icuSources/i18n/unicode/search.h

   1 /*
   2 **********************************************************************
   3 *   Copyright (C) 2001-2011 IBM and others. All rights reserved.
   4 **********************************************************************
   5 *   Date        Name        Description
   6 *  03/22/2000   helena      Creation.
   7 **********************************************************************
   8 */
   9
  10 #ifndef SEARCH_H
  11 #define SEARCH_H
  12
  13 #include "unicode/utypes.h"
  14
  15 /**
  16  * \file
  17  * \brief C++ API: SearchIterator object.
  18  */
  19
  20 #if !UCONFIG_NO_COLLATION && !UCONFIG_NO_BREAK_ITERATION
  21
  22 #include "unicode/uobject.h"
  23 #include "unicode/unistr.h"
  24 #include "unicode/chariter.h"
  25 #include "unicode/brkiter.h"
  26 #include "unicode/usearch.h"
  27
  28 /**
  29 * @stable ICU 2.0
  30 */
  31 struct USearch;
  32 /**
  33 * @stable ICU 2.0
  34 */
  35 typedef struct USearch USearch;
  36
  37 U_NAMESPACE_BEGIN
  38
  39 /**
  40  *
  41  * <tt>SearchIterator</tt> is an abstract base class that provides
  42  * methods to search for a pattern within a text string. Instances of
  43  * <tt>SearchIterator</tt> maintain a current position and scans over the
  44  * target text, returning the indices the pattern is matched and the length
  45  * of each match.
  46  * <p>
  47  * <tt>SearchIterator</tt> defines a protocol for text searching.
  48  * Subclasses provide concrete implementations of various search algorithms.
  49  * For example, <tt>StringSearch</tt> implements language-sensitive pattern
  50  * matching based on the comparison rules defined in a
  51  * <tt>RuleBasedCollator</tt> object.
  52  * <p>
  53  * Other options for searching includes using a BreakIterator to restrict
  54  * the points at which matches are detected.
  55  * <p>
  56  * <tt>SearchIterator</tt> provides an API that is similar to that of
  57  * other text iteration classes such as <tt>BreakIterator</tt>. Using
  58  * this class, it is easy to scan through text looking for all occurances of
  59  * a given pattern. The following example uses a <tt>StringSearch</tt>
  60  * object to find all instances of "fox" in the target string. Any other
  61  * subclass of <tt>SearchIterator</tt> can be used in an identical
  62  * manner.
  63  * <pre><code>
  64  * UnicodeString target("The quick brown fox jumped over the lazy fox");
  65  * UnicodeString pattern("fox");
  66  *
  67  * SearchIterator *iter  = new StringSearch(pattern, target);
  68  * UErrorCode      error = U_ZERO_ERROR;
  69  * for (int pos = iter->first(error); pos != USEARCH_DONE;
  70  *                               pos = iter->next(error)) {
  71  *     printf("Found match at %d pos, length is %d\n", pos,
  72  *                                             iter.getMatchLength());
  73  * }
  74  * </code></pre>
  75  *
  76  * @see StringSearch
  77  * @see RuleBasedCollator
  78  */
  79 class U_I18N_API SearchIterator : public UObject {
  80
  81 public:
  82
  83     // public constructors and destructors -------------------------------
  84
  85     /**
  86     * Copy constructor that creates a SearchIterator instance with the same
  87     * behavior, and iterating over the same text.
  88     * @param other the SearchIterator instance to be copied.
  89     * @stable ICU 2.0
  90     */
  91     SearchIterator(const SearchIterator &other);
  92
  93     /**
  94      * Destructor. Cleans up the search iterator data struct.
  95      * @stable ICU 2.0
  96      */
  97     virtual ~SearchIterator();
  98
  99     // public get and set methods ----------------------------------------
 100
 101     /**
 102      * Sets the index to point to the given position, and clears any state
 103      * that's affected.
