1 // © 2016 and later: Unicode, Inc. and others.
2 // License & terms of use: http://www.unicode.org/copyright.html#License
3 /*
4  *******************************************************************************
5  * Copyright (C) 1996-2010, International Business Machines Corporation and    *
6  * others. All Rights Reserved.                                                *
7  *******************************************************************************
8  */
9 package com.ibm.icu.text;
10 
11 import com.ibm.icu.impl.Utility;
12 
13 /**
14  * A transliteration rule used by
15  * <code>RuleBasedTransliterator</code>.
16  * <code>TransliterationRule</code> is an immutable object.
17  *
18  * <p>A rule consists of an input pattern and an output string.  When
19  * the input pattern is matched, the output string is emitted.  The
20  * input pattern consists of zero or more characters which are matched
21  * exactly (the key) and optional context.  Context must match if it
22  * is specified.  Context may be specified before the key, after the
23  * key, or both.  The key, preceding context, and following context
24  * may contain variables.  Variables represent a set of Unicode
25  * characters, such as the letters <i>a</i> through <i>z</i>.
26  * Variables are detected by looking up each character in a supplied
27  * variable list to see if it has been so defined.
28  *
29  * <p>A rule may contain segments in its input string and segment
30  * references in its output string.  A segment is a substring of the
31  * input pattern, indicated by an offset and limit.  The segment may
32  * be in the preceding or following context.  It may not span a
33  * context boundary.  A segment reference is a special character in
34  * the output string that causes a segment of the input string (not
35  * the input pattern) to be copied to the output string.  The range of
36  * special characters that represent segment references is defined by
37  * RuleBasedTransliterator.Data.
38  *
39  * <p>Example: The rule "([a-z]) . ([0-9]) > $2 . $1" will change the input
40  * string "abc.123" to "ab1.c23".
41  *
42  * <p>Copyright &copy; IBM Corporation 1999.  All rights reserved.
43  *
44  * @author Alan Liu
45  */
46 class TransliterationRule {
47 
48     // TODO Eliminate the pattern and keyLength data members.  They
49     // are used only by masks() and getIndexValue() which are called
50     // only during build time, not during run-time.  Perhaps these
51     // methods and pattern/keyLength can be isolated into a separate
52     // object.
53 
54     /**
55      * The match that must occur before the key, or null if there is no
56      * preceding context.
57      */
58     private StringMatcher anteContext;
59 
60     /**
61      * The matcher object for the key.  If null, then the key is empty.
62      */
63     private StringMatcher key;
64 
65     /**
66      * The match that must occur after the key, or null if there is no
67      * following context.
68      */
69     private StringMatcher postContext;
70 
71     /**
72      * The object that performs the replacement if the key,
73      * anteContext, and postContext are matched.  Never null.
74      */
75     private UnicodeReplacer output;
76 
77     /**
78      * The string that must be matched, consisting of the anteContext, key,
79      * and postContext, concatenated together, in that order.  Some components
80      * may be empty (zero length).
81      * @see anteContextLength
82      * @see keyLength
83      */
84     private String pattern;
85 
86     /**
87      * An array of matcher objects corresponding to the input pattern
88      * segments.  If there are no segments this is null.  N.B. This is
89      * a UnicodeMatcher for generality, but in practice it is always a
90      * StringMatcher.  In the future we may generalize this, but for
91      * now we sometimes cast down to StringMatcher.
92      */
93     UnicodeMatcher[] segments;
94 
95     /**
96      * The length of the string that must match before the key.  If
97      * zero, then there is no matching requirement before the key.
98      * Substring [0,anteContextLength) of pattern is the anteContext.
99      */
100     private int anteContextLength;
101 
102     /**
103      * The length of the key.  Substring [anteContextLength,
104      * anteContextLength + keyLength) is the key.
105      */
106     private int keyLength;
107 
108     /**
109      * Miscellaneous attributes.
110      */
111     byte flags;
112 
113     /**
114      * Flag attributes.
115      */
116     static final int ANCHOR_START = 1;
117     static final int ANCHOR_END   = 2;
118 
119     /**
120      * An alias pointer to the data for this rule.  The data provides
121      * lookup services for matchers and segments.
122      */
123     private final RuleBasedTransliterator.Data data;
124 
125 
126     /**
127      * Construct a new rule with the given input, output text, and other
128      * attributes.  A cursor position may be specified for the output text.
129      * @param input input string, including key and optional ante and
130      * post context
131      * @param anteContextPos offset into input to end of ante context, or -1 if
132      * none.  Must be <= input.length() if not -1.
