Home | History | Annotate | Download | only in i18n
      1 /*
      2  **********************************************************************
      3  *   Copyright (C) 1999-2008, International Business Machines
      4  *   Corporation and others.  All Rights Reserved.
      5  **********************************************************************
      6  *   Date        Name        Description
      7  *   11/17/99    aliu        Creation.
      8  **********************************************************************
      9  */
     10 
     11 #include "unicode/utypes.h"
     12 
     13 #if !UCONFIG_NO_TRANSLITERATION
     14 
     15 #include "unicode/rep.h"
     16 #include "unicode/unifilt.h"
     17 #include "unicode/uniset.h"
     18 #include "rbt_rule.h"
     19 #include "rbt_data.h"
     20 #include "cmemory.h"
     21 #include "strmatch.h"
     22 #include "strrepl.h"
     23 #include "util.h"
     24 #include "putilimp.h"
     25 
     26 static const UChar FORWARD_OP[] = {32,62,32,0}; // " > "
     27 
     28 U_NAMESPACE_BEGIN
     29 
     30 /**
     31  * Construct a new rule with the given input, output text, and other
     32  * attributes.  A cursor position may be specified for the output text.
     33  * @param input input string, including key and optional ante and
     34  * post context
     35  * @param anteContextPos offset into input to end of ante context, or -1 if
     36  * none.  Must be <= input.length() if not -1.
     37  * @param postContextPos offset into input to start of post context, or -1
     38  * if none.  Must be <= input.length() if not -1, and must be >=
     39  * anteContextPos.
     40  * @param output output string
     41  * @param cursorPosition offset into output at which cursor is located, or -1 if
     42  * none.  If less than zero, then the cursor is placed after the
     43  * <code>output</code>; that is, -1 is equivalent to
     44  * <code>output.length()</code>.  If greater than
     45  * <code>output.length()</code> then an exception is thrown.
     46  * @param segs array of UnicodeFunctors corresponding to input pattern
     47  * segments, or null if there are none.  The array itself is adopted,
     48  * but the pointers within it are not.
     49  * @param segsCount number of elements in segs[]
     50  * @param anchorStart TRUE if the the rule is anchored on the left to
     51  * the context start
     52  * @param anchorEnd TRUE if the rule is anchored on the right to the
     53  * context limit
     54  */
     55 TransliterationRule::TransliterationRule(const UnicodeString& input,
     56                                          int32_t anteContextPos, int32_t postContextPos,
     57                                          const UnicodeString& outputStr,
     58                                          int32_t cursorPosition, int32_t cursorOffset,
     59                                          UnicodeFunctor** segs,
     60                                          int32_t segsCount,
     61                                          UBool anchorStart, UBool anchorEnd,
     62                                          const TransliterationRuleData* theData,
     63                                          UErrorCode& status) :
     64     UMemory(),
     65     segments(0),
     66     data(theData) {
     67 
     68     if (U_FAILURE(status)) {
     69         return;
     70     }
     71     // Do range checks only when warranted to save time
     72     if (anteContextPos < 0) {
     73         anteContextLength = 0;
     74     } else {
     75         if (anteContextPos > input.length()) {
     76             // throw new IllegalArgumentException("Invalid ante context");
     77             status = U_ILLEGAL_ARGUMENT_ERROR;
     78             return;
     79         }
     80         anteContextLength = anteContextPos;
     81     }
     82     if (postContextPos < 0) {
     83         keyLength = input.length() - anteContextLength;
     84     } else {
     85         if (postContextPos < anteContextLength ||
     86             postContextPos > input.length()) {
     87             // throw new IllegalArgumentException("Invalid post context");
     88             status = U_ILLEGAL_ARGUMENT_ERROR;
     89             return;
     90         }
     91         keyLength = postContextPos - anteContextLength;
     92     }
     93     if (cursorPosition < 0) {
     94         cursorPosition = outputStr.length();
     95     } else if (cursorPosition > outputStr.length()) {
     96         // throw new IllegalArgumentException("Invalid cursor position");
     97         status = U_ILLEGAL_ARGUMENT_ERROR;
     98         return;
     99     }
    100     // We don't validate the segments array.  The caller must
    101     // guarantee that the segments are well-formed (that is, that
    102     // all $n references in the output refer to indices of this
    103     // array, and that no array elements are null).
