eSpeak NG is an open source speech synthesizer that supports more than hundred languages and accents.
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

tr_languages.c 67KB

12345678910111213141516171819202122232425262728293031323334353637383940414243444546474849505152535455565758596061626364656667686970717273747576777879808182838485868788899091929394959697989910010110210310410510610710810911011111211311411511611711811912012112212312412512612712812913013113213313413513613713813914014114214314414514614714814915015115215315415515615715815916016116216316416516616716816917017117217317417517617717817918018118218318418518618718818919019119219319419519619719819920020120220320420520620720820921021121221321421521621721821922022122222322422522622722822923023123223323423523623723823924024124224324424524624724824925025125225325425525625725825926026126226326426526626726826927027127227327427527627727827928028128228328428528628728828929029129229329429529629729829930030130230330430530630730830931031131231331431531631731831932032132232332432532632732832933033133233333433533633733833934034134234334434534634734834935035135235335435535635735835936036136236336436536636736836937037137237337437537637737837938038138238338438538638738838939039139239339439539639739839940040140240340440540640740840941041141241341441541641741841942042142242342442542642742842943043143243343443543643743843944044144244344444544644744844945045145245345445545645745845946046146246346446546646746846947047147247347447547647747847948048148248348448548648748848949049149249349449549649749849950050150250350450550650750850951051151251351451551651751851952052152252352452552652752852953053153253353453553653753853954054154254354454554654754854955055155255355455555655755855956056156256356456556656756856957057157257357457557657757857958058158258358458558658758858959059159259359459559659759859960060160260360460560660760860961061161261361461561661761861962062162262362462562662762862963063163263363463563663763863964064164264364464564664764864965065165265365465565665765865966066166266366466566666766866967067167267367467567667767867968068168268368468568668768868969069169269369469569669769869970070170270370470570670770870971071171271371471571671771871972072172272372472572672772872973073173273373473573673773873974074174274374474574674774874975075175275375475575675775875976076176276376476576676776876977077177277377477577677777877978078178278378478578678778878979079179279379479579679779879980080180280380480580680780880981081181281381481581681781881982082182282382482582682782882983083183283383483583683783883984084184284384484584684784884985085185285385485585685785885986086186286386486586686786886987087187287387487587687787887988088188288388488588688788888989089189289389489589689789889990090190290390490590690790890991091191291391491591691791891992092192292392492592692792892993093193293393493593693793893994094194294394494594694794894995095195295395495595695795895996096196296396496596696796896997097197297397497597697797897998098198298398498598698798898999099199299399499599699799899910001001100210031004100510061007100810091010101110121013101410151016101710181019102010211022102310241025102610271028102910301031103210331034103510361037103810391040104110421043104410451046104710481049105010511052105310541055105610571058105910601061106210631064106510661067106810691070107110721073107410751076107710781079108010811082108310841085108610871088108910901091109210931094109510961097109810991100110111021103110411051106110711081109111011111112111311141115111611171118111911201121112211231124112511261127112811291130113111321133113411351136113711381139114011411142114311441145114611471148114911501151115211531154115511561157115811591160116111621163116411651166116711681169117011711172117311741175117611771178117911801181118211831184118511861187118811891190119111921193119411951196119711981199120012011202120312041205120612071208120912101211121212131214121512161217121812191220122112221223122412251226122712281229123012311232123312341235123612371238123912401241124212431244124512461247124812491250125112521253125412551256125712581259126012611262126312641265126612671268126912701271127212731274127512761277127812791280128112821283128412851286128712881289129012911292129312941295129612971298129913001301130213031304130513061307130813091310131113121313131413151316131713181319132013211322132313241325132613271328132913301331133213331334133513361337133813391340134113421343134413451346134713481349135013511352135313541355135613571358135913601361136213631364136513661367136813691370137113721373137413751376137713781379138013811382138313841385138613871388138913901391139213931394139513961397139813991400140114021403140414051406140714081409141014111412141314141415141614171418141914201421142214231424142514261427142814291430143114321433143414351436143714381439144014411442144314441445144614471448144914501451145214531454145514561457145814591460146114621463146414651466146714681469147014711472147314741475147614771478147914801481148214831484148514861487148814891490149114921493149414951496149714981499150015011502150315041505150615071508150915101511151215131514151515161517151815191520152115221523152415251526152715281529153015311532153315341535153615371538153915401541154215431544154515461547154815491550155115521553155415551556155715581559156015611562156315641565156615671568156915701571157215731574157515761577157815791580158115821583158415851586158715881589159015911592159315941595
  1. /*
  2. * Copyright (C) 2005 to 2015 by Jonathan Duddington
  3. * email: [email protected]
  4. * Copyright (C) 2015-2016 Reece H. Dunn
  5. *
  6. * This program is free software; you can redistribute it and/or modify
  7. * it under the terms of the GNU General Public License as published by
  8. * the Free Software Foundation; either version 3 of the License, or
  9. * (at your option) any later version.
  10. *
  11. * This program is distributed in the hope that it will be useful,
  12. * but WITHOUT ANY WARRANTY; without even the implied warranty of
  13. * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  14. * GNU General Public License for more details.
  15. *
  16. * You should have received a copy of the GNU General Public License
  17. * along with this program; if not, see: <http://www.gnu.org/licenses/>.
  18. */
  19. #include "config.h"
  20. #include <ctype.h>
  21. #include <locale.h>
  22. #include <stdint.h>
  23. #include <stdio.h>
  24. #include <stdlib.h>
  25. #include <string.h>
  26. #include <espeak-ng/espeak_ng.h>
  27. #include <espeak-ng/speak_lib.h>
  28. #include <espeak-ng/encoding.h>
  29. #include "speech.h"
  30. #include "phoneme.h"
  31. #include "voice.h"
  32. #include "synthesize.h"
  33. #include "translate.h"
  34. #define L_grc 0x677263 // grc Ancient Greek
  35. #define L_jbo 0x6a626f // jbo Lojban
  36. #define L_mni 0x627079 // bpy Manipuri
  37. #define L_pap 0x706170 // pap Papiamento]
  38. #define L_zhy 0x7a6879 // zhy Chinese (Cantonese)
  39. #define L_cmn 0x636D6E // cmn Chinese (Mandarin)
  40. // start of unicode pages for character sets
  41. #define OFFSET_GREEK 0x380
  42. #define OFFSET_CYRILLIC 0x420
  43. #define OFFSET_ARMENIAN 0x530
  44. #define OFFSET_HEBREW 0x590
  45. #define OFFSET_ARABIC 0x600
  46. #define OFFSET_SYRIAC 0x700
  47. #define OFFSET_THAANA 0x780 // Divehi/Maldives
  48. #define OFFSET_DEVANAGARI 0x900
  49. #define OFFSET_BENGALI 0x980
  50. #define OFFSET_GURMUKHI 0xa00
  51. #define OFFSET_GUJARATI 0xa80
  52. #define OFFSET_ORIYA 0xb00
  53. #define OFFSET_TAMIL 0xb80
  54. #define OFFSET_TELUGU 0xc00
  55. #define OFFSET_KANNADA 0xc80
  56. #define OFFSET_MALAYALAM 0xd00
  57. #define OFFSET_SINHALA 0x0d80
  58. #define OFFSET_THAI 0x0e00
  59. #define OFFSET_LAO 0x0e80
  60. #define OFFSET_TIBET 0x0f00
  61. #define OFFSET_MYANMAR 0x1000
  62. #define OFFSET_GEORGIAN 0x10a0
  63. #define OFFSET_KOREAN 0x1100
  64. #define OFFSET_ETHIOPIC 0x1200
  65. // character ranges must be listed in ascending unicode order
  66. ALPHABET alphabets[] = {
  67. { "_el", OFFSET_GREEK, 0x380, 0x3ff, L('e', 'l'), AL_DONT_NAME | AL_NOT_LETTERS | AL_WORDS },
  68. { "_cyr", OFFSET_CYRILLIC, 0x400, 0x52f, 0, 0 },
  69. { "_hy", OFFSET_ARMENIAN, 0x530, 0x58f, L('h', 'y'), AL_WORDS },
  70. { "_he", OFFSET_HEBREW, 0x590, 0x5ff, 0, 0 },
  71. { "_ar", OFFSET_ARABIC, 0x600, 0x6ff, 0, 0 },
  72. { "_syc", OFFSET_SYRIAC, 0x700, 0x74f, 0, 0 },
  73. { "_dv", OFFSET_THAANA, 0x780, 0x7bf, 0, 0 },
  74. { "_hi", OFFSET_DEVANAGARI, 0x900, 0x97f, L('h', 'i'), AL_WORDS },
  75. { "_bn", OFFSET_BENGALI, 0x0980, 0x9ff, L('b', 'n'), AL_WORDS },
  76. { "_gur", OFFSET_GURMUKHI, 0xa00, 0xa7f, L('p', 'a'), AL_WORDS },
  77. { "_gu", OFFSET_GUJARATI, 0xa80, 0xaff, L('g', 'u'), AL_WORDS },
  78. { "_or", OFFSET_ORIYA, 0xb00, 0xb7f, 0, 0 },
  79. { "_ta", OFFSET_TAMIL, 0xb80, 0xbff, L('t', 'a'), AL_WORDS },
  80. { "_te", OFFSET_TELUGU, 0xc00, 0xc7f, L('t', 'e'), 0 },
  81. { "_kn", OFFSET_KANNADA, 0xc80, 0xcff, L('k', 'n'), AL_WORDS },
  82. { "_ml", OFFSET_MALAYALAM, 0xd00, 0xd7f, L('m', 'l'), AL_WORDS },
  83. { "_si", OFFSET_SINHALA, 0xd80, 0xdff, L('s', 'i'), AL_WORDS },
  84. { "_th", OFFSET_THAI, 0xe00, 0xe7f, 0, 0 },
  85. { "_lo", OFFSET_LAO, 0xe80, 0xeff, 0, 0 },
  86. { "_ti", OFFSET_TIBET, 0xf00, 0xfff, 0, 0 },
  87. { "_my", OFFSET_MYANMAR, 0x1000, 0x109f, 0, 0 },
  88. { "_ka", OFFSET_GEORGIAN, 0x10a0, 0x10ff, L('k', 'a'), AL_WORDS },
  89. { "_ko", OFFSET_KOREAN, 0x1100, 0x11ff, L('k', 'o'), AL_WORDS },
  90. { "_eth", OFFSET_ETHIOPIC, 0x1200, 0x139f, 0, 0 },
  91. { "_braille", 0x2800, 0x2800, 0x28ff, 0, AL_NO_SYMBOL },
  92. { "_ja", 0x3040, 0x3040, 0x30ff, 0, AL_NOT_CODE },
  93. { "_zh", 0x3100, 0x3100, 0x9fff, 0, AL_NOT_CODE },
  94. { "_ko", 0xa700, 0xa700, 0xd7ff, L('k', 'o'), AL_NOT_CODE | AL_WORDS },
  95. { NULL, 0, 0, 0, 0, 0 }
  96. };
  97. ALPHABET *AlphabetFromChar(int c)
  98. {
  99. // Find the alphabet from a character.
  100. ALPHABET *alphabet = alphabets;
  101. while (alphabet->name != NULL) {
  102. if (c <= alphabet->range_max) {
  103. if (c >= alphabet->range_min)
  104. return alphabet;
  105. else
  106. break;
  107. }
  108. alphabet++;
  109. }
  110. return NULL;
  111. }
  112. static void Translator_Russian(Translator *tr);
  113. static void SetLetterVowel(Translator *tr, int c)
  114. {
  115. tr->letter_bits[c] = (tr->letter_bits[c] & 0x40) | 0x81; // keep value for group 6 (front vowels e,i,y)
  116. }
  117. static void ResetLetterBits(Translator *tr, int groups)
  118. {
  119. // Clear all the specified groups
  120. unsigned int ix;
  121. unsigned int mask;
  122. mask = ~groups;
  123. for (ix = 0; ix < sizeof(tr->letter_bits); ix++)
  124. tr->letter_bits[ix] &= mask;
  125. }
  126. static void SetLetterBits(Translator *tr, int group, const char *string)
  127. {
  128. int bits;
  129. unsigned char c;
  130. bits = (1L << group);
  131. while ((c = *string++) != 0)
  132. tr->letter_bits[c] |= bits;
  133. }
  134. static void SetLetterBitsRange(Translator *tr, int group, int first, int last)
  135. {
  136. int bits;
  137. int ix;
  138. bits = (1L << group);
  139. for (ix = first; ix <= last; ix++)
  140. tr->letter_bits[ix] |= bits;
  141. }
  142. // ignore these characters
  143. static const unsigned short chars_ignore_default[] = {
  144. // U+00AD SOFT HYPHEN
  145. // Used to mark hyphenation points in words for where to split a
  146. // word at the end of a line to provide readable justified text.
