]> git.saurik.com Git - apple/icu.git/blame - icuSources/layout/TibetanReordering.cpp
projects / apple / icu.git / blame
? search:
summary | shortlog | log | commit | commitdiff | tree
blob | blame (incremental) | history | HEAD
ICU-8.11.tar.gz
[apple/icu.git] / icuSources / layout / TibetanReordering.cpp
CommitLineData
73c04bcf
A		 1/*
2 *
3 * (C) Copyright IBM Corp. 1998-2006 - All Rights Reserved
4 *
5 * Developed at DIT - Government of Bhutan
6 *
7 * Contact person: Pema Geyleg - <pema_geyleg@druknet.bt>
8 *
9 * This file is a modification of the ICU file KhmerReordering.cpp
10 * by Jens Herden and Javier Sola who have given all their possible rights to IBM and the Governement of Bhutan
11 * A first module for Dzongkha was developed by Karunakar under Panlocalisation funding.
12 * Assistance for this module has been received from Namgay Thinley, Christopher Fynn and Javier Sola
13 *
14 */
15
16//#include <stdio.h>
17#include "LETypes.h"
18#include "OpenTypeTables.h"
19#include "TibetanReordering.h"
20#include "LEGlyphStorage.h"
21
22
23U_NAMESPACE_BEGIN
24
25// Characters that get refered to by name...
26enum
27{
28 C_DOTTED_CIRCLE = 0x25CC,
29 C_PRE_NUMBER_MARK = 0x0F3F
30 };
31
32
33enum
34{
35 // simple classes, they are used in the statetable (in this file) to control the length of a syllable
36 // they are also used to know where a character should be placed (location in reference to the base character)
37 // and also to know if a character, when independtly displayed, should be displayed with a dotted-circle to
38 // indicate error in syllable construction
39 _xx = TibetanClassTable::CC_RESERVED,
40 _ba = TibetanClassTable::CC_BASE,
41 _sj = TibetanClassTable::CC_SUBJOINED | TibetanClassTable::CF_DOTTED_CIRCLE | TibetanClassTable::CF_POS_BELOW,
42 _tp = TibetanClassTable::CC_TSA_PHRU | TibetanClassTable::CF_DOTTED_CIRCLE | TibetanClassTable::CF_POS_ABOVE,
43 _ac = TibetanClassTable::CC_A_CHUNG | TibetanClassTable::CF_DOTTED_CIRCLE | TibetanClassTable::CF_POS_BELOW,
44 _cs = TibetanClassTable::CC_COMP_SANSKRIT | TibetanClassTable::CF_DOTTED_CIRCLE | TibetanClassTable::CF_POS_BELOW,
45 _ha = TibetanClassTable::CC_HALANTA | TibetanClassTable::CF_DOTTED_CIRCLE | TibetanClassTable::CF_POS_BELOW,
46 _bv = TibetanClassTable::CC_BELOW_VOWEL | TibetanClassTable::CF_DOTTED_CIRCLE | TibetanClassTable::CF_POS_BELOW,
47 _av = TibetanClassTable::CC_ABOVE_VOWEL | TibetanClassTable::CF_DOTTED_CIRCLE | TibetanClassTable::CF_POS_ABOVE,
48 _an = TibetanClassTable::CC_ANUSVARA | TibetanClassTable::CF_DOTTED_CIRCLE | TibetanClassTable::CF_POS_ABOVE,
49 _cb = TibetanClassTable::CC_CANDRABINDU | TibetanClassTable::CF_DOTTED_CIRCLE | TibetanClassTable::CF_POS_ABOVE,
50 _vs = TibetanClassTable::CC_VISARGA | TibetanClassTable::CF_DOTTED_CIRCLE| TibetanClassTable::CF_POS_AFTER,
51 _as = TibetanClassTable::CC_ABOVE_S_MARK | TibetanClassTable::CF_DOTTED_CIRCLE | TibetanClassTable::CF_POS_ABOVE,
52 _bs = TibetanClassTable::CC_BELOW_S_MARK | TibetanClassTable::CF_DOTTED_CIRCLE | TibetanClassTable::CF_POS_BELOW,
53 _di = TibetanClassTable::CC_DIGIT | TibetanClassTable::CF_DIGIT,
54 _pd = TibetanClassTable::CC_PRE_DIGIT_MARK | TibetanClassTable::CF_DOTTED_CIRCLE | TibetanClassTable::CF_PREDIGIT | TibetanClassTable::CF_POS_BEFORE ,
55 _bd = TibetanClassTable::CC_POST_BELOW_DIGIT_M | TibetanClassTable::CF_DOTTED_CIRCLE | TibetanClassTable::CF_POS_AFTER
56};
57
58
59// Character class tables