    104     this->segments = segs;
    105     this->segmentsCount = segsCount;
    106 
    107     pattern = input;
    108     flags = 0;
    109     if (anchorStart) {
    110         flags |= ANCHOR_START;
    111     }
    112     if (anchorEnd) {
    113         flags |= ANCHOR_END;
    114     }
    115 
    116     anteContext = NULL;
    117     if (anteContextLength > 0) {
    118         anteContext = new StringMatcher(pattern, 0, anteContextLength,
    119                                         FALSE, *data);
    120         /* test for NULL */
    121         if (anteContext == 0) {
    122             status = U_MEMORY_ALLOCATION_ERROR;
    123             return;
    124         }
    125     }
    126 
    127     key = NULL;
    128     if (keyLength > 0) {
    129         key = new StringMatcher(pattern, anteContextLength, anteContextLength + keyLength,
    130                                 FALSE, *data);
    131         /* test for NULL */
    132         if (key == 0) {
    133             status = U_MEMORY_ALLOCATION_ERROR;
    134             return;
    135         }
    136     }
    137 
    138     int32_t postContextLength = pattern.length() - keyLength - anteContextLength;
    139     postContext = NULL;
    140     if (postContextLength > 0) {
    141         postContext = new StringMatcher(pattern, anteContextLength + keyLength, pattern.length(),
    142                                         FALSE, *data);
    143         /* test for NULL */
    144         if (postContext == 0) {
    145             status = U_MEMORY_ALLOCATION_ERROR;
    146             return;
    147         }
    148     }
    149 
    150     this->output = new StringReplacer(outputStr, cursorPosition + cursorOffset, data);
    151     /* test for NULL */
    152     if (this->output == 0) {
    153         status = U_MEMORY_ALLOCATION_ERROR;
    154         return;
    155     }
    156 }
    157 
    158 /**
    159  * Copy constructor.
    160  */
    161 TransliterationRule::TransliterationRule(TransliterationRule& other) :
    162     UMemory(other),
    163     anteContext(NULL),
    164     key(NULL),
    165     postContext(NULL),
    166     pattern(other.pattern),
    167     anteContextLength(other.anteContextLength),
    168     keyLength(other.keyLength),
    169     flags(other.flags),
    170     data(other.data) {
    171 
    172     segments = NULL;
    173     segmentsCount = 0;
    174     if (other.segmentsCount > 0) {
    175         segments = (UnicodeFunctor **)uprv_malloc(other.segmentsCount * sizeof(UnicodeFunctor *));
    176         uprv_memcpy(segments, other.segments, other.segmentsCount*sizeof(segments[0]));
    177     }
    178 
    179     if (other.anteContext != NULL) {
    180         anteContext = (StringMatcher*) other.anteContext->clone();
    181     }
    182     if (other.key != NULL) {
    183         key = (StringMatcher*) other.key->clone();
    184     }
    185     if (other.postContext != NULL) {
    186         postContext = (StringMatcher*) other.postContext->clone();
    187     }
    188     output = other.output->clone();
    189 }
    190 
    191 TransliterationRule::~TransliterationRule() {
    192     uprv_free(segments);
    193     delete anteContext;
    194     delete key;
    195     delete postContext;
    196     delete output;
    197 }
    198 
    199 /**
    200  * Return the preceding context length.  This method is needed to
    201  * support the <code>Transliterator</code> method
    202  * <code>getMaximumContextLength()</code>.  Internally, this is
    203  * implemented as the anteContextLength, optionally plus one if
    204  * there is a start anchor.  The one character anchor gap is
    205  * needed to make repeated incremental transliteration with
    206  * anchors work.
    207  */
    208 int32_t TransliterationRule::getContextLength(void) const {
    209     return anteContextLength + ((flags & ANCHOR_START) ? 1 : 0);
    210 }
    211 
    212 /**
    213  * Internal method.  Returns 8-bit index value for this rule.
    214  * This is the low byte of the first character of the key,
    215  * unless the first character of the key is a set.  If it's a
    216  * set, or otherwise can match multiple keys, the index value is -1.
    217  */
    218 int16_t TransliterationRule::getIndexValue() const {
    219     if (anteContextLength == pattern.length()) {
    220         // A pattern with just ante context {such as foo)>bar} can
    221         // match any key.
    222         return -1;
    223     }
    224     UChar32 c = pattern.char32At(anteContextLength);
    225     return (int16_t)(data->lookupMatcher(c) == NULL ? (c & 0xFF) : -1);
    226 }
    227 
    228 /**
    229  * Internal method.  Returns true if this rule matches the given
    230  * index value.  The index value is an 8-bit integer, 0..255,
    231  * representing the low byte of the first character of the key.
