• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
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  */
TransliterationRule(const UnicodeString & input,int32_t anteContextPos,int32_t postContextPos,const UnicodeString & outputStr,int32_t cursorPosition,int32_t cursorOffset,UnicodeFunctor ** segs,int32_t segsCount,UBool anchorStart,UBool anchorEnd,const TransliterationRuleData * theData,UErrorCode & status)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  */
TransliterationRule(TransliterationRule & other)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 
~TransliterationRule()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  */
getContextLength(void) const208 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  */
getIndexValue() const218 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  */
matchesIndexValue(uint8_t v) const238 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  */
masks(const TransliterationRule & r2) const251 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 
posBefore(const Replaceable & str,int32_t pos)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 
posAfter(const Replaceable & str,int32_t pos)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  */
matchAndReplace(Replaceable & text,UTransPosition & pos,UBool incremental) const349 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  */
toRule(UnicodeString & rule,UBool escapeUnprintable) const466 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 
setData(const TransliterationRuleData * d)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  */
addSourceSetTo(UnicodeSet & toUnionTo) const530 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  */
addTargetSetTo(UnicodeSet & toUnionTo) const548 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