60//_xx Non Combining characters
61//_ba Base Consonants
62//_sj Subjoined consonants
63//_tp Tsa - phru
64//_ac A-chung, Vowel Lengthening mark
65//_cs Precomposed Sanskrit vowel + subjoined consonants
66//_ha Halanta/Virama
67//_bv Below vowel
68//_av above vowel
69//_an Anusvara
70//_cb Candrabindu
71//_vs Visaraga/Post mark
72//_as Upper Stress marks
73//_bs Lower Stress marks
74//_di Digit
75//_pd Number pre combining, Needs reordering
76//_bd Other number combining marks
77
78static const TibetanClassTable::CharClass tibetanCharClasses[] =
79{
80 // 0 1 2 3 4 5 6 7 8 9 a b c d e f
81 _xx, _ba, _xx, _xx, _ba, _ba, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, // 0F00 - 0F0F 0
82 _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _bd, _bd, _xx, _xx, _xx, _xx, _xx, _xx, // 0F10 - 0F1F 1
83 _di, _di, _di, _di, _di, _di, _di, _di, _di, _di, _xx, _xx, _xx, _xx, _xx, _xx, // 0F20 - 0F2F 2
84 _xx, _xx, _xx, _xx, _xx, _bs, _xx, _bs, _xx, _tp, _xx, _xx, _xx, _xx, _bd, _pd, // 0F30 - 0F3F 3
85 _ba, _ba, _ba, _ba, _ba, _ba, _ba, _ba, _xx, _ba, _ba, _ba, _ba, _ba, _ba, _ba, // 0F40 - 0F4F 4
86 _ba, _ba, _ba, _ba, _ba, _ba, _ba, _ba, _ba, _ba, _ba, _ba, _ba, _ba, _ba, _ba, // 0F50 - 0F5F 5
87 _ba, _ba, _ba, _ba, _ba, _ba, _ba, _ba, _ba, _ba, _ba, _xx, _xx, _xx, _xx, _xx, // 0F60 - 0F6F 6
88 _xx, _ac, _av, _cs, _bv, _bv, _cs, _cs, _cs, _cs, _av, _av, _av, _av, _an, _vs, // 0F70 - 0F7F 7
89 _av, _cs, _cb, _cb, _ha, _xx, _as, _as, _ba, _ba, _ba, _ba, _xx, _xx, _xx, _xx, // 0F80 - 0F8F 8
90 _sj, _sj, _sj, _sj, _sj, _sj, _sj, _sj, _xx, _sj, _sj, _sj, _sj, _sj, _sj, _sj, // 0F90 - 0F9F 9
91 _sj, _sj, _sj, _sj, _sj, _sj, _sj, _sj, _sj, _sj, _sj, _sj, _sj, _sj, _sj, _sj, // 0FA0 - 0FAF a
92 _sj, _sj, _sj, _sj, _sj, _sj, _sj, _sj, _sj, _sj, _sj, _sj, _sj, _xx, _sj, _sj, // 0FB0 - 0FBF b
93 _xx, _xx, _xx, _xx, _xx, _xx, _bs, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, // 0FC0 - 0FCF c
94 _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx,// 0FD0 - 0FDF d
95 _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, // 0FE0 - 0FEF e
96 _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, // 0FF0 - 0FFF f
97};
98
99
100//
101// Tibetan Class Tables
102//
103
104//
105// The range of characters defined in the above table is defined here. For Tibetan 0F00 to 0FFF
106// Even if the Tibetan range is bigger, most of the characters are not combinable, and therefore treated
107// as _xx
108static const TibetanClassTable tibetanClassTable = {0x0F00, 0x0FFF, tibetanCharClasses};
109
110
111// Below we define how a character in the input string is either in the tibetanCharClasses table
112// (in which case we get its type back), or an unknown object in which case we get _xx (CC_RESERVED) back
113TibetanClassTable::CharClass TibetanClassTable::getCharClass(LEUnicode ch) const
114{
115 if (ch < firstChar || ch > lastChar) {
116 return CC_RESERVED;
117 }
118
119 return classTable[ch - firstChar];
120}
121
122const TibetanClassTable *TibetanClassTable::getTibetanClassTable()
123{
124 return &tibetanClassTable;
125}
126
127
128
129class TibetanReorderingOutput : public UMemory {
130private:
131 le_int32 fOutIndex;
132 LEUnicode *fOutChars;
133
134 LEGlyphStorage &fGlyphStorage;
135
136
137public:
138 TibetanReorderingOutput(LEUnicode *outChars, LEGlyphStorage &glyphStorage)
139 : fOutIndex(0), fOutChars(outChars), fGlyphStorage(glyphStorage)
140 {
141 // nothing else to do...