  147. 0xad, 1,
  148. // U+200C ZERO WIDTH NON-JOINER
  149. // Used to prevent combined ligatures being displayed in their
  150. // combined form.
  151. 0x200c, 1,
  152. // U+200D ZERO WIDTH JOINER
  153. // Used to indicate an alternative connected form made up of the
  154. // characters surrounding the ZWJ in Devanagari, Kannada, Malayalam
  155. // and Emoji.
  156. // 0x200d, 1, // Not ignored.
  157. // End of the ignored character list.
  158. 0, 0
  159. };
  160. // alternatively, ignore characters but allow zero-width-non-joiner (lang-fa)
  161. static const unsigned short chars_ignore_zwnj_hyphen[] = {
  162. // U+00AD SOFT HYPHEN
  163. // Used to mark hyphenation points in words for where to split a
  164. // word at the end of a line to provide readable justified text.
  165. 0xad, 1,
  166. // U+0640 TATWEEL (KASHIDA)
  167. // Used in Arabic scripts to stretch characters for justifying
  168. // the text.
  169. 0x640, 1,
  170. // U+200C ZERO WIDTH NON-JOINER
  171. // Used to prevent combined ligatures being displayed in their
  172. // combined form.
  173. 0x200c, '-',
  174. // U+200D ZERO WIDTH JOINER
  175. // Used to indicate an alternative connected form made up of the
  176. // characters surrounding the ZWJ in Devanagari, Kannada, Malayalam
  177. // and Emoji.
  178. // 0x200d, 1, // Not ignored.
  179. // End of the ignored character list.
  180. 0, 0
  181. };
  182. const unsigned char utf8_ordinal[] = { 0xc2, 0xba, 0 }; // masculine ordinal character, UTF-8
  183. const unsigned char utf8_null[] = { 0 }; // null string, UTF-8
  184. static Translator *NewTranslator(void)
  185. {
  186. Translator *tr;
  187. int ix;
  188. static const unsigned char stress_amps2[] = { 18, 18, 20, 20, 20, 22, 22, 20 };
  189. static const short stress_lengths2[8] = { 182, 140, 220, 220, 220, 240, 260, 280 };
  190. static const wchar_t empty_wstring[1] = { 0 };
  191. static const wchar_t punct_in_word[2] = { '\'', 0 }; // allow hyphen within words
  192. static const unsigned char default_tunes[6] = { 0, 1, 2, 3, 0, 0 };
  193. // Translates character codes in the range transpose_min to transpose_max to
  194. // a number in the range 1 to 63. 0 indicates there is no translation.
  195. // Used up to 57 (max of 63)
  196. static const char transpose_map_latin[] = {
  197. 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, // 0x60
  198. 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 0, 0, 0, 0, 0, // 0x70
  199. 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, // 0x80
  200. 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, // 0x90
  201. 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, // 0xa0
  202. 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, // 0xb0
  203. 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, // 0xc0
  204. 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, // 0xd0
  205. 27, 28, 29, 0, 0, 30, 31, 32, 33, 34, 35, 36, 0, 37, 38, 0, // 0xe0
  206. 0, 0, 0, 39, 0, 0, 40, 0, 41, 0, 42, 0, 43, 0, 0, 0, // 0xf0
  207. 0, 0, 0, 44, 0, 45, 0, 46, 0, 0, 0, 0, 0, 47, 0, 0, // 0x100
  208. 0, 48, 0, 0, 0, 0, 0, 0, 0, 49, 0, 0, 0, 0, 0, 0, // 0x110
  209. 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, // 0x120
  210. 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, // 0x130
  211. 0, 0, 50, 0, 51, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, // 0x140
  212. 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 52, 0, 0, 0, 0, // 0x150
  213. 0, 53, 0, 54, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, // 0x160
  214. 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 55, 0, 56, 0, 57, 0, // 0x170
  215. };
  216. if ((tr = (Translator *)malloc(sizeof(Translator))) == NULL)
  217. return NULL;
  218. tr->encoding = ESPEAKNG_ENCODING_ISO_8859_1;
  219. dictionary_name[0] = 0;
  220. tr->dictionary_name[0] = 0;
  221. tr->dict_condition = 0;
  222. tr->dict_min_size = 0;
  223. tr->data_dictrules = NULL; // language_1 translation rules file
  224. tr->data_dictlist = NULL; // language_2 dictionary lookup file
  225. tr->transpose_min = 0x60;
  226. tr->transpose_max = 0x17f;
  227. tr->transpose_map = transpose_map_latin;
  228. tr->frequent_pairs = NULL;
  229. // only need lower case
  230. tr->letter_bits_offset = 0;
  231. memset(tr->letter_bits, 0, sizeof(tr->letter_bits));
  232. memset(tr->letter_groups, 0, sizeof(tr->letter_groups));
  233. // 0-6 sets of characters matched by A B C H F G Y in pronunciation rules
  234. // these may be set differently for different languages
  235. SetLetterBits(tr, 0, "aeiou"); // A vowels, except y
  236. SetLetterBits(tr, 1, "bcdfgjklmnpqstvxz"); // B hard consonants, excluding h,r,w
  237. SetLetterBits(tr, 2, "bcdfghjklmnpqrstvwxz"); // C all consonants
  238. SetLetterBits(tr, 3, "hlmnr"); // H 'soft' consonants
  239. SetLetterBits(tr, 4, "cfhkpqstx"); // F voiceless consonants
  240. SetLetterBits(tr, 5, "bdgjlmnrvwyz"); // G voiced
  241. SetLetterBits(tr, 6, "eiy"); // Letter group Y, front vowels
  242. SetLetterBits(tr, 7, "aeiouy"); // vowels, including y
  243. tr->char_plus_apostrophe = empty_wstring;
  244. tr->punct_within_word = punct_in_word;
  245. tr->chars_ignore = chars_ignore_default;
  246. for (ix = 0; ix < 8; ix++) {
  247. tr->stress_amps[ix] = stress_amps2[ix];
  248. tr->stress_amps_r[ix] = stress_amps2[ix] - 1;
  249. tr->stress_lengths[ix] = stress_lengths2[ix];
  250. }
  251. memset(&(tr->langopts), 0, sizeof(tr->langopts));
  252. tr->langopts.max_lengthmod = 500;
  253. tr->langopts.lengthen_tonic = 20;
  254. tr->langopts.stress_rule = STRESSPOSN_2R;
  255. tr->langopts.unstressed_wd1 = 1;
  256. tr->langopts.unstressed_wd2 = 3;
  257. tr->langopts.param[LOPT_SONORANT_MIN] = 95;
  258. tr->langopts.param[LOPT_LONG_VOWEL_THRESHOLD] = 190/2;
  259. tr->langopts.param[LOPT_MAXAMP_EOC] = 19;
  260. tr->langopts.param[LOPT_UNPRONOUNCABLE] = 's'; // don't count this character at start of word
  261. tr->langopts.param[LOPT_BRACKET_PAUSE] = 4; // pause at bracket
  262. tr->langopts.param2[LOPT_BRACKET_PAUSE] = 2; // pauses when announcing bracket names
  263. tr->langopts.max_initial_consonants = 3;
  264. tr->langopts.replace_chars = NULL;
  265. tr->langopts.alt_alphabet_lang = L('e', 'n');
  266. tr->langopts.roman_suffix = utf8_null;
  267. SetLengthMods(tr, 201);
  268. tr->langopts.long_stop = 100;
  269. tr->langopts.max_roman = 49;
  270. tr->langopts.min_roman = 2;
  271. tr->langopts.thousands_sep = ',';
  272. tr->langopts.decimal_sep = '.';
  273. tr->langopts.break_numbers = BREAK_THOUSANDS; // 1000, 1000,000 1,000,000 etc
  274. tr->langopts.max_digits = 14;
  275. memcpy(tr->punct_to_tone, punctuation_to_tone, sizeof(tr->punct_to_tone));
  276. memcpy(tr->langopts.tunes, default_tunes, sizeof(tr->langopts.tunes));
  277. return tr;
  278. }
  279. // common letter pairs, encode these as a single byte
  280. // 2 bytes, using the transposed character codes
  281. static const short pairs_ru[] = {
  282. 0x010c, // ла 21052 0x23
  283. 0x010e, // на 18400
  284. 0x0113, // та 14254
  285. 0x0301, // ав 31083
  286. 0x030f, // ов 13420
  287. 0x060e, // не 21798
  288. 0x0611, // ре 19458
  289. 0x0903, // ви 16226
  290. 0x0b01, // ак 14456
  291. 0x0b0f, // ок 17836
  292. 0x0c01, // ал 13324
  293. 0x0c09, // ил 16877
  294. 0x0e01, // ан 15359
  295. 0x0e06, // ен 13543 0x30
  296. 0x0e09, // ин 17168
  297. 0x0e0e, // нн 15973
  298. 0x0e0f, // он 22373
  299. 0x0e1c, // ын 15052
  300. 0x0f03, // во 24947
  301. 0x0f11, // ро 13552
  302. 0x0f12, // со 16368
  303. 0x100f, // оп 19054
  304. 0x1011, // рп 17067
  305. 0x1101, // ар 23967
  306. 0x1106, // ер 18795
  307. 0x1109, // ир 13797
  308. 0x110f, // ор 21737
  309. 0x1213, // тс 25076
  310. 0x1220, // яс 14310
  311. 0x7fff
  312. };
  313. static const unsigned int replace_cyrillic_latin[] =
  314. { 0x430, 'a',
  315. 0x431, 'b',
  316. 0x446, 'c',
  317. 0x45b, 0x107,
  318. 0x447, 0x10d,
  319. 0x45f, 'd'+(0x17e<<16),
  320. 0x455, 'd'+('z'<<16),
  321. 0x434, 'd',
  322. 0x452, 0x111,
  323. 0x435, 'e',
  324. 0x444, 'f',
  325. 0x433, 'g',
  326. 0x445, 'h',
  327. 0x438, 'i',
  328. 0x458, 'j',
  329. 0x43a, 'k',
  330. 0x459, 'l'+('j'<<16),
  331. 0x43b, 'l',
  332. 0x43c, 'm',
  333. 0x45a, 'n'+('j'<<16),
  334. 0x43d, 'n',
  335. 0x43e, 'o',
  336. 0x43f, 'p',
  337. 0x440, 'r',
  338. 0x441, 's',
  339. 0x448, 0x161,
  340. 0x442, 't',
  341. 0x443, 'u',
  342. 0x432, 'v',
  343. 0x437, 'z',
  344. 0x436, 0x17e,
  345. 0x453, 0x111,
  346. 0x45c, 0x107,
  347. 0 }; // ѓ ѕ ќ
  348. static const unsigned char ru_vowels[] = { 0x10, 0x15, 0x31, 0x18, 0x1e, 0x23, 0x2b, 0x2d, 0x2e, 0x2f, 0xb9, 0xc9, 0x91, 0x8f, 0x36, 0 }; // also kazakh
  349. static const unsigned char ru_consonants[] = { 0x11, 0x12, 0x13, 0x14, 0x16, 0x17, 0x19, 0x1a, 0x1b, 0x1c, 0x1d, 0x1f, 0x20, 0x21, 0x22, 0x24, 0x25, 0x26, 0x27, 0x28, 0x29, 0x2a, 0x2c, 0x73, 0x7b, 0x83, 0x9b, 0 };
  350. static void SetCyrillicLetters(Translator *tr)
  351. {
  352. // character codes offset by 0x420
  353. static const char ru_soft[] = { 0x2c, 0x19, 0x27, 0x29, 0 }; // letter group B [k ts; s;]
  354. static const char ru_hard[] = { 0x2a, 0x16, 0x26, 0x28, 0 }; // letter group H [S Z ts]
  355. static const char ru_nothard[] = { 0x11, 0x12, 0x13, 0x14, 0x17, 0x19, 0x1a, 0x1b, 0x1c, 0x1d, 0x1f, 0x20, 0x21, 0x22, 0x24, 0x25, 0x27, 0x29, 0x2c, 0 };
