offapi/com/sun/star/util/XTextSearch.idl

   1 /*************************************************************************
   2  *
   3  * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
   4  *
   5  * Copyright 2008 by Sun Microsystems, Inc.
   6  *
   7  * OpenOffice.org - a multi-platform office productivity suite
   8  *
   9  * $RCSfile: XTextSearch.idl,v $
  10  * $Revision: 1.13 $
  11  *
  12  * This file is part of OpenOffice.org.
  13  *
  14  * OpenOffice.org is free software: you can redistribute it and/or modify
  15  * it under the terms of the GNU Lesser General Public License version 3
  16  * only, as published by the Free Software Foundation.
  17  *
  18  * OpenOffice.org is distributed in the hope that it will be useful,
  19  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  20  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  21  * GNU Lesser General Public License version 3 for more details
  22  * (a copy is included in the LICENSE file that accompanied this code).
  23  *
  24  * You should have received a copy of the GNU Lesser General Public License
  25  * version 3 along with OpenOffice.org.  If not, see
  26  * <http://www.openoffice.org/license.html>
  27  * for a copy of the LGPLv3 License.
  28  *
  29  ************************************************************************/
  30
  31 #ifndef __com_sun_star_lang_XTextSearch_idl__
  32 #define __com_sun_star_lang_XTextSearch_idl__
  33
  34
  35 #include <com/sun/star/lang/Locale.idl>
  36 #include <com/sun/star/uno/XInterface.idl>
  37 //#include <com/sun/star/lang/CascadeTransliterator.idl>
  38
  39 //=============================================================================
  40
  41 module com { module sun { module star { module util {
  42
  43 //=============================================================================
  44
  45
  46 published enum SearchAlgorithms {
  47     ABSOLUTE,           // "normal" a kind of Boyer-Moore
  48     REGEXP,                 // regular expression
  49     APPROXIMATE         // Leveinstein distance
  50 };
  51
  52 published constants SearchFlags
  53 {
  54     //Flag for all search methods
  55
  56     /**
  57         @deprecated : the constant ALL_IGNORE_CASE is never supported -
  58                         it must use the transliteration flags of
  59                         the SearchOptions.
  60                         <p>@see TransliterationModulesNew
  61     */
  62     const long  ALL_IGNORE_CASE         = 0x00000001;
  63
  64     /// Flag for normal (Boyer-Moore) search
  65     const long  NORM_WORD_ONLY          = 0x00000010;
  66
  67     /// Flag for "regular expression" search / interpret as extended regular expression
  68     const long  REG_EXTENDED            = 0x00000100;
  69     /** Flag for "regular expression" search / No replace, i.e., avoid sub regular
  70          expresions, return true/false to match
  71
  72 <!-- JRH: Check this for sense of the expression. -->
  73     */
  74     const long  REG_NOSUB               = 0x00000200;
  75
  76     /// Flag for "regular expression" search / Special new  line treatment
  77     const long  REG_NEWLINE             = 0x00000400;
  78
  79     /** A NEWLINE character in string will not be matched by  a period outside bracket
  80         expression or by any form of a non matching list.
  81         A circumflex (^) in pattern when used to specify expression anchoring
  82 <!-- JRH: anhoring to anchoring.  -->
  83         will match the zero length string immediately after a newline in string,
  84         regardless of the setting of REG_NOTBOL
  85         A dollar-sign ($) in pattern when used to specify expression anchoring,
  86         will match zero-length string immediately before a new line in string,
  87         regardless of the setting of REG_NOTEOL
  88 */
  89     const long  REG_NOT_BEGINOFLINE     = 0x00000800;
  90
  91     /** The first character in the string is not the beginning of the line therefore ^  will not
  92         match with first character of the string
  93     */
  94     const long  REG_NOT_ENDOFLINE       = 0x00001000;
  95
  96
  97     /// Flags for "Weight Levenshtein-Distance" search
  98     const long  LEV_RELAXED             = 0x00010000;
  99 };
 100
 101
 102 published  struct SearchOptions  {
 103     //-------------------------------------------------------------------------
 104     /// search type, can be: ABSOLUTE, REGEXP, APPROXIMATE
 105     SearchAlgorithms    algorithmType;
 106
 107     /** some flags - can be mixed
 108
 109         @see SearchFlags
 110     */
 111     long                        searchFlag;
 112
 113     /// the search text
 114     string                      searchString;
 115
 116     /** the replace text
 117     (is for optional replacing - SearchOption is only the data container for it)*/
 118     string                      replaceString;
 119
 120     /// this is the language for case insensitive search
 121     ::com::sun::star::lang::Locale  Locale;
 122
 123     /** this many characters can be different between the found and search word
 124     in a "Weight Levenshtein-Distance"*/
 125     long                        changedChars;
 126
 127     /** this many characters can be missed in the found word
 128     in a "Weight Levenshtein-Distance"*/
 129     long                        deletedChars;
 130
 131     /// this many characters can be additional in the found word in a "Weight Levenshtein-Distance"
 132     long                        insertedChars;
 133
 134     /** asian flags for the transliteration. Same meaning as the enum of TransliteratorModule
 135
 136         @see com/sun/star/i18n/XTransliteration.idl
 137     */
 138     long                        transliterateFlags;
 139 };
 140
 141
 142 published  struct SearchResult  {
 143     //-------------------------------------------------------------------------
 144     /** Number of subexpressions,
 145     if it is 0, then no match found; this value is 1 for ABSOLUTE and APPROXIMATE match.
 146     The start and endOffset are always dependent on the search direction.
 147     For example:
 148     if you search "X" in the text "-X-" the offset are:
 149         for forward:    start = 1, end = 2
 150         for backward:   start = 2, end = 1
 151     Forward, the startOffset is inclusive, the endOffset exclusive.
 152     Backward, the startOffset is exclusive, the endOffset inclusive.
 153
 154     For regular expressions it can be greater than 1.
 155     If the value is 1, startoffset[0] and endoffset[0] points to the matching sub string
 156     if value is > 1, still startoffset[0] and endoffset[0] points to the matching substring for whole regular expression
 157     startoffset[i] and endoffset[i] points to the matching substring of i th matching substring.
 158     */
 159     long subRegExpressions;
 160     sequence<long> startOffset;         // inclusive
 161     sequence<long> endOffset;           // exclusive
 162 };
 163
 164
 165
 166 /** enables an object to search in its content.
 167  */
 168 published interface XTextSearch : com::sun::star::uno::XInterface
 169 {
 170     //-------------------------------------------------------------------------
 171     /** set the options for the forward or backward search.
 172
 173     */
 174     void setOptions ([in] SearchOptions options);
 175     //-------------------------------------------------------------------------
 176     /** search forward in the searchStr, starts at startPos and ends by endpos.
 177         The result is returned in the SearchResult.
 178
 179     */
 180     SearchResult  searchForward  ([in] string searchStr, [in] long startPos, [in] long endPos );
 181     //-------------------------------------------------------------------------
 182     /** search backward in the searchStr, starts at startPos and ends by endpos.
 183         The endpos must be lower then the startpos, because the function searches backward!
 184         The result is returned in the SearchResult.
 185
 186     */
 187     SearchResult  searchBackward ([in] string searchStr, [in] long startPos, [in] long endPos );
 188 };
 189
 190 //=============================================================================
 191 }; }; }; };
 192
 193 #endif