133      * @param postContextPos offset into input to start of post context, or -1
134      * if none.  Must be <= input.length() if not -1, and must be >=
135      * anteContextPos.
136      * @param output output string
137      * @param cursorPos offset into output at which cursor is located, or -1 if
138      * none.  If less than zero, then the cursor is placed after the
139      * <code>output</code>; that is, -1 is equivalent to
140      * <code>output.length()</code>.  If greater than
141      * <code>output.length()</code> then an exception is thrown.
142      * @param cursorOffset an offset to be added to cursorPos to position the
143      * cursor either in the ante context, if < 0, or in the post context, if >
144      * 0.  For example, the rule "abc{def} > | @@@ xyz;" changes "def" to
145      * "xyz" and moves the cursor to before "a".  It would have a cursorOffset
146      * of -3.
147      * @param segs array of UnicodeMatcher corresponding to input pattern
148      * segments, or null if there are none
149      * @param anchorStart true if the the rule is anchored on the left to
150      * the context start
151      * @param anchorEnd true if the rule is anchored on the right to the
152      * context limit
153      */
TransliterationRule(String input, int anteContextPos, int postContextPos, String output, int cursorPos, int cursorOffset, UnicodeMatcher[] segs, boolean anchorStart, boolean anchorEnd, RuleBasedTransliterator.Data theData)154     public TransliterationRule(String input,
155                                int anteContextPos, int postContextPos,
156                                String output,
157                                int cursorPos, int cursorOffset,
158                                UnicodeMatcher[] segs,
159                                boolean anchorStart, boolean anchorEnd,
160                                RuleBasedTransliterator.Data theData) {
161         data = theData;
162 
163         // Do range checks only when warranted to save time
164         if (anteContextPos < 0) {
165             anteContextLength = 0;
166         } else {
167             if (anteContextPos > input.length()) {
168                 throw new IllegalArgumentException("Invalid ante context");
169             }
170             anteContextLength = anteContextPos;
171         }
172         if (postContextPos < 0) {
173             keyLength = input.length() - anteContextLength;
174         } else {
175             if (postContextPos < anteContextLength ||
176                 postContextPos > input.length()) {
177                 throw new IllegalArgumentException("Invalid post context");
178             }
179             keyLength = postContextPos - anteContextLength;
180         }
181         if (cursorPos < 0) {
182             cursorPos = output.length();
183         } else if (cursorPos > output.length()) {
184             throw new IllegalArgumentException("Invalid cursor position");
185         }
186 
187         // We don't validate the segments array.  The caller must
188         // guarantee that the segments are well-formed (that is, that
189         // all $n references in the output refer to indices of this
190         // array, and that no array elements are null).
191         this.segments = segs;
192 
193         pattern = input;
194         flags = 0;
195         if (anchorStart) {
196             flags |= ANCHOR_START;
197         }
198         if (anchorEnd) {
199             flags |= ANCHOR_END;
200         }
201 
202         anteContext = null;
203         if (anteContextLength > 0) {
204             anteContext = new StringMatcher(pattern.substring(0, anteContextLength),
205                                             0, data);
206         }
207 
208         key = null;
209         if (keyLength > 0) {
210             key = new StringMatcher(pattern.substring(anteContextLength, anteContextLength + keyLength),
211                                     0, data);
212         }
213 
214         int postContextLength = pattern.length() - keyLength - anteContextLength;
215         postContext = null;
216         if (postContextLength > 0) {
217             postContext = new StringMatcher(pattern.substring(anteContextLength + keyLength),
218                                             0, data);
219         }
220 
221         this.output = new StringReplacer(output, cursorPos + cursorOffset, data);
222     }
223 
224     /**
225      * Return the preceding context length.  This method is needed to
226      * support the <code>Transliterator</code> method
227      * <code>getMaximumContextLength()</code>.
228      */
getAnteContextLength()229     public int getAnteContextLength() {
230         return anteContextLength + (((flags & ANCHOR_START) != 0) ? 1 : 0);
231     }
232 
233     /**
234      * Internal method.  Returns 8-bit index value for this rule.
235      * This is the low byte of the first character of the key,
236      * unless the first character of the key is a set.  If it's a
237      * set, or otherwise can match multiple keys, the index value is -1.
238      */
getIndexValue()239     final int getIndexValue() {
240         if (anteContextLength == pattern.length()) {
241             // A pattern with just ante context {such as foo)>bar} can
242             // match any key.