  356. static const char ru_voiced[] = { 0x11, 0x12, 0x13, 0x14, 0x16, 0x17, 0 }; // letter group G (voiced obstruents)
  357. static const char ru_ivowels[] = { 0x2c, 0x2e, 0x2f, 0x31, 0 }; // letter group Y (iotated vowels & soft-sign)
  358. tr->encoding = ESPEAKNG_ENCODING_KOI8_R;
  359. tr->transpose_min = 0x430; // convert cyrillic from unicode into range 0x01 to 0x22
  360. tr->transpose_max = 0x451;
  361. tr->transpose_map = NULL;
  362. tr->frequent_pairs = pairs_ru;
  363. tr->letter_bits_offset = OFFSET_CYRILLIC;
  364. memset(tr->letter_bits, 0, sizeof(tr->letter_bits));
  365. SetLetterBits(tr, LETTERGP_A, (char *)ru_vowels);
  366. SetLetterBits(tr, LETTERGP_B, ru_soft);
  367. SetLetterBits(tr, LETTERGP_C, (char *)ru_consonants);
  368. SetLetterBits(tr, LETTERGP_H, ru_hard);
  369. SetLetterBits(tr, LETTERGP_F, ru_nothard);
  370. SetLetterBits(tr, LETTERGP_G, ru_voiced);
  371. SetLetterBits(tr, LETTERGP_Y, ru_ivowels);
  372. SetLetterBits(tr, LETTERGP_VOWEL2, (char *)ru_vowels);
  373. }
  374. static void SetIndicLetters(Translator *tr)
  375. {
  376. // Set letter types for Indic scripts, Devanagari, Tamill, etc
  377. static const char dev_consonants2[] = { 0x02, 0x03, 0x58, 0x59, 0x5a, 0x5b, 0x5c, 0x5d, 0x5e, 0x5f, 0x7b, 0x7c, 0x7e, 0x7f, 0 };
  378. static const char dev_vowels2[] = { 0x60, 0x61, 0x55, 0x56, 0x57, 0x62, 0x63, 0 }; // non-consecutive vowels and vowel-signs
  379. memset(tr->letter_bits, 0, sizeof(tr->letter_bits));
  380. SetLetterBitsRange(tr, LETTERGP_A, 0x04, 0x14); // vowel letters
  381. SetLetterBitsRange(tr, LETTERGP_A, 0x3e, 0x4d); // + vowel signs, and virama
  382. SetLetterBits(tr, LETTERGP_A, dev_vowels2); // + extra vowels and vowel signs
  383. SetLetterBitsRange(tr, LETTERGP_B, 0x3e, 0x4d); // vowel signs, and virama
  384. SetLetterBits(tr, LETTERGP_B, dev_vowels2); // + extra vowels and vowel signs
  385. SetLetterBitsRange(tr, LETTERGP_C, 0x15, 0x39); // the main consonant range
  386. SetLetterBits(tr, LETTERGP_C, dev_consonants2); // + additional consonants
  387. SetLetterBitsRange(tr, LETTERGP_Y, 0x04, 0x14); // vowel letters
  388. SetLetterBitsRange(tr, LETTERGP_Y, 0x3e, 0x4c); // + vowel signs
  389. SetLetterBits(tr, LETTERGP_Y, dev_vowels2); // + extra vowels and vowel signs
  390. tr->langopts.param[LOPT_UNPRONOUNCABLE] = 1; // disable check for unpronouncable words
  391. tr->langopts.suffix_add_e = tr->letter_bits_offset + 0x4d; // virama
  392. }
  393. static void SetupTranslator(Translator *tr, const short *lengths, const unsigned char *amps)
  394. {
  395. if (lengths != NULL)
  396. memcpy(tr->stress_lengths, lengths, sizeof(tr->stress_lengths));
  397. if (amps != NULL)
  398. memcpy(tr->stress_amps, amps, sizeof(tr->stress_amps));
  399. }
  400. Translator *SelectTranslator(const char *name)
  401. {
  402. int name2 = 0;
  403. Translator *tr;
  404. static const short stress_lengths_equal[8] = { 230, 230, 230, 230, 0, 0, 230, 230 };
  405. static const unsigned char stress_amps_equal[8] = { 19, 19, 19, 19, 19, 19, 19, 19 };
  406. static const short stress_lengths_fr[8] = { 190, 170, 190, 200, 0, 0, 190, 240 };
  407. static const unsigned char stress_amps_fr[8] = { 18, 16, 18, 18, 18, 18, 18, 18 };
  408. static const unsigned char stress_amps_sk[8] = { 17, 16, 20, 20, 20, 22, 22, 21 };
  409. static const short stress_lengths_sk[8] = { 190, 190, 210, 210, 0, 0, 210, 210 };
  410. static const short stress_lengths_ta[8] = { 200, 200, 210, 210, 0, 0, 230, 230 };
  411. static const short stress_lengths_ta2[8] = { 230, 230, 240, 240, 0, 0, 260, 260 };
  412. static const unsigned char stress_amps_ta[8] = { 18, 18, 18, 18, 20, 20, 22, 22 };
  413. tr = NewTranslator();
  414. strcpy(tr->dictionary_name, name);
  415. // convert name string into a word of up to 4 characters, for the switch()
  416. while (*name != 0)
  417. name2 = (name2 << 8) + *name++;
  418. switch (name2)
  419. {
  420. case L('a', 'f'):
  421. {
  422. static const short stress_lengths_af[8] = { 170, 140, 220, 220, 0, 0, 250, 270 };
  423. SetupTranslator(tr, stress_lengths_af, NULL);
  424. tr->langopts.stress_rule = STRESSPOSN_1L;
  425. tr->langopts.vowel_pause = 0x30;
  426. tr->langopts.param[LOPT_DIERESES] = 1;
  427. tr->langopts.param[LOPT_PREFIXES] = 1;
  428. SetLetterVowel(tr, 'y'); // add 'y' to vowels
  429. tr->langopts.numbers = NUM_SWAP_TENS | NUM_HUNDRED_AND | NUM_SINGLE_AND | NUM_ROMAN | NUM_1900;
  430. tr->langopts.accents = 1;
  431. }
  432. break;
  433. case L('a', 'm'): // Amharic, Ethiopia
  434. {
  435. SetupTranslator(tr, stress_lengths_fr, stress_amps_fr);
  436. tr->letter_bits_offset = OFFSET_ETHIOPIC;
  437. tr->langopts.stress_rule = STRESSPOSN_1L;
  438. tr->langopts.stress_flags = S_NO_AUTO_2 | S_FINAL_DIM; // don't use secondary stress
  439. tr->langopts.length_mods0 = tr->langopts.length_mods; // don't lengthen vowels in the last syllable
  440. tr->langopts.param[LOPT_UNPRONOUNCABLE] = 1; // disable check for unpronouncable words
  441. tr->langopts.numbers = NUM_OMIT_1_HUNDRED;
  442. }
  443. break;
  444. case L('a', 'r'): // Arabic
  445. tr->transpose_min = 0x620; // for ar_list, use 6-bit character codes
  446. tr->transpose_max = 0x65f;
  447. tr->transpose_map = NULL;
  448. tr->letter_bits_offset = OFFSET_ARABIC;
  449. tr->langopts.numbers = NUM_SWAP_TENS | NUM_AND_UNITS | NUM_HUNDRED_AND | NUM_OMIT_1_HUNDRED | NUM_AND_HUNDRED | NUM_THOUSAND_AND | NUM_OMIT_1_THOUSAND;
  450. tr->langopts.param[LOPT_UNPRONOUNCABLE] = 1; // disable check for unpronouncable words
  451. tr->encoding = ESPEAKNG_ENCODING_ISO_8859_6;
  452. break;
  453. case L('b', 'g'): // Bulgarian
  454. {
  455. SetCyrillicLetters(tr);
  456. SetLetterVowel(tr, 0x2a);
  457. tr->encoding = ESPEAKNG_ENCODING_ISO_8859_5;
  458. tr->langopts.param[LOPT_UNPRONOUNCABLE] = 0x432; // [v] don't count this character at start of word
  459. tr->langopts.param[LOPT_REGRESSIVE_VOICING] = 0x107; // devoice at end of word, and change voicing to match a following consonant (except v)
  460. tr->langopts.param[LOPT_REDUCE] = 2;
  461. tr->langopts.stress_rule = STRESSPOSN_2R;
  462. tr->langopts.numbers = NUM_DECIMAL_COMMA | NUM_ALLOW_SPACE | NUM_OMIT_1_HUNDRED | NUM_HUNDRED_AND | NUM_AND_UNITS | NUM_SINGLE_AND | NUM_ROMAN | NUM_ROMAN_ORDINAL | NUM_ROMAN_CAPITALS;
  463. tr->langopts.thousands_sep = ' '; // don't allow dot as thousands separator
  464. }
  465. break;
  466. case L('b', 'n'): // Bengali
  467. case L('a', 's'): // Assamese
  468. case L_mni: // Manipuri (temporary placement - it's not indo-european)
  469. {
  470. static const short stress_lengths_bn[8] = { 180, 180, 210, 210, 0, 0, 230, 240 };
  471. static const unsigned char stress_amps_bn[8] = { 18, 18, 18, 18, 20, 20, 22, 22 };
  472. static const char bn_consonants2[3] = { 0x70, 0x71, 0 };
  473. SetupTranslator(tr, stress_lengths_bn, stress_amps_bn);
  474. tr->langopts.length_mods0 = tr->langopts.length_mods; // don't lengthen vowels in the last syllable
  475. tr->langopts.stress_rule = STRESSPOSN_1L;
  476. tr->langopts.stress_flags = S_MID_DIM | S_FINAL_DIM; // use 'diminished' for unstressed final syllable
  477. tr->letter_bits_offset = OFFSET_BENGALI;
  478. SetIndicLetters(tr); // call this after setting OFFSET_BENGALI
  479. SetLetterBitsRange(tr, LETTERGP_B, 0x01, 0x01); // candranindu
  480. SetLetterBitsRange(tr, LETTERGP_F, 0x3e, 0x4c); // vowel signs, but not virama
  481. SetLetterBits(tr, LETTERGP_C, bn_consonants2);
  482. tr->langopts.numbers = NUM_SWAP_TENS;
  483. tr->langopts.break_numbers = 0x24924aa8; // for languages which have numbers for 100,000 and 100,00,000, eg Hindi
  484. if (name2 == L_mni) {
  485. tr->langopts.numbers = 1;
  486. tr->langopts.numbers2 = NUM2_SWAP_THOUSANDS;
  487. }
  488. }
  489. break;
  490. case L('b', 'o'): // Tibet
  491. {
  492. tr->langopts.stress_rule = STRESSPOSN_1L;
  493. tr->letter_bits_offset = OFFSET_TIBET;
  494. SetLetterBitsRange(tr, LETTERGP_A, 0x71, 0x7d); // vowel signs
  495. SetLetterBitsRange(tr, LETTERGP_B, 0x71, 0x81); // vowel signs and subjoined letters
  496. SetLetterBitsRange(tr, LETTERGP_B, 0x90, 0xbc);
  497. SetLetterBitsRange(tr, LETTERGP_C, 0x40, 0x6c); // consonant letters (not subjoined)
  498. tr->langopts.param[LOPT_UNPRONOUNCABLE] = 1; // disable check for unpronouncable words
  499. tr->langopts.numbers = 1;
  500. }
  501. break;
  502. case L('c', 'y'): // Welsh
  503. {
  504. static const short stress_lengths_cy[8] = { 170, 220, 180, 180, 0, 0, 250, 270 };
  505. static const unsigned char stress_amps_cy[8] = { 17, 15, 18, 18, 0, 0, 22, 20 }; // 'diminished' is used to mark a quieter, final unstressed syllable
  506. SetupTranslator(tr, stress_lengths_cy, stress_amps_cy);
  507. tr->encoding = ESPEAKNG_ENCODING_ISO_8859_14;
  508. tr->langopts.stress_rule = STRESSPOSN_2R;
  509. // 'diminished' is an unstressed final syllable
  510. tr->langopts.stress_flags = S_FINAL_DIM_ONLY | S_FINAL_NO_2;
  511. tr->langopts.unstressed_wd1 = 0;
  512. tr->langopts.unstressed_wd2 = 2;
  513. tr->langopts.param[LOPT_SONORANT_MIN] = 120; // limit the shortening of sonorants before short vowels
  514. tr->langopts.numbers = NUM_OMIT_1_HUNDRED;
  515. SetLetterVowel(tr, 'w'); // add letter to vowels and remove from consonants
  516. SetLetterVowel(tr, 'y');
  517. }
  518. break;
