• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1 // © 2016 and later: Unicode, Inc. and others.
2 // License & terms of use: http://www.unicode.org/copyright.html
3 /*
4  **********************************************************************
5  *   Copyright (C) 1999-2011, International Business Machines
6  *   Corporation and others.  All Rights Reserved.
7  **********************************************************************
8  *   Date        Name        Description
9  *   11/17/99    aliu        Creation.
10  **********************************************************************
11  */
12 
13 #include "unicode/utypes.h"
14 
15 #if !UCONFIG_NO_TRANSLITERATION
16 
17 #include "unicode/rep.h"
18 #include "unicode/unifilt.h"
19 #include "unicode/uniset.h"
20 #include "unicode/utf16.h"
21 #include "rbt_rule.h"
22 #include "rbt_data.h"
23 #include "cmemory.h"
24 #include "strmatch.h"
25 #include "strrepl.h"
26 #include "util.h"
27 #include "putilimp.h"
28 
29 static const UChar FORWARD_OP[] = {32,62,32,0}; // " > "
30 
31 U_NAMESPACE_BEGIN
32 
33 /**
34  * Construct a new rule with the given input, output text, and other
35  * attributes.  A cursor position may be specified for the output text.
36  * @param input input string, including key and optional ante and
37  * post context
38  * @param anteContextPos offset into input to end of ante context, or -1 if
39  * none.  Must be <= input.length() if not -1.
40  * @param postContextPos offset into input to start of post context, or -1
41  * if none.  Must be <= input.length() if not -1, and must be >=
42  * anteContextPos.
43  * @param output output string
44  * @param cursorPosition offset into output at which cursor is located, or -1 if
45  * none.  If less than zero, then the cursor is placed after the
46  * <code>output</code>; that is, -1 is equivalent to
47  * <code>output.length()</code>.  If greater than
48  * <code>output.length()</code> then an exception is thrown.
49  * @param segs array of UnicodeFunctors corresponding to input pattern
50  * segments, or null if there are none.  The array itself is adopted,
51  * but the pointers within it are not.
52  * @param segsCount number of elements in segs[]
53  * @param anchorStart TRUE if the the rule is anchored on the left to
54  * the context start
55  * @param anchorEnd TRUE if the rule is anchored on the right to the
56  * context limit
57  */
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)58 TransliterationRule::TransliterationRule(const UnicodeString& input,
59                                          int32_t anteContextPos, int32_t postContextPos,
60                                          const UnicodeString& outputStr,
61                                          int32_t cursorPosition, int32_t cursorOffset,
62                                          UnicodeFunctor** segs,
63                                          int32_t segsCount,
64                                          UBool anchorStart, UBool anchorEnd,
65                                          const TransliterationRuleData* theData,
66                                          UErrorCode& status) :
67     UMemory(),
68     segments(0),
69     data(theData) {
70 
71     if (U_FAILURE(status)) {
72         return;
73     }
74     // Do range checks only when warranted to save time
75     if (anteContextPos < 0) {
76         anteContextLength = 0;
77     } else {
78         if (anteContextPos > input.length()) {
79             // throw new IllegalArgumentException("Invalid ante context");
80             status = U_ILLEGAL_ARGUMENT_ERROR;
81             return;
82         }
83         anteContextLength = anteContextPos;
84     }
85     if (postContextPos < 0) {
86         keyLength = input.length() - anteContextLength;
87     } else {
88         if (postContextPos < anteContextLength ||
89             postContextPos > input.length()) {
90             // throw new IllegalArgumentException("Invalid post context");
91             status = U_ILLEGAL_ARGUMENT_ERROR;
92             return;
93         }
94         keyLength = postContextPos - anteContextLength;
95     }
96     if (cursorPosition < 0) {
97         cursorPosition = outputStr.length();
98     } else if (cursorPosition > outputStr.length()) {
99         // throw new IllegalArgumentException("Invalid cursor position");
100         status = U_ILLEGAL_ARGUMENT_ERROR;
101         return;
102     }
103     // We don't validate the segments array.  The caller must
104     // guarantee that the segments are well-formed (that is, that
105     // all $n references in the output refer to indices of this
106     // array, and that no array elements are null).