 104      * <p>
 105      * This method takes the argument index and sets the position in the text
 106      * string accordingly without checking if the index is pointing to a
 107      * valid starting point to begin searching.
 108      * @param position within the text to be set. If position is less
 109      *             than or greater than the text range for searching,
 110      *          an U_INDEX_OUTOFBOUNDS_ERROR will be returned
 111      * @param status for errors if it occurs
 112      * @stable ICU 2.0
 113      */
 114     virtual void setOffset(int32_t position, UErrorCode &status) = 0;
 115
 116     /**
 117      * Return the current index in the text being searched.
 118      * If the iteration has gone past the end of the text
 119      * (or past the beginning for a backwards search), USEARCH_DONE
 120      * is returned.
 121      * @return current index in the text being searched.
 122      * @stable ICU 2.0
 123      */
 124     virtual int32_t getOffset(void) const = 0;
 125
 126     /**
 127     * Sets the text searching attributes located in the enum
 128     * USearchAttribute with values from the enum USearchAttributeValue.
 129     * USEARCH_DEFAULT can be used for all attributes for resetting.
 130     * @param attribute text attribute (enum USearchAttribute) to be set
 131     * @param value text attribute value
 132     * @param status for errors if it occurs
 133     * @stable ICU 2.0
 134     */
 135     void setAttribute(USearchAttribute       attribute,
 136                       USearchAttributeValue  value,
 137                       UErrorCode            &status);
 138
 139     /**
 140     * Gets the text searching attributes
 141     * @param attribute text attribute (enum USearchAttribute) to be retrieve
 142     * @return text attribute value
 143     * @stable ICU 2.0
 144     */
 145     USearchAttributeValue getAttribute(USearchAttribute  attribute) const;
 146
 147     /**
 148     * Returns the index to the match in the text string that was searched.
 149     * This call returns a valid result only after a successful call to
 150     * <tt>first</tt>, <tt>next</tt>, <tt>previous</tt>, or <tt>last</tt>.
 151     * Just after construction, or after a searching method returns
 152     * <tt>USEARCH_DONE</tt>, this method will return <tt>USEARCH_DONE</tt>.
 153     * <p>
 154     * Use getMatchedLength to get the matched string length.
 155     * @return index of a substring within the text string that is being
 156     *         searched.
 157     * @see #first
 158     * @see #next
 159     * @see #previous
 160     * @see #last
 161     * @stable ICU 2.0
 162     */
 163     int32_t getMatchedStart(void) const;
 164
 165     /**
 166      * Returns the length of text in the string which matches the search
 167      * pattern. This call returns a valid result only after a successful call
 168      * to <tt>first</tt>, <tt>next</tt>, <tt>previous</tt>, or <tt>last</tt>.
 169      * Just after construction, or after a searching method returns
 170      * <tt>USEARCH_DONE</tt>, this method will return 0.
 171      * @return The length of the match in the target text, or 0 if there
 172      *         is no match currently.
 173      * @see #first
 174      * @see #next
 175      * @see #previous
 176      * @see #last
 177      * @stable ICU 2.0
 178      */
 179     int32_t getMatchedLength(void) const;
 180
 181     /**
 182      * Returns the text that was matched by the most recent call to
 183      * <tt>first</tt>, <tt>next</tt>, <tt>previous</tt>, or <tt>last</tt>.
 184      * If the iterator is not pointing at a valid match (e.g. just after
 185      * construction or after <tt>USEARCH_DONE</tt> has been returned,
 186      * returns an empty string.
 187      * @param result stores the matched string or an empty string if a match
 188      *        is not found.
 189      * @see #first
 190      * @see #next
 191      * @see #previous
 192      * @see #last
 193      * @stable ICU 2.0
 194      */
 195     void getMatchedText(UnicodeString &result) const;
 196
 197     /**
 198      * Set the BreakIterator that will be used to restrict the points
 199      * at which matches are detected. The user is responsible for deleting
 200      * the breakiterator.