  519. case L('d', 'a'): // Danish
  520. {
  521. static const short stress_lengths_da[8] = { 160, 140, 200, 200, 0, 0, 220, 230 };
  522. SetupTranslator(tr, stress_lengths_da, NULL);
  523. tr->langopts.stress_rule = STRESSPOSN_1L;
  524. tr->langopts.param[LOPT_PREFIXES] = 1;
  525. SetLetterVowel(tr, 'y');
  526. tr->langopts.numbers = NUM_DECIMAL_COMMA | NUM_SWAP_TENS | NUM_HUNDRED_AND | NUM_OMIT_1_HUNDRED | NUM_ORDINAL_DOT | NUM_1900 | NUM_ROMAN | NUM_ROMAN_CAPITALS | NUM_ROMAN_ORDINAL;
  527. }
  528. break;
  529. case L('d', 'e'):
  530. {
  531. static const short stress_lengths_de[8] = { 150, 130, 200, 200, 0, 0, 270, 270 };
  532. static const unsigned char stress_amps_de[] = { 20, 20, 20, 20, 20, 22, 22, 20 };
  533. SetupTranslator(tr, stress_lengths_de, stress_amps_de);
  534. tr->langopts.stress_rule = STRESSPOSN_1L;
  535. tr->langopts.word_gap = 0x8; // don't use linking phonemes
  536. tr->langopts.vowel_pause = 0x30;
  537. tr->langopts.param[LOPT_PREFIXES] = 1;
  538. tr->langopts.param[LOPT_REGRESSIVE_VOICING] = 0x100; // devoice at end of word
  539. tr->langopts.param[LOPT_LONG_VOWEL_THRESHOLD] = 175/2;
  540. tr->langopts.numbers = NUM_DECIMAL_COMMA | NUM_SWAP_TENS | NUM_ALLOW_SPACE | NUM_ORDINAL_DOT | NUM_ROMAN;
  541. SetLetterVowel(tr, 'y');
  542. tr->langopts.param[LOPT_UNPRONOUNCABLE] = 2; // use de_rules for unpronouncable rules
  543. }
  544. break;
  545. case L('d', 'v'): // Divehi (Maldives)
  546. {
  547. SetupTranslator(tr, stress_lengths_ta, stress_amps_ta);
  548. tr->langopts.param[LOPT_UNPRONOUNCABLE] = 1; // disable check for unpronouncable words
  549. tr->langopts.length_mods0 = tr->langopts.length_mods; // don't lengthen vowels in the last syllable
  550. tr->letter_bits_offset = OFFSET_THAANA;
  551. tr->langopts.stress_rule = STRESSPOSN_1L;
  552. tr->langopts.stress_flags = S_MID_DIM | S_FINAL_DIM; // use 'diminished' for unstressed final syllable
  553. SetLetterBitsRange(tr, LETTERGP_B, 0x26, 0x30); // vowel signs, and virama
  554. tr->langopts.break_numbers = 0x14a8; // 1000, 100,000 10,000,000
  555. tr->langopts.numbers = 1;
  556. }
  557. break;
  558. case L('e', 'n'):
  559. {
  560. static const short stress_lengths_en[8] = { 182, 140, 220, 220, 0, 0, 248, 275 };
  561. SetupTranslator(tr, stress_lengths_en, NULL);
  562. tr->langopts.stress_rule = STRESSPOSN_1L;
  563. tr->langopts.stress_flags = 0x08;
  564. tr->langopts.numbers = NUM_HUNDRED_AND | NUM_ROMAN | NUM_1900;
  565. tr->langopts.param[LOPT_COMBINE_WORDS] = 2; // allow "mc" to cmbine with the following word
  566. tr->langopts.suffix_add_e = 'e';
  567. tr->langopts.param[LOPT_UNPRONOUNCABLE] = 2; // use en_rules for unpronouncable rules
  568. SetLetterBits(tr, 6, "aeiouy"); // Group Y: vowels, including y
  569. }
  570. break;
  571. case L('e', 'l'): // Greek
  572. case L_grc: // Ancient Greek
  573. {
  574. static const short stress_lengths_el[8] = { 155, 180, 210, 210, 0, 0, 270, 300 };
  575. static const unsigned char stress_amps_el[8] = { 15, 12, 20, 20, 20, 22, 22, 21 }; // 'diminished' is used to mark a quieter, final unstressed syllable
  576. // character codes offset by 0x380
  577. static const char el_vowels[] = { 0x10, 0x2c, 0x2d, 0x2e, 0x2f, 0x30, 0x31, 0x35, 0x37, 0x39, 0x3f, 0x45, 0x49, 0x4a, 0x4b, 0x4c, 0x4d, 0x4e, 0x4f, 0 };
  578. static const char el_fvowels[] = { 0x2d, 0x2e, 0x2f, 0x35, 0x37, 0x39, 0x45, 0x4d, 0 }; // ε η ι υ έ ή ί ύ _
  579. static const char el_voiceless[] = { 0x38, 0x3a, 0x3e, 0x40, 0x42, 0x43, 0x44, 0x46, 0x47, 0 }; // θ κ ξ π ς σ τ φ χ _
  580. static const char el_consonants[] = { 0x32, 0x33, 0x34, 0x36, 0x38, 0x3a, 0x3b, 0x3c, 0x3d, 0x3e, 0x40, 0x41, 0x42, 0x43, 0x44, 0x46, 0x47, 0x48, 0 };
  581. static const wchar_t el_char_apostrophe[] = { 0x3c3, 0 }; // σ _
  582. SetupTranslator(tr, stress_lengths_el, stress_amps_el);
  583. tr->encoding = ESPEAKNG_ENCODING_ISO_8859_7;
  584. tr->char_plus_apostrophe = el_char_apostrophe;
  585. tr->letter_bits_offset = OFFSET_GREEK;
  586. memset(tr->letter_bits, 0, sizeof(tr->letter_bits));
  587. SetLetterBits(tr, LETTERGP_A, el_vowels);
  588. SetLetterBits(tr, LETTERGP_VOWEL2, el_vowels);
  589. SetLetterBits(tr, LETTERGP_B, el_voiceless);
  590. SetLetterBits(tr, LETTERGP_C, el_consonants);
  591. SetLetterBits(tr, LETTERGP_Y, el_fvowels); // front vowels: ε η ι υ _
  592. tr->langopts.length_mods0 = tr->langopts.length_mods; // don't lengthen vowels in the last syllable
  593. tr->langopts.stress_rule = STRESSPOSN_2R;
  594. tr->langopts.stress_flags = S_FINAL_DIM_ONLY; // mark unstressed final syllables as diminished
  595. tr->langopts.unstressed_wd1 = 0;
  596. tr->langopts.unstressed_wd2 = 2;
  597. tr->langopts.param[LOPT_SONORANT_MIN] = 130; // limit the shortening of sonorants before short vowels
  598. tr->langopts.numbers = NUM_SINGLE_STRESS | NUM_DECIMAL_COMMA;
  599. tr->langopts.numbers2 = 0x2 | NUM2_MULTIPLE_ORDINAL | NUM2_ORDINAL_NO_AND; // variant form of numbers before thousands
  600. if (name2 == L_grc) {
  601. // ancient greek
  602. tr->langopts.param[LOPT_UNPRONOUNCABLE] = 1;
  603. }
  604. }
  605. break;
  606. case L('e', 'o'):
  607. {
  608. static const short stress_lengths_eo[8] = { 150, 140, 180, 180, 0, 0, 200, 200 };
  609. static const unsigned char stress_amps_eo[] = { 16, 14, 20, 20, 20, 22, 22, 21 };
  610. static const wchar_t eo_char_apostrophe[2] = { 'l', 0 };
  611. SetupTranslator(tr, stress_lengths_eo, stress_amps_eo);
  612. tr->encoding = ESPEAKNG_ENCODING_ISO_8859_3;
  613. tr->char_plus_apostrophe = eo_char_apostrophe;
  614. tr->langopts.vowel_pause = 2;
  615. tr->langopts.stress_rule = STRESSPOSN_2R;
  616. tr->langopts.stress_flags = S_FINAL_DIM_ONLY | S_FINAL_NO_2;
  617. tr->langopts.unstressed_wd2 = 2;
  618. tr->langopts.numbers = NUM_DECIMAL_COMMA | NUM_OMIT_1_HUNDRED | NUM_ALLOW_SPACE | NUM_ROMAN;
  619. }
  620. break;
  621. case L('e', 's'): // Spanish
  622. case L('a', 'n'): // Aragonese
  623. case L('c', 'a'): // Catalan
  624. case L('i', 'a'): // Interlingua
  625. case L_pap: // Papiamento
  626. {
  627. static const short stress_lengths_es[8] = { 160, 145, 155, 150, 0, 0, 200, 245 };
  628. static const unsigned char stress_amps_es[8] = { 16, 14, 15, 16, 20, 20, 22, 22 }; // 'diminished' is used to mark a quieter, final unstressed syllable
  629. static const wchar_t ca_punct_within_word[] = { '\'', 0xb7, 0 }; // ca: allow middle-dot within word
  630. SetupTranslator(tr, stress_lengths_es, stress_amps_es);
  631. tr->langopts.length_mods0 = tr->langopts.length_mods; // don't lengthen vowels in the last syllable
  632. tr->langopts.stress_rule = STRESSPOSN_2R;
  633. // stress last syllable if it doesn't end in vowel or "s" or "n"
  634. // 'diminished' is an unstressed final syllable
  635. tr->langopts.stress_flags = S_FINAL_SPANISH | S_FINAL_DIM_ONLY | S_FINAL_NO_2;
  636. tr->langopts.unstressed_wd1 = 0;
  637. tr->langopts.unstressed_wd2 = 2;
  638. tr->langopts.param[LOPT_SONORANT_MIN] = 120; // limit the shortening of sonorants before short vowels
  639. tr->langopts.numbers = NUM_SINGLE_STRESS | NUM_DECIMAL_COMMA | NUM_AND_UNITS | NUM_OMIT_1_HUNDRED | NUM_OMIT_1_THOUSAND | NUM_ROMAN | NUM_ROMAN_AFTER | NUM_DFRACTION_4;
  640. tr->langopts.numbers2 = NUM2_MULTIPLE_ORDINAL | NUM2_ORDINAL_NO_AND;
  641. if (name2 == L('c', 'a')) {
  642. // stress last syllable unless word ends with a vowel
  643. tr->punct_within_word = ca_punct_within_word;
  644. tr->langopts.stress_flags = S_FINAL_SPANISH | S_FINAL_DIM_ONLY | S_FINAL_NO_2 | S_NO_AUTO_2;
  645. } else if (name2 == L('i', 'a')) {
  646. tr->langopts.stress_flags = S_FINAL_SPANISH | S_FINAL_DIM_ONLY | S_FINAL_NO_2;
  647. tr->langopts.numbers = NUM_DECIMAL_COMMA | NUM_OMIT_1_HUNDRED | NUM_OMIT_1_THOUSAND | NUM_ROMAN | NUM_ROMAN_AFTER;
  648. } else if (name2 == L('a', 'n')) {
  649. tr->langopts.stress_flags = S_FINAL_SPANISH | S_FINAL_DIM_ONLY | S_FINAL_NO_2;
  650. tr->langopts.numbers = NUM_SINGLE_STRESS | NUM_DECIMAL_COMMA | NUM_AND_UNITS | NUM_OMIT_1_HUNDRED | NUM_OMIT_1_THOUSAND | NUM_ROMAN | NUM_ROMAN_ORDINAL;
  651. tr->langopts.numbers2 = NUM2_ORDINAL_NO_AND;
  652. tr->langopts.roman_suffix = utf8_ordinal;
  653. } else if (name2 == L_pap) {
  654. // stress last syllable unless word ends with a vowel
  655. tr->langopts.stress_rule = STRESSPOSN_1R;
  656. tr->langopts.stress_flags = S_FINAL_VOWEL_UNSTRESSED | S_FINAL_DIM_ONLY | S_FINAL_NO_2 | S_NO_AUTO_2;
  657. } else
  658. tr->langopts.param[LOPT_UNPRONOUNCABLE] = 2; // use es_rules for unpronouncable rules
  659. }
  660. break;
  661. case L('e', 'u'): // basque
  662. {
  663. static const short stress_lengths_eu[8] = { 200, 200, 200, 200, 0, 0, 210, 230 }; // very weak stress
  664. static const unsigned char stress_amps_eu[8] = { 16, 16, 18, 18, 18, 18, 18, 18 };
  665. SetupTranslator(tr, stress_lengths_eu, stress_amps_eu);
  666. tr->langopts.stress_rule = STRESSPOSN_2L; // ?? second syllable, but not on a word-final vowel
  667. tr->langopts.stress_flags = S_FINAL_VOWEL_UNSTRESSED;
  668. tr->langopts.param[LOPT_SUFFIX] = 1;
  669. tr->langopts.numbers = NUM_SINGLE_STRESS | NUM_DECIMAL_COMMA | NUM_HUNDRED_AND | NUM_OMIT_1_HUNDRED | NUM_OMIT_1_THOUSAND | NUM_VIGESIMAL;
  670. }
  671. break;
  672. case L('f', 'a'): // Farsi
  673. {
  674. // Convert characters in the range 0x620 to 0x6cc to the range 1 to 63.