243             return -1;
244         }
245         int c = UTF16.charAt(pattern, anteContextLength);
246         return data.lookupMatcher(c) == null ? (c & 0xFF) : -1;
247     }
248 
249     /**
250      * Internal method.  Returns true if this rule matches the given
251      * index value.  The index value is an 8-bit integer, 0..255,
252      * representing the low byte of the first character of the key.
253      * It matches this rule if it matches the first character of the
254      * key, or if the first character of the key is a set, and the set
255      * contains any character with a low byte equal to the index
256      * value.  If the rule contains only ante context, as in foo)>bar,
257      * then it will match any key.
258      */
matchesIndexValue(int v)259     final boolean matchesIndexValue(int v) {
260         // Delegate to the key, or if there is none, to the postContext.
261         // If there is neither then we match any key; return true.
262         UnicodeMatcher m = (key != null) ? key : postContext;
263         return (m != null) ? m.matchesIndexValue(v) : true;
264     }
265 
266     /**
267      * Return true if this rule masks another rule.  If r1 masks r2 then
268      * r1 matches any input string that r2 matches.  If r1 masks r2 and r2 masks
269      * r1 then r1 == r2.  Examples: "a>x" masks "ab>y".  "a>x" masks "a[b]>y".
270      * "[c]a>x" masks "[dc]a>y".
271      */
masks(TransliterationRule r2)272     public boolean masks(TransliterationRule r2) {
273         /* Rule r1 masks rule r2 if the string formed of the
274          * antecontext, key, and postcontext overlaps in the following
275          * way:
276          *
277          * r1:      aakkkpppp
278          * r2:     aaakkkkkpppp
279          *            ^
280          *
281          * The strings must be aligned at the first character of the
282          * key.  The length of r1 to the left of the alignment point
283          * must be <= the length of r2 to the left; ditto for the
284          * right.  The characters of r1 must equal (or be a superset
285          * of) the corresponding characters of r2.  The superset
286          * operation should be performed to check for UnicodeSet
287          * masking.
288          *
289          * Anchors:  Two patterns that differ only in anchors only
290          * mask one another if they are exactly equal, and r2 has
291          * all the anchors r1 has (optionally, plus some).  Here Y
292          * means the row masks the column, N means it doesn't.
293          *
294          *         ab   ^ab    ab$  ^ab$
295          *   ab    Y     Y     Y     Y
296          *  ^ab    N     Y     N     Y
297          *   ab$   N     N     Y     Y
298          *  ^ab$   N     N     N     Y
299          *
300          * Post context: {a}b masks ab, but not vice versa, since {a}b
301          * matches everything ab matches, and {a}b matches {|a|}b but ab
302          * does not.  Pre context is different (a{b} does not align with
303          * ab).
304          */
305 
306         /* LIMITATION of the current mask algorithm: Some rule
307          * maskings are currently not detected.  For example,
308          * "{Lu}]a>x" masks "A]a>y".  This can be added later. TODO
309          */
310 
311         int len = pattern.length();
312         int left = anteContextLength;
313         int left2 = r2.anteContextLength;
314         int right = pattern.length() - left;
315         int right2 = r2.pattern.length() - left2;
316 
317         // TODO Clean this up -- some logic might be combinable with the
318         // next statement.
319 
320         // Test for anchor masking
321         if (left == left2 && right == right2 &&
322             keyLength <= r2.keyLength &&
323             r2.pattern.regionMatches(0, pattern, 0, len)) {
324             // The following boolean logic implements the table above
325             return (flags == r2.flags) ||
326                 (!((flags & ANCHOR_START) != 0) && !((flags & ANCHOR_END) != 0)) ||
327                 (((r2.flags & ANCHOR_START) != 0) && ((r2.flags & ANCHOR_END) != 0));
328         }
329 
330         return left <= left2 &&
331             (right < right2 ||
332              (right == right2 && keyLength <= r2.keyLength)) &&
333             r2.pattern.regionMatches(left2 - left, pattern, 0, len);
334     }
335 
posBefore(Replaceable str, int pos)336     static final int posBefore(Replaceable str, int pos) {
337         return (pos > 0) ?
338             pos - UTF16.getCharCount(str.char32At(pos-1)) :
339             pos - 1;
340     }
341 
posAfter(Replaceable str, int pos)342     static final int posAfter(Replaceable str, int pos) {
343         return (pos >= 0 && pos < str.length()) ?