142 }
143
144 ~TibetanReorderingOutput()
145 {
146 // nothing to do here...
147 }
148
149 void writeChar(LEUnicode ch, le_uint32 charIndex, FeatureMask featureMask)
150 {
151 LEErrorCode success = LE_NO_ERROR;
152
153 fOutChars[fOutIndex] = ch;
154
155 fGlyphStorage.setCharIndex(fOutIndex, charIndex, success);
156 fGlyphStorage.setAuxData(fOutIndex, featureMask, success);
157
158 fOutIndex += 1;
159 }
160
161 le_int32 getOutputIndex()
162 {
163 return fOutIndex;
164 }
165};
166
167
168//TODO remove unused flags
169#define ccmpFeatureTag LE_CCMP_FEATURE_TAG
170#define blwfFeatureTag LE_BLWF_FEATURE_TAG
171#define pstfFeatureTag LE_PSTF_FEATURE_TAG
172#define presFeatureTag LE_PRES_FEATURE_TAG
173#define blwsFeatureTag LE_BLWS_FEATURE_TAG
174#define abvsFeatureTag LE_ABVS_FEATURE_TAG
175#define pstsFeatureTag LE_PSTS_FEATURE_TAG
176
177#define blwmFeatureTag LE_BLWM_FEATURE_TAG
178#define abvmFeatureTag LE_ABVM_FEATURE_TAG
179#define distFeatureTag LE_DIST_FEATURE_TAG
180
181#define prefFeatureTag LE_PREF_FEATURE_TAG
182#define abvfFeatureTag LE_ABVF_FEATURE_TAG
183#define cligFeatureTag LE_CLIG_FEATURE_TAG
184#define mkmkFeatureTag LE_MKMK_FEATURE_TAG
185
186// Shaping features
187#define prefFeatureMask 0x80000000UL
188#define blwfFeatureMask 0x40000000UL
189#define abvfFeatureMask 0x20000000UL
190#define pstfFeatureMask 0x10000000UL
191#define presFeatureMask 0x08000000UL
192#define blwsFeatureMask 0x04000000UL
193#define abvsFeatureMask 0x02000000UL
194#define pstsFeatureMask 0x01000000UL
195#define cligFeatureMask 0x00800000UL
196#define ccmpFeatureMask 0x00040000UL
197
198// Positioning features
199#define distFeatureMask 0x00400000UL
200#define blwmFeatureMask 0x00200000UL
201#define abvmFeatureMask 0x00100000UL
202#define mkmkFeatureMask 0x00080000UL
203
204#define tagPref (ccmpFeatureMask | prefFeatureMask | presFeatureMask | cligFeatureMask | distFeatureMask)
205#define tagAbvf (ccmpFeatureMask | abvfFeatureMask | abvsFeatureMask | cligFeatureMask | distFeatureMask | abvmFeatureMask | mkmkFeatureMask)
206#define tagPstf (ccmpFeatureMask | blwfFeatureMask | blwsFeatureMask | prefFeatureMask | presFeatureMask | pstfFeatureMask | pstsFeatureMask | cligFeatureMask | distFeatureMask | blwmFeatureMask)
207#define tagBlwf (ccmpFeatureMask | blwfFeatureMask | blwsFeatureMask | cligFeatureMask | distFeatureMask | blwmFeatureMask | mkmkFeatureMask)
208#define tagDefault (ccmpFeatureMask | prefFeatureMask | blwfFeatureMask | presFeatureMask | blwsFeatureMask | cligFeatureMask | distFeatureMask | abvmFeatureMask | blwmFeatureMask | mkmkFeatureMask)
209
210
211
212// These are in the order in which the features need to be applied
213// for correct processing
214static const FeatureMap featureMap[] =
215{
216 // Shaping features
217 {ccmpFeatureTag, ccmpFeatureMask},
218 {prefFeatureTag, prefFeatureMask},
219 {blwfFeatureTag, blwfFeatureMask},
220 {abvfFeatureTag, abvfFeatureMask},
221 {pstfFeatureTag, pstfFeatureMask},
222 {presFeatureTag, presFeatureMask},
223 {blwsFeatureTag, blwsFeatureMask},
224 {abvsFeatureTag, abvsFeatureMask},
225 {pstsFeatureTag, pstsFeatureMask},
226 {cligFeatureTag, cligFeatureMask},
227
228 // Positioning features
229 {distFeatureTag, distFeatureMask},
230 {blwmFeatureTag, blwmFeatureMask},
231 {abvmFeatureTag, abvmFeatureMask},
232 {mkmkFeatureTag, mkmkFeatureMask},
233};
234
235static const le_int32 featureMapCount = LE_ARRAY_SIZE(featureMap);
236
237// The stateTable is used to calculate the end (the length) of a well
238// formed Tibetan Syllable.