107     this->segments = segs;
108     this->segmentsCount = segsCount;
109 
110     pattern = input;
111     flags = 0;
112     if (anchorStart) {
113         flags |= ANCHOR_START;
114     }
115     if (anchorEnd) {
116         flags |= ANCHOR_END;
117     }
118 
119     anteContext = NULL;
120     if (anteContextLength > 0) {
121         anteContext = new StringMatcher(pattern, 0, anteContextLength,
122                                         FALSE, *data);
123         /* test for NULL */
124         if (anteContext == 0) {
125             status = U_MEMORY_ALLOCATION_ERROR;
126             return;
127         }
128     }
129 
130     key = NULL;
131     if (keyLength > 0) {
132         key = new StringMatcher(pattern, anteContextLength, anteContextLength + keyLength,
133                                 FALSE, *data);
134         /* test for NULL */
135         if (key == 0) {
136             status = U_MEMORY_ALLOCATION_ERROR;
137             return;
138         }
139     }
140 
141     int32_t postContextLength = pattern.length() - keyLength - anteContextLength;
142     postContext = NULL;
143     if (postContextLength > 0) {
144         postContext = new StringMatcher(pattern, anteContextLength + keyLength, pattern.length(),
145                                         FALSE, *data);
146         /* test for NULL */
147         if (postContext == 0) {
148             status = U_MEMORY_ALLOCATION_ERROR;
149             return;
150         }
151     }
152 
153     this->output = new StringReplacer(outputStr, cursorPosition + cursorOffset, data);
154     /* test for NULL */
155     if (this->output == 0) {
156         status = U_MEMORY_ALLOCATION_ERROR;
157         return;
158     }
159 }
160 
161 /**
162  * Copy constructor.
163  */
TransliterationRule(TransliterationRule & other)164 TransliterationRule::TransliterationRule(TransliterationRule& other) :
165     UMemory(other),
166     anteContext(NULL),
167     key(NULL),
168     postContext(NULL),
169     pattern(other.pattern),
170     anteContextLength(other.anteContextLength),
171     keyLength(other.keyLength),
172     flags(other.flags),
173     data(other.data) {
174 
175     segments = NULL;
176     segmentsCount = 0;
177     if (other.segmentsCount > 0) {
178         segments = (UnicodeFunctor **)uprv_malloc(other.segmentsCount * sizeof(UnicodeFunctor *));
179         uprv_memcpy(segments, other.segments, (size_t)other.segmentsCount*sizeof(segments[0]));
180     }
181 
182     if (other.anteContext != NULL) {
183         anteContext = other.anteContext->clone();
184     }
185     if (other.key != NULL) {
186         key = other.key->clone();
187     }
188     if (other.postContext != NULL) {
189         postContext = other.postContext->clone();
190     }
191     output = other.output->clone();
192 }
193 
~TransliterationRule()194 TransliterationRule::~TransliterationRule() {
195     uprv_free(segments);
196     delete anteContext;
197     delete key;
198     delete postContext;
199     delete output;
200 }
201 
202 /**
203  * Return the preceding context length.  This method is needed to
204  * support the <code>Transliterator</code> method
205  * <code>getMaximumContextLength()</code>.  Internally, this is
206  * implemented as the anteContextLength, optionally plus one if
207  * there is a start anchor.  The one character anchor gap is
208  * needed to make repeated incremental transliteration with
209  * anchors work.
210  */
getContextLength(void) const211 int32_t TransliterationRule::getContextLength(void) const {
212     return anteContextLength + ((flags & ANCHOR_START) ? 1 : 0);
213 }
214 
215 /**
216  * Internal method.  Returns 8-bit index value for this rule.
217  * This is the low byte of the first character of the key,
218  * unless the first character of the key is a set.  If it's a
219  * set, or otherwise can match multiple keys, the index value is -1.
220  */
getIndexValue() const221 int16_t TransliterationRule::getIndexValue() const {
222     if (anteContextLength == pattern.length()) {
223         // A pattern with just ante context {such as foo)>bar} can
224         // match any key.
225         return -1;
226     }
227     UChar32 c = pattern.char32At(anteContextLength);
228     return (int16_t)(data->lookupMatcher(c) == NULL ? (c & 0xFF) : -1);
229 }
230 
231 /**
232  * Internal method.  Returns true if this rule matches the given
233  * index value.  The index value is an 8-bit integer, 0..255,
234  * representing the low byte of the first character of the key.