344             pos + UTF16.getCharCount(str.char32At(pos)) :
345             pos + 1;
346     }
347 
348     /**
349      * Attempt a match and replacement at the given position.  Return
350      * the degree of match between this rule and the given text.  The
351      * degree of match may be mismatch, a partial match, or a full
352      * match.  A mismatch means at least one character of the text
353      * does not match the context or key.  A partial match means some
354      * context and key characters match, but the text is not long
355      * enough to match all of them.  A full match means all context
356      * and key characters match.
357      *
358      * If a full match is obtained, perform a replacement, update pos,
359      * and return U_MATCH.  Otherwise both text and pos are unchanged.
360      *
361      * @param text the text
362      * @param pos the position indices
363      * @param incremental if TRUE, test for partial matches that may
364      * be completed by additional text inserted at pos.limit.
365      * @return one of <code>U_MISMATCH</code>,
366      * <code>U_PARTIAL_MATCH</code>, or <code>U_MATCH</code>.  If
367      * incremental is FALSE then U_PARTIAL_MATCH will not be returned.
368      */
matchAndReplace(Replaceable text, Transliterator.Position pos, boolean incremental)369     public int matchAndReplace(Replaceable text,
370                                Transliterator.Position pos,
371                                boolean incremental) {
372         // Matching and replacing are done in one method because the
373         // replacement operation needs information obtained during the
374         // match.  Another way to do this is to have the match method
375         // create a match result struct with relevant offsets, and to pass
376         // this into the replace method.
377 
378         // ============================ MATCH ===========================
379 
380         // Reset segment match data
381         if (segments != null) {
382             for (int i=0; i<segments.length; ++i) {
383                 ((StringMatcher) segments[i]).resetMatch();
384             }
385         }
386 
387         int keyLimit;
388         int[] intRef = new int[1];
389 
390         // ------------------------ Ante Context ------------------------
391 
392         // A mismatch in the ante context, or with the start anchor,
393         // is an outright U_MISMATCH regardless of whether we are
394         // incremental or not.
395         int oText; // offset into 'text'
396         int minOText;
397 
398         // Note (1): We process text in 16-bit code units, rather than
399         // 32-bit code points.  This works because stand-ins are
400         // always in the BMP and because we are doing a literal match
401         // operation, which can be done 16-bits at a time.
402 
403         int anteLimit = posBefore(text, pos.contextStart);
404 
405         int match;
406 
407         // Start reverse match at char before pos.start
408         intRef[0] = posBefore(text, pos.start);
409 
410         if (anteContext != null) {
411             match = anteContext.matches(text, intRef, anteLimit, false);
412             if (match != UnicodeMatcher.U_MATCH) {
413                 return UnicodeMatcher.U_MISMATCH;
414             }
415         }
416 
417         oText = intRef[0];
418 
419         minOText = posAfter(text, oText);
420 
421         // ------------------------ Start Anchor ------------------------
422 
423         if (((flags & ANCHOR_START) != 0) && oText != anteLimit) {
424             return UnicodeMatcher.U_MISMATCH;
425         }
426 
427         // -------------------- Key and Post Context --------------------
428 
429         intRef[0] = pos.start;
430 
431         if (key != null) {
432             match = key.matches(text, intRef, pos.limit, incremental);
433             if (match != UnicodeMatcher.U_MATCH) {
434                 return match;
435             }
436         }
437 
438         keyLimit = intRef[0];
439 
440         if (postContext != null) {
441             if (incremental && keyLimit == pos.limit) {
442                 // The key matches just before pos.limit, and there is
443                 // a postContext.  Since we are in incremental mode,
444                 // we must assume more characters may be inserted at
445                 // pos.limit -- this is a partial match.
446                 return UnicodeMatcher.U_PARTIAL_MATCH;
447             }
448 
449             match = postContext.matches(text, intRef, pos.contextLimit, incremental);
450             if (match != UnicodeMatcher.U_MATCH) {
451                 return match;
452             }
453         }
454 
455         oText = intRef[0];
456 
457         // ------------------------- Stop Anchor ------------------------
458 
459         if (((flags & ANCHOR_END)) != 0) {
460             if (oText != pos.contextLimit) {
461                 return UnicodeMatcher.U_MISMATCH;
462             }
463             if (incremental) {
464                 return UnicodeMatcher.U_PARTIAL_MATCH;
465             }
466         }
467 
468         // =========================== REPLACE ==========================
469 
470         // We have a full match.  The key is between pos.start and
471         // keyLimit.
472 
473         int newLength = output.replace(text, pos.start, keyLimit, intRef);
474         int lenDelta = newLength - (keyLimit - pos.start);
475         int newStart = intRef[0];
476 
477         oText += lenDelta;
478         pos.limit += lenDelta;
479         pos.contextLimit += lenDelta;
480         // Restrict new value of start to [minOText, min(oText, pos.limit)].