239//
240// Each horizontal line is ordered exactly the same way as the values in TibetanClassTable
241// CharClassValues in TibetanReordering.h This coincidence of values allows the
242// follow up of the table.
243//
244// Each line corresponds to a state, which does not necessarily need to be a type
245// of component... for example, state 2 is a base, with is always a first character
246// in the syllable, but the state could be produced a consonant of any type when
247// it is the first character that is analysed (in ground state).
248//
249static const le_int8 tibetanStateTable[][TibetanClassTable::CC_COUNT] =
250{
251
252
253 //Dzongkha state table
254 //xx ba sj tp ac cs ha bv av an cb vs as bs di pd bd
255 { 1, 2, 4, 3, 8, 7, 9, 10, 14, 13, 17, 18, 19, 19, 20, 21, 21,}, // 0 - ground state
256 {-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,}, // 1 - exit state (or sign to the right of the syllable)
257 {-1, -1, 4, 3, 8, 7, 9, 10, 14, 13, 17, 18, 19, 19, -1, -1, -1,}, // 2 - Base consonant
258 {-1, -1, 5, -1, 8, 7, -1, 10, 14, 13, 17, 18, 19, 19, -1, -1, -1,}, // 3 - Tsa phru after base
259 {-1, -1, 4, 6, 8, 7, 9, 10, 14, 13, 17, 18, 19, 19, -1, -1, -1,}, // 4 - Subjoined consonant after base
260 {-1, -1, 5, -1, 8, 7, -1, 10, 14, 13, 17, 18, 19, 19, -1, -1, -1,}, // 5 - Subjoined consonant after tsa phru
261 {-1, -1, -1, -1, 8, 7, -1, 10, 14, 13, 17, 18, 19, 19, -1, -1, -1,}, // 6 - Tsa phru after subjoined consonant
262 {-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, 19, 19, -1, -1, -1,}, // 7 - Pre Composed Sanskrit
263 {-1, -1, -1, -1, -1, -1, -1, 10, 14, 13, 17, 18, 19, 19, -1, -1, -1,}, // 8 - A-chung
264 {-1, -1, -1, -1, -1, -1, -1, -1, 14, 13, 17, -1, 19, 19, -1, -1, -1,}, // 9 - Halanta
265 {-1, -1, -1, -1, -1, -1, -1, 11, 14, 13, 17, 18, 19, 19, -1, -1, -1,}, // 10 - below vowel 1
266 {-1, -1, -1, -1, -1, -1, -1, 12, 14, 13, 17, 18, 19, 19, -1, -1, -1,}, // 11 - below vowel 2
267 {-1, -1, -1, -1, -1, -1, -1, -1, 14, 13, 17, 18, 19, 19, -1, -1, -1,}, // 12 - below vowel 3
268 {-1, -1, -1, -1, -1, -1, -1, -1, 14, 17, 17, 18, 19, 19, -1, -1, -1,}, // 13 - Anusvara before vowel
269 {-1, -1, -1, -1, -1, -1, -1, -1, 15, 17, 17, 18, 19, 19, -1, -1, -1,}, // 14 - above vowel 1
270 {-1, -1, -1, -1, -1, -1, -1, -1, 16, 17, 17, 18, 19, 19, -1, -1, -1,}, // 15 - above vowel 2
271 {-1, -1, -1, -1, -1, -1, -1, -1, -1, 17, 17, 18, 19, 19, -1, -1, -1,}, // 16 - above vowel 3
272 {-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, 18, 19, 19, -1, -1, -1,}, // 17 - Anusvara or Candrabindu after vowel
273 {-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, 19, 19, -1, -1, -1,}, // 18 - Visarga
274 {-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,}, // 19 - strss mark
275 {-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, 21, 21,}, // 20 - digit
276 {-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,}, // 21 - digit mark
277
278
279};
280
281
282const FeatureMap *TibetanReordering::getFeatureMap(le_int32 &count)
283{
284 count = featureMapCount;
285
286 return featureMap;
287}
288
289
290// Given an input string of characters and a location in which to start looking
291// calculate, using the state table, which one is the last character of the syllable
292// that starts in the starting position.