  675. // 0 indicates no translation for this character
  676. static const char transpose_map_fa[] = {
  677. 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, // 0x620
  678. 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 0, 0, 0, 0, 0, // 0x630
  679. 0, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, // 0x640
  680. 42, 43, 0, 0, 44, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, // 0x650
  681. 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, // 0x660
  682. 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 45, 0, // 0x670
  683. 0, 0, 0, 0, 0, 0, 46, 0, 0, 0, 0, 0, 0, 0, 0, 0, // 0x680
  684. 0, 0, 0, 0, 0, 0, 0, 0, 47, 0, 0, 0, 0, 0, 0, 0, // 0x690
  685. 0, 0, 0, 0, 0, 0, 0, 0, 0, 48, 0, 0, 0, 0, 0, 49, // 0x6a0
  686. 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, // 0x6b0
  687. 50, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 51 // 0x6c0
  688. };
  689. tr->transpose_min = 0x620;
  690. tr->transpose_max = 0x6cc;
  691. tr->transpose_map = transpose_map_fa;
  692. tr->letter_bits_offset = OFFSET_ARABIC;
  693. tr->langopts.numbers = NUM_AND_UNITS | NUM_HUNDRED_AND;
  694. tr->langopts.param[LOPT_UNPRONOUNCABLE] = 1; // disable check for unpronouncable words
  695. tr->chars_ignore = chars_ignore_zwnj_hyphen; // replace ZWNJ by hyphen
  696. }
  697. break;
  698. case L('e', 't'): // Estonian
  699. tr->encoding = ESPEAKNG_ENCODING_ISO_8859_4;
  700. // fallthrough:
  701. case L('f', 'i'): // Finnish
  702. {
  703. static const unsigned char stress_amps_fi[8] = { 18, 16, 22, 22, 20, 22, 22, 22 };
  704. static const short stress_lengths_fi[8] = { 150, 180, 200, 200, 0, 0, 210, 250 };
  705. SetupTranslator(tr, stress_lengths_fi, stress_amps_fi);
  706. tr->langopts.stress_rule = STRESSPOSN_1L;
  707. tr->langopts.stress_flags = S_FINAL_DIM_ONLY | S_FINAL_NO_2 | S_2_TO_HEAVY; // move secondary stress from light to a following heavy syllable
  708. tr->langopts.param[LOPT_IT_DOUBLING] = 1;
  709. tr->langopts.long_stop = 130;
  710. tr->langopts.numbers = NUM_DECIMAL_COMMA + NUM_ALLOW_SPACE;
  711. SetLetterVowel(tr, 'y');
  712. tr->langopts.spelling_stress = 1;
  713. tr->langopts.intonation_group = 3; // less intonation, don't raise pitch at comma
  714. }
  715. break;
  716. case L('f', 'r'): // french
  717. {
  718. SetupTranslator(tr, stress_lengths_fr, stress_amps_fr);
  719. tr->langopts.stress_rule = STRESSPOSN_1R; // stress on final syllable
  720. tr->langopts.stress_flags = S_NO_AUTO_2 | S_FINAL_DIM; // don't use secondary stress
  721. tr->langopts.param[LOPT_IT_LENGTHEN] = 1; // remove lengthen indicator from unstressed syllables
  722. tr->langopts.length_mods0 = tr->langopts.length_mods; // don't lengthen vowels in the last syllable
  723. tr->langopts.accents = 2; // Say "Capital" after the letter.
  724. tr->langopts.numbers = NUM_SINGLE_STRESS | NUM_DECIMAL_COMMA | NUM_ALLOW_SPACE | NUM_OMIT_1_HUNDRED | NUM_NOPAUSE | NUM_ROMAN | NUM_ROMAN_CAPITALS | NUM_ROMAN_AFTER | NUM_VIGESIMAL | NUM_DFRACTION_4;
  725. SetLetterVowel(tr, 'y');
  726. }
  727. break;
  728. case L('g', 'a'): // irish
  729. case L('g', 'd'): // scots gaelic
  730. {
  731. tr->langopts.stress_rule = STRESSPOSN_1L;
  732. tr->langopts.stress_flags = S_NO_AUTO_2; // don't use secondary stress
  733. tr->langopts.numbers = NUM_OMIT_1_HUNDRED | NUM_OMIT_1_THOUSAND;
  734. tr->langopts.accents = 2; // 'capital' after letter name
  735. tr->langopts.param[LOPT_UNPRONOUNCABLE] = 3; // don't count apostrophe
  736. tr->langopts.param[LOPT_IT_LENGTHEN] = 1; // remove [:] phoneme from non-stressed syllables (Lang=gd)
  737. }
  738. break;
  739. case L('g','n'): // guarani
  740. {
  741. tr->langopts.stress_rule = STRESSPOSN_1R; // stress on final syllable
  742. tr->langopts.length_mods0 = tr->langopts.length_mods; // don't lengthen vowels in the last syllable
  743. }
  744. break;
  745. case L('h', 'i'): // Hindi
  746. case L('n', 'e'): // Nepali
  747. case L('o', 'r'): // Oriya
  748. case L('p', 'a'): // Punjabi
  749. case L('g', 'u'): // Gujarati
  750. case L('m', 'r'): // Marathi
  751. {
  752. static const short stress_lengths_hi[8] = { 190, 190, 210, 210, 0, 0, 230, 250 };
  753. static const unsigned char stress_amps_hi[8] = { 17, 14, 20, 19, 20, 22, 22, 21 };
  754. SetupTranslator(tr, stress_lengths_hi, stress_amps_hi);
  755. tr->encoding = ESPEAKNG_ENCODING_ISCII;
  756. tr->langopts.length_mods0 = tr->langopts.length_mods; // don't lengthen vowels in the last syllable
  757. tr->langopts.stress_rule = 6; // stress on last heaviest syllable, excluding final syllable
  758. tr->langopts.stress_flags = S_MID_DIM | S_FINAL_DIM; // use 'diminished' for unstressed final syllable
  759. tr->langopts.numbers = NUM_SWAP_TENS;
  760. tr->langopts.break_numbers = 0x14aa8; // for languages which have numbers for 100,000 and 100,00,000, eg Hindi
  761. tr->letter_bits_offset = OFFSET_DEVANAGARI;
  762. if (name2 == L('p', 'a'))
  763. tr->letter_bits_offset = OFFSET_GURMUKHI;
  764. else if (name2 == L('g', 'u')) {
  765. SetupTranslator(tr, stress_lengths_equal, stress_amps_equal);
  766. tr->letter_bits_offset = OFFSET_GUJARATI;
  767. tr->langopts.stress_rule = STRESSPOSN_2R;
  768. } else if (name2 == L('n', 'e')) {
  769. SetupTranslator(tr, stress_lengths_equal, stress_amps_equal);
  770. tr->langopts.break_numbers = 0x2aaaa8;
  771. tr->langopts.max_digits = 22;
  772. tr->langopts.numbers2 |= NUM2_ENGLISH_NUMERALS;
  773. } else if (name2 == L('o', 'r'))
  774. tr->letter_bits_offset = OFFSET_ORIYA;
  775. SetIndicLetters(tr);
  776. }
  777. break;
  778. case L('h', 'r'): // Croatian
  779. case L('b', 's'): // Bosnian
  780. case L('s', 'r'): // Serbian
  781. {
  782. static const unsigned char stress_amps_hr[8] = { 17, 17, 20, 20, 20, 22, 22, 21 };
  783. static const short stress_lengths_hr[8] = { 180, 160, 200, 200, 0, 0, 220, 230 };
  784. static const short stress_lengths_sr[8] = { 160, 150, 200, 200, 0, 0, 250, 260 };
  785. strcpy(tr->dictionary_name, "hbs");
  786. if (name2 == L('s', 'r'))
  787. SetupTranslator(tr, stress_lengths_sr, stress_amps_hr);
  788. else
  789. SetupTranslator(tr, stress_lengths_hr, stress_amps_hr);
  790. tr->encoding = ESPEAKNG_ENCODING_ISO_8859_2;
  791. tr->langopts.stress_rule = STRESSPOSN_1L;
  792. tr->langopts.stress_flags = S_FINAL_NO_2;
  793. tr->langopts.param[LOPT_REGRESSIVE_VOICING] = 0x3;
  794. tr->langopts.max_initial_consonants = 5;
  795. tr->langopts.spelling_stress = 1;
  796. tr->langopts.accents = 1;
  797. tr->langopts.numbers = NUM_SINGLE_STRESS | NUM_HUNDRED_AND | NUM_OMIT_1_HUNDRED | NUM_DECIMAL_COMMA | NUM_THOUS_SPACE | NUM_DFRACTION_2 | NUM_ROMAN_CAPITALS;
  798. tr->langopts.numbers2 = 0xa + NUM2_THOUSANDS_VAR5; // variant numbers before thousands,milliards
  799. tr->langopts.replace_chars = replace_cyrillic_latin;
  800. tr->langopts.our_alphabet = OFFSET_CYRILLIC; // don't say "cyrillic" before letter names
  801. SetLetterVowel(tr, 'y');
  802. SetLetterVowel(tr, 'r');
  803. }
  804. break;
  805. case L('h', 't'): // Haitian Creole
  806. tr->langopts.stress_rule = STRESSPOSN_1R; // stress on final syllable
  807. tr->langopts.stress_flags = S_NO_AUTO_2 | S_FINAL_DIM; // don't use secondary stress
  808. tr->langopts.numbers = NUM_SINGLE_STRESS | NUM_OMIT_1_HUNDRED | NUM_NOPAUSE | NUM_ROMAN | NUM_VIGESIMAL | NUM_DFRACTION_4;
  809. break;
  810. case L('h', 'u'): // Hungarian
  811. {
  812. static const unsigned char stress_amps_hu[8] = { 17, 17, 19, 19, 20, 22, 22, 21 };
  813. static const short stress_lengths_hu[8] = { 185, 195, 195, 190, 0, 0, 210, 220 };
  814. SetupTranslator(tr, stress_lengths_hu, stress_amps_hu);
  815. tr->encoding = ESPEAKNG_ENCODING_ISO_8859_2;
  816. tr->langopts.vowel_pause = 0x20;
  817. tr->langopts.stress_rule = STRESSPOSN_1L;
  818. tr->langopts.stress_flags = S_FINAL_DIM_ONLY | S_FINAL_NO_2 | S_NO_AUTO_2 | 0x8000 | S_HYPEN_UNSTRESS;
  819. tr->langopts.unstressed_wd1 = 2;
  820. tr->langopts.param[LOPT_IT_DOUBLING] = 1;
  821. tr->langopts.param[LOPT_ANNOUNCE_PUNCT] = 2; // don't break clause before announcing . ? !