    232  * It matches this rule if it matches the first character of the
    233  * key, or if the first character of the key is a set, and the set
    234  * contains any character with a low byte equal to the index
    235  * value.  If the rule contains only ante context, as in foo)>bar,
    236  * then it will match any key.
    237  */
    238 UBool TransliterationRule::matchesIndexValue(uint8_t v) const {
    239     // Delegate to the key, or if there is none, to the postContext.
    240     // If there is neither then we match any key; return true.
    241     UnicodeMatcher *m = (key != NULL) ? key : postContext;
    242     return (m != NULL) ? m->matchesIndexValue(v) : TRUE;
    243 }
    244 
    245 /**
    246  * Return true if this rule masks another rule.  If r1 masks r2 then
    247  * r1 matches any input string that r2 matches.  If r1 masks r2 and r2 masks
    248  * r1 then r1 == r2.  Examples: "a>x" masks "ab>y".  "a>x" masks "a[b]>y".
    249  * "[c]a>x" masks "[dc]a>y".
    250  */
    251 UBool TransliterationRule::masks(const TransliterationRule& r2) const {
    252     /* Rule r1 masks rule r2 if the string formed of the
    253      * antecontext, key, and postcontext overlaps in the following
    254      * way:
    255      *
    256      * r1:      aakkkpppp
    257      * r2:     aaakkkkkpppp
    258      *            ^
    259      *
    260      * The strings must be aligned at the first character of the
    261      * key.  The length of r1 to the left of the alignment point
    262      * must be <= the length of r2 to the left; ditto for the
    263      * right.  The characters of r1 must equal (or be a superset
    264      * of) the corresponding characters of r2.  The superset
    265      * operation should be performed to check for UnicodeSet
    266      * masking.
    267      *
    268      * Anchors:  Two patterns that differ only in anchors only
    269      * mask one another if they are exactly equal, and r2 has
    270      * all the anchors r1 has (optionally, plus some).  Here Y
    271      * means the row masks the column, N means it doesn't.
    272      *
    273      *         ab   ^ab    ab$  ^ab$
    274      *   ab    Y     Y     Y     Y
    275      *  ^ab    N     Y     N     Y
    276      *   ab$   N     N     Y     Y
    277      *  ^ab$   N     N     N     Y
    278      *
    279      * Post context: {a}b masks ab, but not vice versa, since {a}b
    280      * matches everything ab matches, and {a}b matches {|a|}b but ab
    281      * does not.  Pre context is different (a{b} does not align with
    282      * ab).
    283      */
    284 
    285     /* LIMITATION of the current mask algorithm: Some rule
    286      * maskings are currently not detected.  For example,
    287      * "{Lu}]a>x" masks "A]a>y".  This can be added later. TODO
    288      */
    289 
    290     int32_t len = pattern.length();
    291     int32_t left = anteContextLength;
    292     int32_t left2 = r2.anteContextLength;
    293     int32_t right = len - left;
    294     int32_t right2 = r2.pattern.length() - left2;
    295     int32_t cachedCompare = r2.pattern.compare(left2 - left, len, pattern);
    296 
    297     // TODO Clean this up -- some logic might be combinable with the
    298     // next statement.
    299 
    300     // Test for anchor masking
    301     if (left == left2 && right == right2 &&
    302         keyLength <= r2.keyLength &&
    303         0 == cachedCompare) {
    304         // The following boolean logic implements the table above
    305         return (flags == r2.flags) ||
    306             (!(flags & ANCHOR_START) && !(flags & ANCHOR_END)) ||
    307             ((r2.flags & ANCHOR_START) && (r2.flags & ANCHOR_END));
    308     }
    309 
    310     return left <= left2 &&
    311         (right < right2 ||
    312          (right == right2 && keyLength <= r2.keyLength)) &&
    313          (0 == cachedCompare);
    314 }
    315 
    316 static inline int32_t posBefore(const Replaceable& str, int32_t pos) {
    317     return (pos > 0) ?
    318         pos - UTF_CHAR_LENGTH(str.char32At(pos-1)) :
    319         pos - 1;
    320 }
    321 
    322 static inline int32_t posAfter(const Replaceable& str, int32_t pos) {
    323     return (pos >= 0 && pos < str.length()) ?