293le_int32 TibetanReordering::findSyllable(const TibetanClassTable *classTable, const LEUnicode *chars, le_int32 prev, le_int32 charCount)
294{
295 le_int32 cursor = prev;
296 le_int8 state = 0;
297
298 while (cursor < charCount) {
299 TibetanClassTable::CharClass charClass = (classTable->getCharClass(chars[cursor]) & TibetanClassTable::CF_CLASS_MASK);
300
301 state = tibetanStateTable[state][charClass];
302
303 if (state < 0) {
304 break;
305 }
306
307 cursor += 1;
308 }
309
310 return cursor;
311}
312
313
314// This is the real reordering function as applied to the Tibetan language
315
316le_int32 TibetanReordering::reorder(const LEUnicode *chars, le_int32 charCount, le_int32,
317 LEUnicode *outChars, LEGlyphStorage &glyphStorage)
318{
319 const TibetanClassTable *classTable = TibetanClassTable::getTibetanClassTable();
320
321 TibetanReorderingOutput output(outChars, glyphStorage);
322 TibetanClassTable::CharClass charClass;
323 le_int32 i, prev = 0;
324
325 // This loop only exits when we reach the end of a run, which may contain
326 // several syllables.
327 while (prev < charCount) {
328 le_int32 syllable = findSyllable(classTable, chars, prev, charCount);
329
330 // shall we add a dotted circle?
331 // If in the position in which the base should be (first char in the string) there is
332 // a character that has the Dotted circle flag (a character that cannot be a base)
333 // then write a dotted circle
334 if (classTable->getCharClass(chars[prev]) & TibetanClassTable::CF_DOTTED_CIRCLE) {
335 output.writeChar(C_DOTTED_CIRCLE, prev, tagDefault);
336 }
337
338 // copy the rest to output, inverting the pre-number mark if present after a digit.
339 for (i = prev; i < syllable; i += 1) {
340 charClass = classTable->getCharClass(chars[i]);
341
342 if ((TibetanClassTable::CF_DIGIT & charClass)
343 && ( classTable->getCharClass(chars[i+1]) & TibetanClassTable::CF_PREDIGIT))
344 {
345 output.writeChar(C_PRE_NUMBER_MARK, i, tagPref);
346 output.writeChar(chars[i], i+1 , tagPref);
347 i += 1;
348 } else {
349 switch (charClass & TibetanClassTable::CF_POS_MASK) {
350
351 // If the present character is a number, and the next character is a pre-number combining mark
352 // then the two characters are reordered
353
354 case TibetanClassTable::CF_POS_ABOVE :
355 output.writeChar(chars[i], i, tagAbvf);
356 break;
357
358 case TibetanClassTable::CF_POS_AFTER :
359 output.writeChar(chars[i], i, tagPstf);
360 break;
361
362 case TibetanClassTable::CF_POS_BELOW :
363 output.writeChar(chars[i], i, tagBlwf);
364 break;
365
366 default:
367 // default - any other characters
368 output.writeChar(chars[i], i, tagDefault);
369 break;
370 } // switch
371 } // if
372 } // for
373
374 prev = syllable; // move the pointer to the start of next syllable
375 }
376
377 return output.getOutputIndex();
378}
379
380
381U_NAMESPACE_END
mirror of ICU