 201      * @param breakiter A BreakIterator that will be used to restrict the
 202      *                points at which matches are detected. If a match is
 203      *                found, but the match's start or end index is not a
 204      *                boundary as determined by the <tt>BreakIterator</tt>,
 205      *                the match will be rejected and another will be searched
 206      *                for. If this parameter is <tt>NULL</tt>, no break
 207      *                detection is attempted.
 208      * @param status for errors if it occurs
 209      * @see BreakIterator
 210      * @stable ICU 2.0
 211      */
 212     void setBreakIterator(BreakIterator *breakiter, UErrorCode &status);
 213
 214     /**
 215      * Returns the BreakIterator that is used to restrict the points at
 216      * which matches are detected.  This will be the same object that was
 217      * passed to the constructor or to <tt>setBreakIterator</tt>.
 218      * Note that <tt>NULL</tt> is a legal value; it means that break
 219      * detection should not be attempted.
 220      * @return BreakIterator used to restrict matchings.
 221      * @see #setBreakIterator
 222      * @stable ICU 2.0
 223      */
 224     const BreakIterator * getBreakIterator(void) const;
 225
 226     /**
 227      * Set the string text to be searched. Text iteration will hence begin at
 228      * the start of the text string. This method is useful if you want to
 229      * re-use an iterator to search for the same pattern within a different
 230      * body of text. The user is responsible for deleting the text.
 231      * @param text string to be searched.
 232      * @param status for errors. If the text length is 0,
 233      *        an U_ILLEGAL_ARGUMENT_ERROR is returned.
 234      * @stable ICU 2.0
 235      */
 236     virtual void setText(const UnicodeString &text, UErrorCode &status);
 237
 238     /**
 239      * Set the string text to be searched. Text iteration will hence begin at
 240      * the start of the text string. This method is useful if you want to
 241      * re-use an iterator to search for the same pattern within a different
 242      * body of text.
 243      * <p>
 244      * Note: No parsing of the text within the <tt>CharacterIterator</tt>
 245      * will be done during searching for this version. The block of text
 246      * in <tt>CharacterIterator</tt> will be used as it is.
 247      * The user is responsible for deleting the text.
 248      * @param text string iterator to be searched.
 249      * @param status for errors if any. If the text length is 0 then an
 250      *        U_ILLEGAL_ARGUMENT_ERROR is returned.
 251      * @stable ICU 2.0
 252      */
 253     virtual void setText(CharacterIterator &text, UErrorCode &status);
 254
 255     /**
 256      * Return the string text to be searched.
 257      * @return text string to be searched.
 258      * @stable ICU 2.0
 259      */
 260     const UnicodeString & getText(void) const;
 261
 262     // operator overloading ----------------------------------------------
 263
 264     /**
 265      * Equality operator.
 266      * @param that SearchIterator instance to be compared.
 267      * @return TRUE if both BreakIterators are of the same class, have the
 268      *         same behavior, terates over the same text and have the same
 269      *         attributes. FALSE otherwise.
 270      * @stable ICU 2.0
 271      */
 272     virtual UBool operator==(const SearchIterator &that) const;
 273
 274     /**
 275      * Not-equal operator.
 276      * @param that SearchIterator instance to be compared.
 277      * @return FALSE if operator== returns TRUE, and vice versa.
 278      * @stable ICU 2.0
 279      */
 280     UBool operator!=(const SearchIterator &that) const;
 281
 282     // public methods ----------------------------------------------------
 283
 284     /**
 285      * Returns a copy of SearchIterator with the same behavior, and
 286      * iterating over the same text, as this one. Note that all data will be
 287      * replicated, except for the text string to be searched.