235  * It matches this rule if it matches the first character of the
236  * key, or if the first character of the key is a set, and the set
237  * contains any character with a low byte equal to the index
238  * value.  If the rule contains only ante context, as in foo)>bar,
239  * then it will match any key.
240  */
matchesIndexValue(uint8_t v) const241 UBool TransliterationRule::matchesIndexValue(uint8_t v) const {
242     // Delegate to the key, or if there is none, to the postContext.
243     // If there is neither then we match any key; return true.
244     UnicodeMatcher *m = (key != NULL) ? key : postContext;
245     return (m != NULL) ? m->matchesIndexValue(v) : TRUE;
246 }
247 
248 /**
249  * Return true if this rule masks another rule.  If r1 masks r2 then
250  * r1 matches any input string that r2 matches.  If r1 masks r2 and r2 masks
251  * r1 then r1 == r2.  Examples: "a>x" masks "ab>y".  "a>x" masks "a[b]>y".
252  * "[c]a>x" masks "[dc]a>y".
253  */
masks(const TransliterationRule & r2) const254 UBool TransliterationRule::masks(const TransliterationRule& r2) const {
255     /* Rule r1 masks rule r2 if the string formed of the
256      * antecontext, key, and postcontext overlaps in the following
257      * way:
258      *
259      * r1:      aakkkpppp
260      * r2:     aaakkkkkpppp
261      *            ^
262      *
263      * The strings must be aligned at the first character of the
264      * key.  The length of r1 to the left of the alignment point
265      * must be <= the length of r2 to the left; ditto for the
266      * right.  The characters of r1 must equal (or be a superset
267      * of) the corresponding characters of r2.  The superset
268      * operation should be performed to check for UnicodeSet
269      * masking.
270      *
271      * Anchors:  Two patterns that differ only in anchors only
272      * mask one another if they are exactly equal, and r2 has
273      * all the anchors r1 has (optionally, plus some).  Here Y
274      * means the row masks the column, N means it doesn't.
275      *
276      *         ab   ^ab    ab$  ^ab$
277      *   ab    Y     Y     Y     Y
278      *  ^ab    N     Y     N     Y
279      *   ab$   N     N     Y     Y
280      *  ^ab$   N     N     N     Y
281      *
282      * Post context: {a}b masks ab, but not vice versa, since {a}b
283      * matches everything ab matches, and {a}b matches {|a|}b but ab
284      * does not.  Pre context is different (a{b} does not align with
285      * ab).
286      */
287 
288     /* LIMITATION of the current mask algorithm: Some rule
289      * maskings are currently not detected.  For example,
290      * "{Lu}]a>x" masks "A]a>y".  This can be added later. TODO
291      */
292 
293     int32_t len = pattern.length();
294     int32_t left = anteContextLength;
295     int32_t left2 = r2.anteContextLength;
296     int32_t right = len - left;
297     int32_t right2 = r2.pattern.length() - left2;
298     int32_t cachedCompare = r2.pattern.compare(left2 - left, len, pattern);
299 
300     // TODO Clean this up -- some logic might be combinable with the
301     // next statement.
302 
303     // Test for anchor masking
304     if (left == left2 && right == right2 &&
305         keyLength <= r2.keyLength &&
306         0 == cachedCompare) {
307         // The following boolean logic implements the table above
308         return (flags == r2.flags) ||
309             (!(flags & ANCHOR_START) && !(flags & ANCHOR_END)) ||
310             ((r2.flags & ANCHOR_START) && (r2.flags & ANCHOR_END));
311     }
312 
313     return left <= left2 &&
314         (right < right2 ||
315          (right == right2 && keyLength <= r2.keyLength)) &&
316          (0 == cachedCompare);
317 }
318 
posBefore(const Replaceable & str,int32_t pos)319 static inline int32_t posBefore(const Replaceable& str, int32_t pos) {
320     return (pos > 0) ?
321         pos - U16_LENGTH(str.char32At(pos-1)) :
322         pos - 1;
323 }
324 
posAfter(const Replaceable & str,int32_t pos)325 static inline int32_t posAfter(const Replaceable& str, int32_t pos) {
326     return (pos >= 0 && pos < str.length()) ?