    324         pos + UTF_CHAR_LENGTH(str.char32At(pos)) :
    325         pos + 1;
    326 }
    327 
    328 /**
    329  * Attempt a match and replacement at the given position.  Return
    330  * the degree of match between this rule and the given text.  The
    331  * degree of match may be mismatch, a partial match, or a full
    332  * match.  A mismatch means at least one character of the text
    333  * does not match the context or key.  A partial match means some
    334  * context and key characters match, but the text is not long
    335  * enough to match all of them.  A full match means all context
    336  * and key characters match.
    337  *
    338  * If a full match is obtained, perform a replacement, update pos,
    339  * and return U_MATCH.  Otherwise both text and pos are unchanged.
    340  *
    341  * @param text the text
    342  * @param pos the position indices
    343  * @param incremental if TRUE, test for partial matches that may
    344  * be completed by additional text inserted at pos.limit.
    345  * @return one of <code>U_MISMATCH</code>,
    346  * <code>U_PARTIAL_MATCH</code>, or <code>U_MATCH</code>.  If
    347  * incremental is FALSE then U_PARTIAL_MATCH will not be returned.
    348  */
    349 UMatchDegree TransliterationRule::matchAndReplace(Replaceable& text,
    350                                                   UTransPosition& pos,
    351                                                   UBool incremental) const {
    352     // Matching and replacing are done in one method because the
    353     // replacement operation needs information obtained during the
    354     // match.  Another way to do this is to have the match method
    355     // create a match result struct with relevant offsets, and to pass
    356     // this into the replace method.
    357 
    358     // ============================ MATCH ===========================
    359 
    360     // Reset segment match data
    361     if (segments != NULL) {
    362         for (int32_t i=0; i<segmentsCount; ++i) {
    363             ((StringMatcher*) segments[i])->resetMatch();
    364         }
    365     }
    366 
    367 //    int32_t lenDelta, keyLimit;
    368     int32_t keyLimit;
    369 
    370     // ------------------------ Ante Context ------------------------
    371 
    372     // A mismatch in the ante context, or with the start anchor,
    373     // is an outright U_MISMATCH regardless of whether we are
    374     // incremental or not.
    375     int32_t oText; // offset into 'text'
    376 //    int32_t newStart = 0;
    377     int32_t minOText;
    378 
    379     // Note (1): We process text in 16-bit code units, rather than
    380     // 32-bit code points.  This works because stand-ins are
    381     // always in the BMP and because we are doing a literal match
    382     // operation, which can be done 16-bits at a time.
    383 
    384     int32_t anteLimit = posBefore(text, pos.contextStart);
    385 
    386     UMatchDegree match;
    387 
    388     // Start reverse match at char before pos.start
    389     oText = posBefore(text, pos.start);
    390 
    391     if (anteContext != NULL) {
    392         match = anteContext->matches(text, oText, anteLimit, FALSE);
    393         if (match != U_MATCH) {
    394             return U_MISMATCH;
    395         }
    396     }
    397 
    398     minOText = posAfter(text, oText);
    399 
    400     // ------------------------ Start Anchor ------------------------
    401 
    402     if (((flags & ANCHOR_START) != 0) && oText != anteLimit) {
    403         return U_MISMATCH;
    404     }
    405 
    406     // -------------------- Key and Post Context --------------------
    407 
    408     oText = pos.start;
    409 
    410     if (key != NULL) {
    411         match = key->matches(text, oText, pos.limit, incremental);
    412         if (match != U_MATCH) {
    413             return match;
    414         }
    415     }
    416 
    417     keyLimit = oText;
    418 
    419     if (postContext != NULL) {
    420         if (incremental && keyLimit == pos.limit) {
    421             // The key matches just before pos.limit, and there is
    422             // a postContext.  Since we are in incremental mode,
    423             // we must assume more characters may be inserted at
    424             // pos.limit -- this is a partial match.
    425             return U_PARTIAL_MATCH;
    426         }
    427 
    428         match = postContext->matches(text, oText, pos.contextLimit, incremental);
    429         if (match != U_MATCH) {
    430             return match;
    431         }
    432     }
    433 
    434     // ------------------------- Stop Anchor ------------------------
    435 
    436     if (((flags & ANCHOR_END)) != 0) {
    437         if (oText != pos.contextLimit) {
    438             return U_MISMATCH;
    439         }
    440         if (incremental) {
    441             return U_PARTIAL_MATCH;
    442         }
    443     }
    444 
    445     // =========================== REPLACE ==========================
    446 
    447     // We have a full match.  The key is between pos.start and
    448     // keyLimit.