  822. tr->langopts.numbers = NUM_DFRACTION_5 | NUM_ALLOW_SPACE | NUM_ROMAN | NUM_ROMAN_ORDINAL | NUM_ROMAN_CAPITALS | NUM_ORDINAL_DOT | NUM_OMIT_1_HUNDRED | NUM_OMIT_1_THOUSAND;
  823. tr->langopts.thousands_sep = ' '; // don't allow dot as thousands separator
  824. tr->langopts.decimal_sep = ',';
  825. tr->langopts.max_roman = 899;
  826. tr->langopts.min_roman = 1;
  827. SetLetterVowel(tr, 'y');
  828. tr->langopts.spelling_stress = 1;
  829. SetLengthMods(tr, 3); // all equal
  830. }
  831. break;
  832. case L('h', 'y'): // Armenian
  833. {
  834. static const short stress_lengths_hy[8] = { 250, 200, 250, 250, 0, 0, 250, 250 };
  835. static const char hy_vowels[] = { 0x31, 0x35, 0x37, 0x38, 0x3b, 0x48, 0x55, 0 };
  836. static const char hy_consonants[] = {
  837. 0x32, 0x33, 0x34, 0x36, 0x39, 0x3a, 0x3c, 0x3d, 0x3e, 0x3f, 0x40, 0x41, 0x42, 0x43, 0x44,
  838. 0x46, 0x47, 0x49, 0x4a, 0x4b, 0x4c, 0x4d, 0x4e, 0x4f, 0x50, 0x51, 0x52, 0x53, 0x54, 0x56, 0
  839. };
  840. static const char hy_consonants2[] = { 0x45, 0 };
  841. SetupTranslator(tr, stress_lengths_hy, NULL);
  842. tr->langopts.stress_rule = STRESSPOSN_1R; // default stress on final syllable
  843. tr->letter_bits_offset = OFFSET_ARMENIAN;
  844. memset(tr->letter_bits, 0, sizeof(tr->letter_bits));
  845. SetLetterBits(tr, LETTERGP_A, hy_vowels);
  846. SetLetterBits(tr, LETTERGP_VOWEL2, hy_vowels);
  847. SetLetterBits(tr, LETTERGP_B, hy_consonants); // not including 'j'
  848. SetLetterBits(tr, LETTERGP_C, hy_consonants);
  849. SetLetterBits(tr, LETTERGP_C, hy_consonants2); // add 'j'
  850. tr->langopts.max_initial_consonants = 6;
  851. tr->langopts.numbers = NUM_DECIMAL_COMMA | NUM_ALLOW_SPACE | NUM_OMIT_1_HUNDRED;
  852. }
  853. break;
  854. case L('i', 'd'): // Indonesian
  855. case L('m', 's'): // Malay
  856. {
  857. static const short stress_lengths_id[8] = { 160, 200, 180, 180, 0, 0, 220, 240 };
  858. static const unsigned char stress_amps_id[8] = { 16, 18, 18, 18, 20, 22, 22, 21 };
  859. SetupTranslator(tr, stress_lengths_id, stress_amps_id);
  860. tr->langopts.stress_rule = STRESSPOSN_2R;
  861. tr->langopts.numbers = NUM_DECIMAL_COMMA | NUM_ALLOW_SPACE | NUM_ROMAN;
  862. tr->langopts.stress_flags = S_FINAL_DIM_ONLY | S_FINAL_NO_2;
  863. tr->langopts.accents = 2; // "capital" after letter name
  864. }
  865. break;
  866. case L('i', 's'): // Icelandic
  867. {
  868. static const short stress_lengths_is[8] = { 180, 160, 200, 200, 0, 0, 240, 250 };
  869. static const wchar_t is_lettergroup_B[] = { 'c', 'f', 'h', 'k', 'p', 't', 'x', 0xfe, 0 }; // voiceless conants, including 'þ' ?? 's'
  870. SetupTranslator(tr, stress_lengths_is, NULL);
  871. tr->langopts.stress_rule = STRESSPOSN_1L;
  872. tr->langopts.stress_flags = S_FINAL_NO_2;
  873. tr->langopts.param[LOPT_IT_LENGTHEN] = 0x11; // remove lengthen indicator from unstressed vowels
  874. tr->langopts.param[LOPT_REDUCE] = 2;
  875. ResetLetterBits(tr, 0x18);
  876. SetLetterBits(tr, 4, "kpst"); // Letter group F
  877. SetLetterBits(tr, 3, "jvr"); // Letter group H
  878. tr->letter_groups[1] = is_lettergroup_B;
  879. SetLetterVowel(tr, 'y');
  880. tr->langopts.numbers = NUM_DECIMAL_COMMA | NUM_SINGLE_AND | NUM_HUNDRED_AND | NUM_AND_UNITS | NUM_1900;
  881. tr->langopts.numbers2 = 0x2;
  882. }
  883. break;
  884. case L('i', 't'): // Italian
  885. {
  886. static const short stress_lengths_it[8] = { 160, 140, 150, 165, 0, 0, 218, 305 };
  887. static const unsigned char stress_amps_it[8] = { 17, 15, 18, 16, 20, 22, 22, 22 };
  888. SetupTranslator(tr, stress_lengths_it, stress_amps_it);
  889. tr->langopts.length_mods0 = tr->langopts.length_mods; // don't lengthen vowels in the last syllable
  890. tr->langopts.stress_rule = STRESSPOSN_2R;
  891. tr->langopts.stress_flags = S_NO_AUTO_2 | S_FINAL_DIM_ONLY | S_PRIORITY_STRESS;
  892. tr->langopts.vowel_pause = 1;
  893. tr->langopts.unstressed_wd1 = 0;
  894. tr->langopts.unstressed_wd2 = 2;
  895. tr->langopts.param[LOPT_IT_LENGTHEN] = 2; // remove lengthen indicator from unstressed or non-penultimate syllables
  896. tr->langopts.param[LOPT_IT_DOUBLING] = 1; // double the first consonant if the previous word ends in a stressed vowel (changed to =1, 23.01.2014 - only use if prev.word has $double)
  897. tr->langopts.param[LOPT_SONORANT_MIN] = 130; // limit the shortening of sonorants before short vowels
  898. tr->langopts.param[LOPT_REDUCE] = 1; // reduce vowels even if phonemes are specified in it_list
  899. tr->langopts.param[LOPT_ALT] = 2; // call ApplySpecialAttributes2() if a word has $alt or $alt2
  900. tr->langopts.numbers = NUM_SINGLE_VOWEL | NUM_OMIT_1_HUNDRED |NUM_DECIMAL_COMMA | NUM_DFRACTION_1 | NUM_ROMAN | NUM_ROMAN_CAPITALS | NUM_ROMAN_ORDINAL;
  901. tr->langopts.numbers2 = NUM2_NO_TEEN_ORDINALS;
  902. tr->langopts.roman_suffix = utf8_ordinal;
  903. tr->langopts.accents = 2; // Say "Capital" after the letter.
  904. SetLetterVowel(tr, 'y');
  905. }
  906. break;
  907. case L_jbo: // Lojban
  908. {
  909. static const short stress_lengths_jbo[8] = { 145, 145, 170, 160, 0, 0, 330, 350 };
  910. static const wchar_t jbo_punct_within_word[] = { '.', ',', '\'', 0x2c8, 0 }; // allow period and comma within a word, also stress marker (from LOPT_CAPS_IN_WORD)
  911. SetupTranslator(tr, stress_lengths_jbo, NULL);
  912. tr->langopts.stress_rule = STRESSPOSN_2R;
  913. tr->langopts.vowel_pause = 0x20c; // pause before a word which starts with a vowel, or after a word which ends in a consonant
  914. tr->punct_within_word = jbo_punct_within_word;
  915. tr->langopts.param[LOPT_CAPS_IN_WORD] = 2; // capitals indicate stressed syllables
  916. SetLetterVowel(tr, 'y');
  917. tr->langopts.max_lengthmod = 368;
  918. }
  919. break;
  920. case L('k', 'a'): // Georgian
  921. {
  922. // character codes offset by 0x1080
  923. static const char ka_vowels[] = { 0x30, 0x34, 0x38, 0x3d, 0x43, 0x55, 0x57, 0 };
  924. static const char ka_consonants[] =
  925. { 0x31, 0x32, 0x33, 0x35, 0x36, 0x37, 0x39, 0x3a, 0x3b, 0x3c, 0x3e, 0x3f, 0x40, 0x41, 0x42, 0x44,
  926. 0x45, 0x46, 0x47, 0x48, 0x49, 0x4a, 0x4b, 0x4c, 0x4d, 0x4e, 0x4f, 0x50, 0x51, 0x52, 0x53, 0x54, 0x56, 0 };
  927. SetupTranslator(tr, stress_lengths_ta, stress_amps_ta);
  928. memset(tr->letter_bits, 0, sizeof(tr->letter_bits));
  929. SetLetterBits(tr, LETTERGP_A, ka_vowels);
  930. SetLetterBits(tr, LETTERGP_C, ka_consonants);
  931. SetLetterBits(tr, LETTERGP_VOWEL2, ka_vowels);
  932. tr->langopts.stress_rule = STRESSPOSN_1L;
  933. tr->langopts.stress_flags = S_FINAL_NO_2;
  934. tr->letter_bits_offset = OFFSET_GEORGIAN;
  935. tr->langopts.max_initial_consonants = 7;
  936. tr->langopts.numbers = NUM_VIGESIMAL | NUM_AND_UNITS | NUM_OMIT_1_HUNDRED |NUM_OMIT_1_THOUSAND | NUM_DFRACTION_5 | NUM_ROMAN;
  937. tr->langopts.alt_alphabet = OFFSET_CYRILLIC;
  938. tr->langopts.alt_alphabet_lang = L('r', 'u');
  939. }
  940. break;
  941. case L('k', 'k'): // Kazakh
  942. {
  943. static const unsigned char stress_amps_tr[8] = { 18, 16, 20, 21, 20, 21, 21, 20 };
  944. static const short stress_lengths_tr[8] = { 190, 180, 230, 230, 0, 0, 250, 250 };
  945. tr->letter_bits_offset = OFFSET_CYRILLIC;
  946. memset(tr->letter_bits, 0, sizeof(tr->letter_bits));
  947. SetLetterBits(tr, LETTERGP_A, (char *)ru_vowels);
  948. SetLetterBits(tr, LETTERGP_C, (char *)ru_consonants);
  949. SetLetterBits(tr, LETTERGP_VOWEL2, (char *)ru_vowels);
  950. SetupTranslator(tr, stress_lengths_tr, stress_amps_tr);
  951. tr->langopts.stress_rule = 7; // stress on the last syllable, before any explicitly unstressed syllable
  952. tr->langopts.stress_flags = S_NO_AUTO_2 + S_NO_EOC_LENGTHEN; // no automatic secondary stress, don't lengthen at end-of-clause
  953. tr->langopts.lengthen_tonic = 0;
  954. tr->langopts.param[LOPT_SUFFIX] = 1;
  955. tr->langopts.numbers = NUM_OMIT_1_HUNDRED | NUM_DFRACTION_6;
  956. tr->langopts.max_initial_consonants = 2;
  957. SetLengthMods(tr, 3); // all equal
  958. }
  959. break;
  960. case L('k', 'l'): // Greenlandic
  961. {
  962. SetupTranslator(tr, stress_lengths_equal, stress_amps_equal);
  963. tr->langopts.stress_rule = 12;
  964. tr->langopts.stress_flags = S_NO_AUTO_2;
  965. tr->langopts.numbers = NUM_DECIMAL_COMMA | NUM_SWAP_TENS | NUM_HUNDRED_AND | NUM_OMIT_1_HUNDRED | NUM_ORDINAL_DOT | NUM_1900 | NUM_ROMAN | NUM_ROMAN_CAPITALS | NUM_ROMAN_ORDINAL;
  966. }
  967. break;
  968. case L('k', 'o'): // Korean, TEST
  969. {
  970. static const char ko_ivowels[] = { 0x63, 0x64, 0x67, 0x68, 0x6d, 0x72, 0x74, 0x75, 0 }; // y and i vowels
  971. static const unsigned char ko_voiced[] = { 0x02, 0x05, 0x06, 0xab, 0xaf, 0xb7, 0xbc, 0 }; // voiced consonants, l,m,n,N
  972. tr->letter_bits_offset = OFFSET_KOREAN;
  973. tr->langopts.our_alphabet = 0xa700;
  974. memset(tr->letter_bits, 0, sizeof(tr->letter_bits));
  975. SetLetterBitsRange(tr, LETTERGP_A, 0x61, 0x75);
  976. SetLetterBits(tr, LETTERGP_Y, ko_ivowels);
  977. SetLetterBits(tr, LETTERGP_G, (const char *)ko_voiced);
  978. tr->langopts.stress_rule = 8; // ?? 1st syllable if it is heavy, else 2nd syllable
  979. tr->langopts.param[LOPT_UNPRONOUNCABLE] = 1; // disable check for unpronouncable words