327         pos + U16_LENGTH(str.char32At(pos)) :
328         pos + 1;
329 }
330 
331 /**
332  * Attempt a match and replacement at the given position.  Return
333  * the degree of match between this rule and the given text.  The
334  * degree of match may be mismatch, a partial match, or a full
335  * match.  A mismatch means at least one character of the text
336  * does not match the context or key.  A partial match means some
337  * context and key characters match, but the text is not long
338  * enough to match all of them.  A full match means all context
339  * and key characters match.
340  *
341  * If a full match is obtained, perform a replacement, update pos,
342  * and return U_MATCH.  Otherwise both text and pos are unchanged.
343  *
344  * @param text the text
345  * @param pos the position indices
346  * @param incremental if TRUE, test for partial matches that may
347  * be completed by additional text inserted at pos.limit.
348  * @return one of <code>U_MISMATCH</code>,
349  * <code>U_PARTIAL_MATCH</code>, or <code>U_MATCH</code>.  If
350  * incremental is FALSE then U_PARTIAL_MATCH will not be returned.
351  */
matchAndReplace(Replaceable & text,UTransPosition & pos,UBool incremental) const352 UMatchDegree TransliterationRule::matchAndReplace(Replaceable& text,
353                                                   UTransPosition& pos,
354                                                   UBool incremental) const {
355     // Matching and replacing are done in one method because the
356     // replacement operation needs information obtained during the
357     // match.  Another way to do this is to have the match method
358     // create a match result struct with relevant offsets, and to pass
359     // this into the replace method.
360 
361     // ============================ MATCH ===========================
362 
363     // Reset segment match data
364     if (segments != NULL) {
365         for (int32_t i=0; i<segmentsCount; ++i) {
366             ((StringMatcher*) segments[i])->resetMatch();
367         }
368     }
369 
370 //    int32_t lenDelta, keyLimit;
371     int32_t keyLimit;
372 
373     // ------------------------ Ante Context ------------------------
374 
375     // A mismatch in the ante context, or with the start anchor,
376     // is an outright U_MISMATCH regardless of whether we are
377     // incremental or not.
378     int32_t oText; // offset into 'text'
379 //    int32_t newStart = 0;
380     int32_t minOText;
381 
382     // Note (1): We process text in 16-bit code units, rather than
383     // 32-bit code points.  This works because stand-ins are
384     // always in the BMP and because we are doing a literal match
385     // operation, which can be done 16-bits at a time.
386 
387     int32_t anteLimit = posBefore(text, pos.contextStart);
388 
389     UMatchDegree match;
390 
391     // Start reverse match at char before pos.start
392     oText = posBefore(text, pos.start);
393 
394     if (anteContext != NULL) {
395         match = anteContext->matches(text, oText, anteLimit, FALSE);
396         if (match != U_MATCH) {
397             return U_MISMATCH;
398         }
399     }
400 
401     minOText = posAfter(text, oText);
402 
403     // ------------------------ Start Anchor ------------------------
404 
405     if (((flags & ANCHOR_START) != 0) && oText != anteLimit) {
406         return U_MISMATCH;
407     }
408 
409     // -------------------- Key and Post Context --------------------
410 
411     oText = pos.start;
412 
413     if (key != NULL) {
414         match = key->matches(text, oText, pos.limit, incremental);
415         if (match != U_MATCH) {
416             return match;
417         }
418     }
419 
420     keyLimit = oText;
421 
422     if (postContext != NULL) {
423         if (incremental && keyLimit == pos.limit) {
424             // The key matches just before pos.limit, and there is
425             // a postContext.  Since we are in incremental mode,
426             // we must assume more characters may be inserted at
427             // pos.limit -- this is a partial match.
428             return U_PARTIAL_MATCH;
429         }
430 
431         match = postContext->matches(text, oText, pos.contextLimit, incremental);
432         if (match != U_MATCH) {
433             return match;
434         }
435     }
436 
437     // ------------------------- Stop Anchor ------------------------
438 
439     if (((flags & ANCHOR_END)) != 0) {
440         if (oText != pos.contextLimit) {
441             return U_MISMATCH;
442         }
443         if (incremental) {
444             return U_PARTIAL_MATCH;
445         }
446     }
447 
448     // =========================== REPLACE ==========================
449 
450     // We have a full match.  The key is between pos.start and
451     // keyLimit.
452 
453     int32_t newStart;
454     int32_t newLength = output->toReplacer()->replace(text, pos.start, keyLimit, newStart);
455     int32_t lenDelta = newLength - (keyLimit - pos.start);
456 
457     oText += lenDelta;
458     pos.limit += lenDelta;
459     pos.contextLimit += lenDelta;
460     // Restrict new value of start to [minOText, min(oText, pos.limit)].