481         pos.start = Math.max(minOText, Math.min(Math.min(oText, pos.limit), newStart));
482         return UnicodeMatcher.U_MATCH;
483     }
484 
485     /**
486      * Create a source string that represents this rule.  Append it to the
487      * given string.
488      */
toRule(boolean escapeUnprintable)489     public String toRule(boolean escapeUnprintable) {
490        // int i;
491 
492         StringBuffer rule = new StringBuffer();
493 
494         // Accumulate special characters (and non-specials following them)
495         // into quoteBuf.  Append quoteBuf, within single quotes, when
496         // a non-quoted element must be inserted.
497         StringBuffer quoteBuf = new StringBuffer();
498 
499         // Do not emit the braces '{' '}' around the pattern if there
500         // is neither anteContext nor postContext.
501         boolean emitBraces =
502             (anteContext != null) || (postContext != null);
503 
504         // Emit start anchor
505         if ((flags & ANCHOR_START) != 0) {
506             rule.append('^');
507         }
508 
509         // Emit the input pattern
510         Utility.appendToRule(rule, anteContext, escapeUnprintable, quoteBuf);
511 
512         if (emitBraces) {
513             Utility.appendToRule(rule, '{', true, escapeUnprintable, quoteBuf);
514         }
515 
516         Utility.appendToRule(rule, key, escapeUnprintable, quoteBuf);
517 
518         if (emitBraces) {
519             Utility.appendToRule(rule, '}', true, escapeUnprintable, quoteBuf);
520         }
521 
522         Utility.appendToRule(rule, postContext, escapeUnprintable, quoteBuf);
523 
524         // Emit end anchor
525         if ((flags & ANCHOR_END) != 0) {
526             rule.append('$');
527         }
528 
529         Utility.appendToRule(rule, " > ", true, escapeUnprintable, quoteBuf);
530 
531         // Emit the output pattern
532 
533         Utility.appendToRule(rule, output.toReplacerPattern(escapeUnprintable),
534                      true, escapeUnprintable, quoteBuf);
535 
536         Utility.appendToRule(rule, ';', true, escapeUnprintable, quoteBuf);
537 
538         return rule.toString();
539     }
540 
541     /**
542      * Return a string representation of this object.
543      * @return string representation of this object
544      */
545     @Override
toString()546     public String toString() {
547         return '{' + toRule(true) + '}';
548     }
549 
550     /**
551      * Find the source and target sets, subject to the input filter.
552      * There is a known issue with filters containing multiple characters.
553      */
554     // TODO: Problem: the rule is [{ab}]c > x
555     // The filter is [a{bc}].
556     // If the input is abc, then the rule will work.
557     // However, following code applying the filter won't catch that case.
558 
addSourceTargetSet(UnicodeSet filter, UnicodeSet sourceSet, UnicodeSet targetSet, UnicodeSet revisiting)559     void addSourceTargetSet(UnicodeSet filter, UnicodeSet sourceSet, UnicodeSet targetSet, UnicodeSet revisiting) {
560         int limit = anteContextLength + keyLength;
561         UnicodeSet tempSource = new UnicodeSet();
562         UnicodeSet temp = new UnicodeSet();
563 
564         // We need to walk through the pattern.
565         // Iff some of the characters at ALL of the the positions are matched by the filter, then we add temp to toUnionTo
566         for (int i=anteContextLength; i<limit; ) {
567             int ch = UTF16.charAt(pattern, i);
568             i += UTF16.getCharCount(ch);
569             UnicodeMatcher matcher = data.lookupMatcher(ch);
570             if (matcher == null) {
571                 if (!filter.contains(ch)) {
572                     return;
573                 }
574                 tempSource.add(ch);
575             } else {
576                 try {
577                     if (!filter.containsSome((UnicodeSet) matcher)) {
578                         return;
579                     }
580                     matcher.addMatchSetTo(tempSource);
581                 } catch (ClassCastException e) { // if the matcher is not a UnicodeSet
582                     temp.clear();
583                     matcher.addMatchSetTo(temp);
584                     if (!filter.containsSome(temp)) {
585                         return;
586                     }
587                     tempSource.addAll(temp);
588                 }
589             }
590         }
591         // if we made our way through the gauntlet, add to source/target
592         sourceSet.addAll(tempSource);
593         output.addReplacementSetTo(targetSet);
594     }
595 }
596