  980. tr->langopts.numbers = NUM_OMIT_1_HUNDRED;
  981. tr->langopts.numbers2 = NUM2_MYRIADS;
  982. tr->langopts.break_numbers = 0x1111110;
  983. tr->langopts.max_digits = 20;
  984. }
  985. break;
  986. case L('k', 'u'): // Kurdish
  987. {
  988. static const unsigned char stress_amps_ku[8] = { 18, 18, 20, 20, 20, 22, 22, 21 };
  989. static const short stress_lengths_ku[8] = { 180, 180, 190, 180, 0, 0, 230, 240 };
  990. SetupTranslator(tr, stress_lengths_ku, stress_amps_ku);
  991. tr->encoding = ESPEAKNG_ENCODING_ISO_8859_9;
  992. tr->langopts.stress_rule = 7; // stress on the last syllable, before any explicitly unstressed syllable
  993. tr->langopts.numbers = NUM_HUNDRED_AND | NUM_AND_UNITS | NUM_OMIT_1_HUNDRED | NUM_AND_HUNDRED;
  994. tr->langopts.max_initial_consonants = 2;
  995. }
  996. break;
  997. case L('k', 'y'): // Kyrgyx
  998. tr->langopts.numbers = 1;
  999. break;
  1000. case L('l', 'a'): // Latin
  1001. {
  1002. tr->encoding = ESPEAKNG_ENCODING_ISO_8859_4; // includes a,e,i,o,u-macron
  1003. tr->langopts.stress_rule = STRESSPOSN_2R;
  1004. tr->langopts.stress_flags = S_NO_AUTO_2;
  1005. tr->langopts.unstressed_wd1 = 0;
  1006. tr->langopts.unstressed_wd2 = 2;
  1007. tr->langopts.param[LOPT_DIERESES] = 1;
  1008. tr->langopts.numbers = NUM_ROMAN;
  1009. tr->langopts.max_roman = 5000;
  1010. }
  1011. break;
  1012. case L('l', 't'): // Lithuanian
  1013. {
  1014. tr->encoding = ESPEAKNG_ENCODING_ISO_8859_4;
  1015. tr->langopts.stress_rule = STRESSPOSN_2R;
  1016. tr->langopts.stress_flags = S_NO_AUTO_2;
  1017. tr->langopts.unstressed_wd1 = 0;
  1018. tr->langopts.unstressed_wd2 = 2;
  1019. tr->langopts.param[LOPT_DIERESES] = 1;
  1020. tr->langopts.numbers = NUM_DECIMAL_COMMA | NUM_OMIT_1_HUNDRED | NUM_DFRACTION_4 | NUM_ORDINAL_DOT;
  1021. tr->langopts.numbers2 = NUM2_THOUSANDS_VAR4;
  1022. tr->langopts.max_roman = 5000;
  1023. }
  1024. break;
  1025. case L('l', 'v'): // latvian
  1026. {
  1027. static const unsigned char stress_amps_lv[8] = { 14, 10, 10, 8, 0, 0, 20, 15 };
  1028. static const short stress_lengths_lv[8] = { 180, 180, 180, 160, 0, 0, 230, 180 };
  1029. SetupTranslator(tr, stress_lengths_lv, stress_amps_lv);
  1030. tr->langopts.stress_rule = STRESSPOSN_1L;
  1031. tr->langopts.spelling_stress = 1;
  1032. tr->encoding = ESPEAKNG_ENCODING_ISO_8859_4;
  1033. tr->langopts.numbers = NUM_DECIMAL_COMMA | NUM_OMIT_1_HUNDRED | NUM_DFRACTION_4 | NUM_ORDINAL_DOT;
  1034. tr->langopts.stress_flags = S_NO_AUTO_2 | S_FINAL_DIM | S_FINAL_DIM_ONLY | S_EO_CLAUSE1;
  1035. }
  1036. break;
  1037. case L('m', 'k'): // Macedonian
  1038. {
  1039. static wchar_t vowels_cyrillic[] = {
  1040. // also include 'р' [R]
  1041. 0x440, 0x430, 0x435, 0x438, 0x439, 0x43e, 0x443, 0x44b, 0x44d,
  1042. 0x44e, 0x44f, 0x450, 0x451, 0x456, 0x457, 0x45d, 0x45e, 0
  1043. };
  1044. static const unsigned char stress_amps_mk[8] = { 17, 17, 20, 20, 20, 22, 22, 21 };
  1045. static const short stress_lengths_mk[8] = { 180, 160, 200, 200, 0, 0, 220, 230 };
  1046. SetupTranslator(tr, stress_lengths_mk, stress_amps_mk);
  1047. tr->encoding = ESPEAKNG_ENCODING_ISO_8859_5;
  1048. tr->letter_groups[0] = tr->letter_groups[7] = vowels_cyrillic;
  1049. tr->letter_bits_offset = OFFSET_CYRILLIC;
  1050. tr->langopts.stress_rule = STRESSPOSN_3R; // antipenultimate
  1051. tr->langopts.numbers = NUM_DECIMAL_COMMA | NUM_AND_UNITS | NUM_OMIT_1_HUNDRED | NUM_OMIT_1_THOUSAND | NUM_DFRACTION_2;
  1052. tr->langopts.numbers2 = 0x8a; // variant numbers before thousands,milliards
  1053. }
  1054. break;
  1055. case L('m', 't'): // Maltese
  1056. {
  1057. tr->encoding = ESPEAKNG_ENCODING_ISO_8859_3;
  1058. tr->langopts.param[LOPT_REGRESSIVE_VOICING] = 0x100; // devoice at end of word
  1059. tr->langopts.stress_rule = STRESSPOSN_2R; // penultimate
  1060. tr->langopts.numbers = 1;
  1061. }
  1062. break;
  1063. case L('n', 'l'): // Dutch
  1064. {
  1065. static const short stress_lengths_nl[8] = { 160, 135, 210, 210, 0, 0, 260, 280 };
  1066. tr->langopts.stress_rule = STRESSPOSN_1L;
  1067. tr->langopts.vowel_pause = 0x30; // ??
  1068. tr->langopts.param[LOPT_DIERESES] = 1;
  1069. tr->langopts.param[LOPT_PREFIXES] = 1;
  1070. tr->langopts.param[LOPT_REGRESSIVE_VOICING] = 0x100; // devoice at end of word
  1071. SetLetterVowel(tr, 'y');
  1072. tr->langopts.numbers = NUM_DECIMAL_COMMA | NUM_SWAP_TENS | NUM_OMIT_1_HUNDRED | NUM_OMIT_1_THOUSAND | NUM_ALLOW_SPACE | NUM_1900 | NUM_ORDINAL_DOT;
  1073. tr->langopts.ordinal_indicator = "e";
  1074. tr->langopts.stress_flags = S_FIRST_PRIMARY;
  1075. memcpy(tr->stress_lengths, stress_lengths_nl, sizeof(tr->stress_lengths));
  1076. }
  1077. break;
  1078. case L('n', 'o'): // Norwegian
  1079. {
  1080. static const short stress_lengths_no[8] = { 160, 140, 200, 200, 0, 0, 220, 230 };
  1081. SetupTranslator(tr, stress_lengths_no, NULL);
  1082. tr->langopts.stress_rule = STRESSPOSN_1L;
  1083. SetLetterVowel(tr, 'y');
  1084. tr->langopts.numbers = NUM_DECIMAL_COMMA | NUM_HUNDRED_AND | NUM_ALLOW_SPACE | NUM_1900 | NUM_ORDINAL_DOT;
  1085. }
  1086. break;
  1087. case L('o', 'm'): // Oromo
  1088. {
  1089. static const unsigned char stress_amps_om[] = { 18, 15, 20, 20, 20, 22, 22, 22 };
  1090. static const short stress_lengths_om[8] = { 200, 200, 200, 200, 0, 0, 200, 200 };
  1091. SetupTranslator(tr, stress_lengths_om, stress_amps_om);
  1092. tr->langopts.stress_rule = STRESSPOSN_2R;
  1093. tr->langopts.stress_flags = S_FINAL_DIM_ONLY | S_FINAL_NO_2 | 0x80000;
  1094. tr->langopts.numbers = NUM_OMIT_1_HUNDRED | NUM_HUNDRED_AND;
  1095. tr->langopts.numbers2 = 0x200; // say "thousands" before its number
  1096. }
  1097. break;
  1098. case L('p', 'l'): // Polish
  1099. {
  1100. static const short stress_lengths_pl[8] = { 160, 190, 175, 175, 0, 0, 200, 210 };
  1101. static const unsigned char stress_amps_pl[8] = { 17, 13, 19, 19, 20, 22, 22, 21 }; // 'diminished' is used to mark a quieter, final unstressed syllable
  1102. SetupTranslator(tr, stress_lengths_pl, stress_amps_pl);
  1103. tr->encoding = ESPEAKNG_ENCODING_ISO_8859_2;
  1104. tr->langopts.stress_rule = STRESSPOSN_2R;
  1105. tr->langopts.stress_flags = S_FINAL_DIM_ONLY; // mark unstressed final syllables as diminished
  1106. tr->langopts.param[LOPT_REGRESSIVE_VOICING] = 0x9;
  1107. tr->langopts.max_initial_consonants = 7; // for example: wchrzczony :)
  1108. tr->langopts.numbers = NUM_DECIMAL_COMMA | NUM_ALLOW_SPACE | NUM_DFRACTION_2;
  1109. tr->langopts.numbers2 = NUM2_THOUSANDS_VAR3;
  1110. tr->langopts.param[LOPT_COMBINE_WORDS] = 4 + 0x100; // combine 'nie' (marked with $alt2) with some 1-syllable (and 2-syllable) words (marked with $alt)
  1111. SetLetterVowel(tr, 'y');
  1112. }
  1113. break;
  1114. case L('p', 't'): // Portuguese
  1115. {
  1116. static const short stress_lengths_pt[8] = { 170, 115, 210, 240, 0, 0, 260, 280 };
  1117. static const unsigned char stress_amps_pt[8] = { 16, 11, 19, 21, 20, 22, 22, 21 }; // 'diminished' is used to mark a quieter, final unstressed syllable
  1118. SetupTranslator(tr, stress_lengths_pt, stress_amps_pt);
  1119. tr->langopts.length_mods0 = tr->langopts.length_mods; // don't lengthen vowels in the last syllable
  1120. tr->langopts.stress_rule = STRESSPOSN_1R; // stress on final syllable
  1121. tr->langopts.stress_flags = S_FINAL_DIM_ONLY | S_FINAL_NO_2 | S_INITIAL_2 | S_PRIORITY_STRESS;
  1122. tr->langopts.numbers = NUM_DECIMAL_COMMA | NUM_DFRACTION_2 | NUM_HUNDRED_AND | NUM_AND_UNITS | NUM_ROMAN_CAPITALS;
  1123. tr->langopts.numbers2 = NUM2_MULTIPLE_ORDINAL | NUM2_NO_TEEN_ORDINALS | NUM2_ORDINAL_NO_AND;
  1124. tr->langopts.max_roman = 5000;
  1125. SetLetterVowel(tr, 'y');
  1126. ResetLetterBits(tr, 0x2);
  1127. SetLetterBits(tr, 1, "bcdfgjkmnpqstvxz"); // B hard consonants, excluding h,l,r,w,y
  1128. tr->langopts.param[LOPT_ALT] = 2; // call ApplySpecialAttributes2() if a word has $alt or $alt2
  1129. tr->langopts.accents = 2; // 'capital' after letter name
  1130. }
  1131. break;
  1132. case L('r', 'o'): // Romanian
  1133. {
  1134. static const short stress_lengths_ro[8] = { 170, 170, 180, 180, 0, 0, 240, 260 };
  1135. static const unsigned char stress_amps_ro[8] = { 15, 13, 18, 18, 20, 22, 22, 21 };
  1136. SetupTranslator(tr, stress_lengths_ro, stress_amps_ro);
  1137. tr->langopts.stress_rule = STRESSPOSN_1R;
  1138. tr->langopts.stress_flags = S_FINAL_VOWEL_UNSTRESSED | S_FINAL_DIM_ONLY;
  1139. tr->encoding = ESPEAKNG_ENCODING_ISO_8859_2;
  1140. tr->langopts.numbers = NUM_DECIMAL_COMMA | NUM_ALLOW_SPACE | NUM_DFRACTION_3 | NUM_AND_UNITS | NUM_ROMAN;
  1141. tr->langopts.numbers2 = 0x1e; // variant numbers before all thousandplex
  1142. }
  1143. break;
  1144. case L('r', 'u'): // Russian
  1145. Translator_Russian(tr);
  1146. break;
  1147. case L('r', 'w'): // Kiryarwanda
  1148. {
  1149. tr->langopts.stress_rule = STRESSPOSN_2R;
  1150. tr->langopts.stress_flags = S_FINAL_DIM_ONLY | S_FINAL_NO_2;
  1151. tr->langopts.length_mods0 = tr->langopts.length_mods; // don't lengthen vowels in the last syllable
  1152. tr->langopts.param[LOPT_UNPRONOUNCABLE] = 1; // disable check for unpronouncable words. Need to allow "bw'" prefix