 288      * @return cloned object
 289      * @stable ICU 2.0
 290      */
 291     virtual SearchIterator* safeClone(void) const = 0;
 292
 293     /**
 294      * Returns the first index at which the string text matches the search
 295      * pattern. The iterator is adjusted so that its current index (as
 296      * returned by <tt>getOffset</tt>) is the match position if one
 297      * was found.
 298      * If a match is not found, <tt>USEARCH_DONE</tt> will be returned and
 299      * the iterator will be adjusted to the index USEARCH_DONE
 300      * @param  status for errors if it occurs
 301      * @return The character index of the first match, or
 302      *         <tt>USEARCH_DONE</tt> if there are no matches.
 303      * @see #getOffset
 304      * @stable ICU 2.0
 305      */
 306     int32_t first(UErrorCode &status);
 307
 308     /**
 309      * Returns the first index equal or greater than <tt>position</tt> at which the
 310      * string text matches the search pattern. The iterator is adjusted so
 311      * that its current index (as returned by <tt>getOffset</tt>) is the
 312      * match position if one was found.
 313      * If a match is not found, <tt>USEARCH_DONE</tt> will be returned and the
 314      * iterator will be adjusted to the index <tt>USEARCH_DONE</tt>.
 315      * @param  position where search if to start from. If position is less
 316      *             than or greater than the text range for searching,
 317      *          an U_INDEX_OUTOFBOUNDS_ERROR will be returned
 318      * @param  status for errors if it occurs
 319      * @return The character index of the first match following
 320      *         <tt>position</tt>, or <tt>USEARCH_DONE</tt> if there are no
 321      *         matches.
 322      * @see #getOffset
 323      * @stable ICU 2.0
 324      */
 325     int32_t following(int32_t position, UErrorCode &status);
 326
 327     /**
 328      * Returns the last index in the target text at which it matches the
 329      * search pattern. The iterator is adjusted so that its current index
 330      * (as returned by <tt>getOffset</tt>) is the match position if one was
 331      * found.
 332      * If a match is not found, <tt>USEARCH_DONE</tt> will be returned and
 333      * the iterator will be adjusted to the index USEARCH_DONE.
 334      * @param  status for errors if it occurs
 335      * @return The index of the first match, or <tt>USEARCH_DONE</tt> if
 336      *         there are no matches.
 337      * @see #getOffset
 338      * @stable ICU 2.0
 339      */
 340     int32_t last(UErrorCode &status);
 341
 342     /**
 343      * Returns the first index less than <tt>position</tt> at which the string
 344      * text matches the search pattern. The iterator is adjusted so that its
 345      * current index (as returned by <tt>getOffset</tt>) is the match
 346      * position if one was found. If a match is not found,
 347      * <tt>USEARCH_DONE</tt> will be returned and the iterator will be
 348      * adjusted to the index USEARCH_DONE
 349      * <p>
 350      * When <tt>USEARCH_OVERLAP</tt> option is off, the last index of the
 351      * result match is always less than <tt>position</tt>.
 352      * When <tt>USERARCH_OVERLAP</tt> is on, the result match may span across
 353      * <tt>position</tt>.
 354      *
 355      * @param  position where search is to start from. If position is less
 356      *             than or greater than the text range for searching,
 357      *          an U_INDEX_OUTOFBOUNDS_ERROR will be returned
 358      * @param  status for errors if it occurs
 359      * @return The character index of the first match preceding
 360      *         <tt>position</tt>, or <tt>USEARCH_DONE</tt> if there are
 361      *         no matches.
 362      * @see #getOffset
 363      * @stable ICU 2.0
 364      */
 365     int32_t preceding(int32_t position, UErrorCode &status);
 366
 367     /**
 368      * Returns the index of the next point at which the text matches the
 369      * search pattern, starting from the current position
 370      * The iterator is adjusted so that its current index (as returned by
 371      * <tt>getOffset</tt>) is the match position if one was found.
 372      * If a match is not found, <tt>USEARCH_DONE</tt> will be returned and
 373      * the iterator will be adjusted to a position after the end of the text
 374      * string.