461     pos.start = uprv_max(minOText, uprv_min(uprv_min(oText, pos.limit), newStart));
462     return U_MATCH;
463 }
464 
465 /**
466  * Create a source string that represents this rule.  Append it to the
467  * given string.
468  */
toRule(UnicodeString & rule,UBool escapeUnprintable) const469 UnicodeString& TransliterationRule::toRule(UnicodeString& rule,
470                                            UBool escapeUnprintable) const {
471 
472     // Accumulate special characters (and non-specials following them)
473     // into quoteBuf.  Append quoteBuf, within single quotes, when
474     // a non-quoted element must be inserted.
475     UnicodeString str, quoteBuf;
476 
477     // Do not emit the braces '{' '}' around the pattern if there
478     // is neither anteContext nor postContext.
479     UBool emitBraces =
480         (anteContext != NULL) || (postContext != NULL);
481 
482     // Emit start anchor
483     if ((flags & ANCHOR_START) != 0) {
484         rule.append((UChar)94/*^*/);
485     }
486 
487     // Emit the input pattern
488     ICU_Utility::appendToRule(rule, anteContext, escapeUnprintable, quoteBuf);
489 
490     if (emitBraces) {
491         ICU_Utility::appendToRule(rule, (UChar) 0x007B /*{*/, TRUE, escapeUnprintable, quoteBuf);
492     }
493 
494     ICU_Utility::appendToRule(rule, key, escapeUnprintable, quoteBuf);
495 
496     if (emitBraces) {
497         ICU_Utility::appendToRule(rule, (UChar) 0x007D /*}*/, TRUE, escapeUnprintable, quoteBuf);
498     }
499 
500     ICU_Utility::appendToRule(rule, postContext, escapeUnprintable, quoteBuf);
501 
502     // Emit end anchor
503     if ((flags & ANCHOR_END) != 0) {
504         rule.append((UChar)36/*$*/);
505     }
506 
507     ICU_Utility::appendToRule(rule, UnicodeString(TRUE, FORWARD_OP, 3), TRUE, escapeUnprintable, quoteBuf);
508 
509     // Emit the output pattern
510 
511     ICU_Utility::appendToRule(rule, output->toReplacer()->toReplacerPattern(str, escapeUnprintable),
512                               TRUE, escapeUnprintable, quoteBuf);
513 
514     ICU_Utility::appendToRule(rule, (UChar) 0x003B /*;*/, TRUE, escapeUnprintable, quoteBuf);
515 
516     return rule;
517 }
518 
setData(const TransliterationRuleData * d)519 void TransliterationRule::setData(const TransliterationRuleData* d) {
520     data = d;
521     if (anteContext != NULL) anteContext->setData(d);
522     if (postContext != NULL) postContext->setData(d);
523     if (key != NULL) key->setData(d);
524     // assert(output != NULL);
525     output->setData(d);
526     // Don't have to do segments since they are in the context or key
527 }
528 
529 /**
530  * Union the set of all characters that may be modified by this rule
531  * into the given set.
532  */
addSourceSetTo(UnicodeSet & toUnionTo) const533 void TransliterationRule::addSourceSetTo(UnicodeSet& toUnionTo) const {
534     int32_t limit = anteContextLength + keyLength;
535     for (int32_t i=anteContextLength; i<limit; ) {
536         UChar32 ch = pattern.char32At(i);
537         i += U16_LENGTH(ch);
538         const UnicodeMatcher* matcher = data->lookupMatcher(ch);
539         if (matcher == NULL) {
540             toUnionTo.add(ch);
541         } else {
542             matcher->addMatchSetTo(toUnionTo);
543         }
544     }
545 }
546 
547 /**
548  * Union the set of all characters that may be emitted by this rule
549  * into the given set.
550  */
addTargetSetTo(UnicodeSet & toUnionTo) const551 void TransliterationRule::addTargetSetTo(UnicodeSet& toUnionTo) const {
552     output->toReplacer()->addReplacementSetTo(toUnionTo);
553 }
554 
555 U_NAMESPACE_END
556 
557 #endif /* #if !UCONFIG_NO_TRANSLITERATION */
558 
559 //eof
560