #if !UCONFIG_NO_COLLATION */