    449 
    450     int32_t newStart;
    451     int32_t newLength = output->toReplacer()->replace(text, pos.start, keyLimit, newStart);
    452     int32_t lenDelta = newLength - (keyLimit - pos.start);
    453 
    454     oText += lenDelta;
    455     pos.limit += lenDelta;
    456     pos.contextLimit += lenDelta;
    457     // Restrict new value of start to [minOText, min(oText, pos.limit)].
    458     pos.start = uprv_max(minOText, uprv_min(uprv_min(oText, pos.limit), newStart));
    459     return U_MATCH;
    460 }
    461 
    462 /**
    463  * Create a source string that represents this rule.  Append it to the
    464  * given string.
    465  */
    466 UnicodeString& TransliterationRule::toRule(UnicodeString& rule,
    467                                            UBool escapeUnprintable) const {
    468 
    469     // Accumulate special characters (and non-specials following them)
    470     // into quoteBuf.  Append quoteBuf, within single quotes, when
    471     // a non-quoted element must be inserted.
    472     UnicodeString str, quoteBuf;
    473 
    474     // Do not emit the braces '{' '}' around the pattern if there
    475     // is neither anteContext nor postContext.
    476     UBool emitBraces =
    477         (anteContext != NULL) || (postContext != NULL);
    478 
    479     // Emit start anchor
    480     if ((flags & ANCHOR_START) != 0) {
    481         rule.append((UChar)94/*^*/);
    482     }
    483 
    484     // Emit the input pattern
    485     ICU_Utility::appendToRule(rule, anteContext, escapeUnprintable, quoteBuf);
    486 
    487     if (emitBraces) {
    488         ICU_Utility::appendToRule(rule, (UChar) 0x007B /*{*/, TRUE, escapeUnprintable, quoteBuf);
    489     }
    490 
    491     ICU_Utility::appendToRule(rule, key, escapeUnprintable, quoteBuf);
    492 
    493     if (emitBraces) {
    494         ICU_Utility::appendToRule(rule, (UChar) 0x007D /*}*/, TRUE, escapeUnprintable, quoteBuf);
    495     }
    496 
    497     ICU_Utility::appendToRule(rule, postContext, escapeUnprintable, quoteBuf);
    498 
    499     // Emit end anchor
    500     if ((flags & ANCHOR_END) != 0) {
    501         rule.append((UChar)36/*$*/);
    502     }
    503 
    504     ICU_Utility::appendToRule(rule, FORWARD_OP, TRUE, escapeUnprintable, quoteBuf);
    505 
    506     // Emit the output pattern
    507 
    508     ICU_Utility::appendToRule(rule, output->toReplacer()->toReplacerPattern(str, escapeUnprintable),
    509                               TRUE, escapeUnprintable, quoteBuf);
    510 
    511     ICU_Utility::appendToRule(rule, (UChar) 0x003B /*;*/, TRUE, escapeUnprintable, quoteBuf);
    512 
    513     return rule;
    514 }
    515 
    516 void TransliterationRule::setData(const TransliterationRuleData* d) {
    517     data = d;
    518     if (anteContext != NULL) anteContext->setData(d);
    519     if (postContext != NULL) postContext->setData(d);
    520     if (key != NULL) key->setData(d);
    521     // assert(output != NULL);
    522     output->setData(d);
    523     // Don't have to do segments since they are in the context or key
    524 }
    525 
    526 /**
    527  * Union the set of all characters that may be modified by this rule
    528  * into the given set.
    529  */
    530 void TransliterationRule::addSourceSetTo(UnicodeSet& toUnionTo) const {
    531     int32_t limit = anteContextLength + keyLength;
    532     for (int32_t i=anteContextLength; i<limit; ) {
    533         UChar32 ch = pattern.char32At(i);
    534         i += UTF_CHAR_LENGTH(ch);
    535         const UnicodeMatcher* matcher = data->lookupMatcher(ch);
    536         if (matcher == NULL) {
    537             toUnionTo.add(ch);
    538         } else {
    539             matcher->addMatchSetTo(toUnionTo);
    540         }
    541     }
    542 }
    543 
    544 /**
    545  * Union the set of all characters that may be emitted by this rule
    546  * into the given set.
    547  */
    548 void TransliterationRule::addTargetSetTo(UnicodeSet& toUnionTo) const {
    549     output->toReplacer()->addReplacementSetTo(toUnionTo);
    550 }
    551 
    552 U_NAMESPACE_END
    553 
    554 #endif /* #if !UCONFIG_NO_TRANSLITERATION */
    555 
    556 //eof
    557