 375      * @param  status for errors if it occurs
 376      * @return The index of the next match after the current position,
 377      *          or <tt>USEARCH_DONE</tt> if there are no more matches.
 378      * @see #getOffset
 379      * @stable ICU 2.0
 380      */
 381      int32_t next(UErrorCode &status);
 382
 383     /**
 384      * Returns the index of the previous point at which the string text
 385      * matches the search pattern, starting at the current position.
 386      * The iterator is adjusted so that its current index (as returned by
 387      * <tt>getOffset</tt>) is the match position if one was found.
 388      * If a match is not found, <tt>USEARCH_DONE</tt> will be returned and
 389      * the iterator will be adjusted to the index USEARCH_DONE
 390      * @param  status for errors if it occurs
 391      * @return The index of the previous match before the current position,
 392      *          or <tt>USEARCH_DONE</tt> if there are no more matches.
 393      * @see #getOffset
 394      * @stable ICU 2.0
 395      */
 396     int32_t previous(UErrorCode &status);
 397
 398     /**
 399     * Resets the iteration.
 400     * Search will begin at the start of the text string if a forward
 401     * iteration is initiated before a backwards iteration. Otherwise if a
 402     * backwards iteration is initiated before a forwards iteration, the
 403     * search will begin at the end of the text string.
 404     * @stable ICU 2.0
 405     */
 406     virtual void reset();
 407
 408 protected:
 409     // protected data members ---------------------------------------------
 410
 411     /**
 412     * C search data struct
 413     * @stable ICU 2.0
 414     */
 415     USearch *m_search_;
 416
 417     /**
 418     * Break iterator.
 419     * Currently the C++ breakiterator does not have getRules etc to reproduce
 420     * another in C. Hence we keep the original around and do the verification
 421     * at the end of the match. The user is responsible for deleting this
 422     * break iterator.
 423     * @stable ICU 2.0
 424     */
 425     BreakIterator *m_breakiterator_;
 426
 427     /**
 428     * Unicode string version of the search text
 429     * @stable ICU 2.0
 430     */
 431     UnicodeString  m_text_;
 432
 433     // protected constructors and destructors -----------------------------
 434
 435     /**
 436     * Default constructor.
 437     * Initializes data to the default values.
 438     * @stable ICU 2.0
 439     */
 440     SearchIterator();
 441
 442     /**
 443      * Constructor for use by subclasses.
 444      * @param text The target text to be searched.
 445      * @param breakiter A {@link BreakIterator} that is used to restrict the
 446      *                points at which matches are detected. If
 447      *                <tt>handleNext</tt> or <tt>handlePrev</tt> finds a
 448      *                match, but the match's start or end index is not a
 449      *                boundary as determined by the <tt>BreakIterator</tt>,
 450      *                the match is rejected and <tt>handleNext</tt> or
 451      *                <tt>handlePrev</tt> is called again. If this parameter
 452      *                is <tt>NULL</tt>, no break detection is attempted.
 453      * @see #handleNext
 454      * @see #handlePrev
 455      * @stable ICU 2.0
 456      */
 457     SearchIterator(const UnicodeString &text,
 458                          BreakIterator *breakiter = NULL);
 459
 460     /**
 461      * Constructor for use by subclasses.
 462      * <p>
 463      * Note: No parsing of the text within the <tt>CharacterIterator</tt>
 464      * will be done during searching for this version. The block of text
 465      * in <tt>CharacterIterator</tt> will be used as it is.
 466      * @param text The target text to be searched.
 467      * @param breakiter A {@link BreakIterator} that is used to restrict the
 468      *                points at which matches are detected. If
 469      *                <tt>handleNext</tt> or <tt>handlePrev</tt> finds a
 470      *                match, but the match's start or end index is not a
 471      *                boundary as determined by the <tt>BreakIterator</tt>,
 472      *                the match is rejected and <tt>handleNext</tt> or
 473      *                <tt>handlePrev</tt> is called again. If this parameter
 474      *                is <tt>NULL</tt>, no break detection is attempted.
 475      * @see #handleNext
 476      * @see #handlePrev
 477      * @stable ICU 2.0
 478      */
 479     SearchIterator(CharacterIterator &text, BreakIterator *breakiter = NULL);
 480
 481     // protected methods --------------------------------------------------
 482
 483     /**
 484      * Assignment operator. Sets this iterator to have the same behavior,
 485      * and iterate over the same text, as the one passed in.
 486      * @param that instance to be copied.
 487      * @stable ICU 2.0
 488      */
 489     SearchIterator & operator=(const SearchIterator &that);
 490
 491     /**
 492      * Abstract method which subclasses override to provide the mechanism
 493      * for finding the next match in the target text. This allows different
 494      * subclasses to provide different search algorithms.
 495      * <p>
 496      * If a match is found, the implementation should return the index at
 497      * which the match starts and should call
 498      * <tt>setMatchLength</tt> with the number of characters
 499      * in the target text that make up the match. If no match is found, the
 500      * method should return USEARCH_DONE.
 501      * <p>
 502      * @param position The index in the target text at which the search
 503      *                 should start.
 504      * @param status for error codes if it occurs.
 505      * @return index at which the match starts, else if match is not found
 506      *         USEARCH_DONE is returned
 507      * @see #setMatchLength
 508      * @stable ICU 2.0
 509      */
 510     virtual int32_t handleNext(int32_t position, UErrorCode &status)
 511                                                                          = 0;
 512
 513     /**
 514      * Abstract method which subclasses override to provide the mechanism for
 515      * finding the previous match in the target text. This allows different
 516      * subclasses to provide different search algorithms.
 517      * <p>
 518      * If a match is found, the implementation should return the index at
 519      * which the match starts and should call
 520      * <tt>setMatchLength</tt> with the number of characters
 521      * in the target text that make up the match. If no match is found, the
 522      * method should return USEARCH_DONE.
 523      * <p>
 524      * @param position The index in the target text at which the search
 525      *                 should start.
 526      * @param status for error codes if it occurs.
 527      * @return index at which the match starts, else if match is not found
 528      *         USEARCH_DONE is returned
 529      * @see #setMatchLength
 530      * @stable ICU 2.0
 531      */
 532      virtual int32_t handlePrev(int32_t position, UErrorCode &status)
 533                                                                          = 0;
 534
 535     /**
 536      * Sets the length of the currently matched string in the text string to
 537      * be searched.
 538      * Subclasses' <tt>handleNext</tt> and <tt>handlePrev</tt>
 539      * methods should call this when they find a match in the target text.
 540      * @param length length of the matched text.
 541      * @see #handleNext
 542      * @see #handlePrev
 543      * @stable ICU 2.0
 544      */
 545     virtual void setMatchLength(int32_t length);
 546
 547     /**
 548      * Sets the offset of the currently matched string in the text string to
 549      * be searched.
 550      * Subclasses' <tt>handleNext</tt> and <tt>handlePrev</tt>
 551      * methods should call this when they find a match in the target text.
 552      * @param position start offset of the matched text.
 553      * @see #handleNext
 554      * @see #handlePrev
 555      * @stable ICU 2.0
 556      */
 557     virtual void setMatchStart(int32_t position);
 558
 559     /**
 560     * sets match not found
 561     * @stable ICU 2.0
 562     */
 563     void setMatchNotFound();
 564 };
 565
 566 inline UBool SearchIterator::operator!=(const SearchIterator &that) const
 567 {
 568    return !operator==(that);
 569 }
 570 U_NAMESPACE_END
 571
 572 #endif /* #if !UCONFIG_NO_COLLATION */
 573
 574 #endif
 575