  1153. tr->langopts.numbers = NUM_HUNDRED_AND | NUM_AND_UNITS | NUM_DFRACTION_2 | NUM_AND_HUNDRED;
  1154. tr->langopts.numbers2 = 0x200; // say "thousands" before its number
  1155. }
  1156. break;
  1157. case L('s', 'k'): // Slovak
  1158. case L('c', 's'): // Czech
  1159. {
  1160. static const char *sk_voiced = "bdgjlmnrvwzaeiouy";
  1161. SetupTranslator(tr, stress_lengths_sk, stress_amps_sk);
  1162. tr->encoding = ESPEAKNG_ENCODING_ISO_8859_2;
  1163. tr->langopts.stress_rule = STRESSPOSN_1L;
  1164. tr->langopts.stress_flags = S_FINAL_DIM_ONLY | S_FINAL_NO_2;
  1165. tr->langopts.param[LOPT_REGRESSIVE_VOICING] = 0x3;
  1166. tr->langopts.max_initial_consonants = 5;
  1167. tr->langopts.spelling_stress = 1;
  1168. tr->langopts.param[LOPT_COMBINE_WORDS] = 4; // combine some prepositions with the following word
  1169. tr->langopts.numbers = NUM_OMIT_1_HUNDRED | NUM_DFRACTION_2 | NUM_ROMAN;
  1170. tr->langopts.numbers2 = NUM2_THOUSANDS_VAR2;
  1171. tr->langopts.thousands_sep = STRESSPOSN_1L; // no thousands separator
  1172. tr->langopts.decimal_sep = ',';
  1173. if (name2 == L('c', 's'))
  1174. tr->langopts.numbers2 = 0x108; // variant numbers before milliards
  1175. SetLetterVowel(tr, 'y');
  1176. SetLetterVowel(tr, 'r');
  1177. ResetLetterBits(tr, 0x20);
  1178. SetLetterBits(tr, 5, sk_voiced);
  1179. }
  1180. break;
  1181. case L('s', 'i'): // Sinhala
  1182. {
  1183. SetupTranslator(tr, stress_lengths_ta, stress_amps_ta);
  1184. tr->langopts.length_mods0 = tr->langopts.length_mods; // don't lengthen vowels in the last syllable
  1185. tr->langopts.stress_rule = STRESSPOSN_1L;
  1186. tr->langopts.stress_flags = S_FINAL_DIM_ONLY | S_FINAL_NO_2;
  1187. tr->langopts.spelling_stress = 1;
  1188. tr->letter_bits_offset = OFFSET_SINHALA;
  1189. memset(tr->letter_bits, 0, sizeof(tr->letter_bits));
  1190. SetLetterBitsRange(tr, LETTERGP_A, 0x05, 0x16); // vowel letters
  1191. SetLetterBitsRange(tr, LETTERGP_A, 0x4a, 0x73); // + vowel signs, and virama
  1192. SetLetterBitsRange(tr, LETTERGP_B, 0x4a, 0x73); // vowel signs, and virama
  1193. SetLetterBitsRange(tr, LETTERGP_C, 0x1a, 0x46); // the main consonant range
  1194. tr->langopts.param[LOPT_UNPRONOUNCABLE] = 1; // disable check for unpronouncable words
  1195. tr->langopts.suffix_add_e = tr->letter_bits_offset + 0x4a; // virama
  1196. tr->langopts.numbers = NUM_OMIT_1_THOUSAND | NUM_SINGLE_STRESS_L | NUM_DFRACTION_7;
  1197. tr->langopts.numbers2 = NUM2_PERCENT_BEFORE;
  1198. tr->langopts.break_numbers = 0x14aa8; // for languages which have numbers for 100,000 and 100,00,000, eg Hindi
  1199. }
  1200. break;
  1201. case L('s', 'l'): // Slovenian
  1202. tr->encoding = ESPEAKNG_ENCODING_ISO_8859_2;
  1203. tr->langopts.stress_rule = STRESSPOSN_2R; // Temporary
  1204. tr->langopts.stress_flags = S_NO_AUTO_2;
  1205. tr->langopts.param[LOPT_REGRESSIVE_VOICING] = 0x103;
  1206. tr->langopts.param[LOPT_UNPRONOUNCABLE] = 0x76; // [v] don't count this character at start of word
  1207. tr->langopts.param[LOPT_ALT] = 2; // call ApplySpecialAttributes2() if a word has $alt or $alt2
  1208. tr->langopts.param[LOPT_IT_LENGTHEN] = 1; // remove lengthen indicator from unstressed syllables
  1209. tr->letter_bits[(int)'r'] |= 0x80; // add 'r' to letter group 7, vowels for Unpronouncable test
  1210. tr->langopts.numbers = NUM_DECIMAL_COMMA | NUM_ALLOW_SPACE | NUM_SWAP_TENS | NUM_OMIT_1_HUNDRED | NUM_DFRACTION_2 | NUM_ORDINAL_DOT | NUM_ROMAN;
  1211. tr->langopts.numbers2 = 0x100; // plural forms of millions etc
  1212. tr->langopts.thousands_sep = ' '; // don't allow dot as thousands separator
  1213. tr->langopts.replace_chars = replace_cyrillic_latin;
  1214. break;
  1215. case L('s', 'q'): // Albanian
  1216. {
  1217. static const short stress_lengths_sq[8] = { 150, 150, 180, 180, 0, 0, 300, 300 };
  1218. static const unsigned char stress_amps_sq[8] = { 16, 12, 16, 16, 20, 20, 21, 19 };
  1219. SetupTranslator(tr, stress_lengths_sq, stress_amps_sq);
  1220. tr->langopts.stress_rule = STRESSPOSN_1R;
  1221. tr->langopts.stress_flags = S_FINAL_DIM_ONLY | S_FINAL_NO_2 | S_FINAL_VOWEL_UNSTRESSED;
  1222. SetLetterVowel(tr, 'y');
  1223. tr->langopts.numbers = NUM_DECIMAL_COMMA | NUM_HUNDRED_AND | NUM_AND_UNITS | NUM_DFRACTION_4;
  1224. tr->langopts.accents = 2; // "capital" after letter name
  1225. }
  1226. break;
  1227. case L('s', 'v'): // Swedish
  1228. {
  1229. static const unsigned char stress_amps_sv[] = { 16, 16, 20, 20, 20, 22, 22, 21 };
  1230. static const short stress_lengths_sv[8] = { 160, 135, 220, 220, 0, 0, 250, 280 };
  1231. SetupTranslator(tr, stress_lengths_sv, stress_amps_sv);
  1232. tr->langopts.stress_rule = STRESSPOSN_1L;
  1233. SetLetterVowel(tr, 'y');
  1234. tr->langopts.numbers = NUM_SINGLE_STRESS | NUM_DECIMAL_COMMA | NUM_ALLOW_SPACE | NUM_1900;
  1235. tr->langopts.accents = 1;
  1236. }
  1237. break;
  1238. case L('s', 'w'): // Swahili
  1239. case L('t', 'n'): // Setswana
  1240. {
  1241. static const short stress_lengths_sw[8] = { 160, 170, 200, 200, 0, 0, 320, 340 };
  1242. static const unsigned char stress_amps_sw[] = { 16, 12, 19, 19, 20, 22, 22, 21 };
  1243. SetupTranslator(tr, stress_lengths_sw, stress_amps_sw);
  1244. tr->langopts.length_mods0 = tr->langopts.length_mods; // don't lengthen vowels in the last syllable
  1245. tr->langopts.vowel_pause = 1;
  1246. tr->langopts.stress_rule = STRESSPOSN_2R;
  1247. tr->langopts.stress_flags = S_FINAL_DIM_ONLY | S_FINAL_NO_2;
  1248. tr->langopts.max_initial_consonants = 4; // for example: mwngi
  1249. tr->langopts.numbers = NUM_AND_UNITS | NUM_HUNDRED_AND | NUM_SINGLE_AND | NUM_OMIT_1_HUNDRED;
  1250. tr->langopts.break_numbers = 0x49249268; // for languages which have numbers for 100,000 and 1,000,000
  1251. }
  1252. break;
  1253. case L('t', 'a'): // Tamil
  1254. case L('k', 'n'): // Kannada
  1255. case L('m', 'l'): // Malayalam
  1256. case L('t', 'e'): // Telugu
  1257. {
  1258. SetupTranslator(tr, stress_lengths_ta2, stress_amps_ta);
  1259. tr->langopts.length_mods0 = tr->langopts.length_mods; // don't lengthen vowels in the last syllable
  1260. tr->langopts.stress_rule = STRESSPOSN_1L;
  1261. tr->langopts.stress_flags = S_FINAL_DIM_ONLY | S_FINAL_NO_2; // use 'diminished' for unstressed final syllable
  1262. tr->langopts.spelling_stress = 1;
  1263. tr->langopts.break_numbers = 0x14a8; // 1000, 100,000 10,000,000
  1264. if (name2 == L('t', 'a')) {
  1265. SetupTranslator(tr, stress_lengths_ta, NULL);
  1266. tr->letter_bits_offset = OFFSET_TAMIL;
  1267. tr->langopts.numbers = NUM_OMIT_1_THOUSAND;
  1268. tr->langopts.numbers2 = NUM2_ORDINAL_AND_THOUSANDS;
  1269. tr->langopts.param[LOPT_WORD_MERGE] = 1; // don't break vowels betwen words
  1270. } else if (name2 == L('m', 'l')) {
  1271. static const short stress_lengths_ml[8] = { 180, 160, 240, 240, 0, 0, 260, 260 };
  1272. SetupTranslator(tr, stress_lengths_ml, stress_amps_equal);
  1273. tr->letter_bits_offset = OFFSET_MALAYALAM;
  1274. tr->langopts.numbers = NUM_OMIT_1_THOUSAND | NUM_OMIT_1_HUNDRED;
  1275. tr->langopts.numbers2 = NUM2_OMIT_1_HUNDRED_ONLY;
  1276. tr->langopts.stress_rule = 13; // 1st syllable, unless 1st vowel is short and 2nd is long
  1277. } else if (name2 == L('k', 'n')) {
  1278. tr->letter_bits_offset = OFFSET_KANNADA;
  1279. tr->langopts.numbers = 0x1;
  1280. } else if (name2 == L('t', 'e')) {
  1281. tr->letter_bits_offset = OFFSET_TELUGU;
  1282. tr->langopts.numbers = 0x1;
  1283. tr->langopts.numbers2 = NUM2_ORDINAL_DROP_VOWEL;
  1284. }
  1285. SetIndicLetters(tr); // call this after setting OFFSET_
  1286. SetLetterBitsRange(tr, LETTERGP_B, 0x4e, 0x4e); // chillu-virama (unofficial)
  1287. }
  1288. break;
  1289. case L('t', 'r'): // Turkish
  1290. case L('a', 'z'): // Azerbaijan
  1291. {
  1292. static const unsigned char stress_amps_tr[8] = { 18, 16, 20, 21, 20, 21, 21, 20 };
  1293. static const short stress_lengths_tr[8] = { 190, 180, 200, 230, 0, 0, 240, 250 };
  1294. SetupTranslator(tr, stress_lengths_tr, stress_amps_tr);
  1295. tr->encoding = ESPEAKNG_ENCODING_ISO_8859_9;
  1296. tr->langopts.stress_rule = 7; // stress on the last syllable, before any explicitly unstressed syllable
  1297. tr->langopts.stress_flags = S_NO_AUTO_2; // no automatic secondary stress
  1298. tr->langopts.dotless_i = 1;
  1299. tr->langopts.param[LOPT_SUFFIX] = 1;
  1300. if (name2 == L('a', 'z'))
  1301. tr->langopts.numbers = NUM_SINGLE_STRESS | NUM_DECIMAL_COMMA | NUM_ALLOW_SPACE | NUM_OMIT_1_HUNDRED | NUM_OMIT_1_THOUSAND | NUM_DFRACTION_2;
  1302. else
  1303. tr->langopts.numbers = NUM_SINGLE_STRESS | NUM_DECIMAL_COMMA | NUM_OMIT_1_HUNDRED | NUM_OMIT_1_THOUSAND | NUM_DFRACTION_2;
  1304. tr->langopts.max_initial_consonants = 2;
  1305. }
  1306. break;
  1307. case L('t', 't'): // Tatar
  1308. {
  1309. SetCyrillicLetters(tr);
  1310. SetupTranslator(tr, stress_lengths_fr, stress_amps_fr);
  1311. tr->langopts.stress_rule = STRESSPOSN_1R; // stress on final syllable
  1312. tr->langopts.stress_flags = S_NO_AUTO_2; // no automatic secondary stress
  1313. tr->langopts.numbers = NUM_SINGLE_STRESS | NUM_DECIMAL_COMMA | NUM_OMIT_1_HUNDRED | NUM_OMIT_1_THOUSAND | NUM_DFRACTION_4;
  1314. }
  1315. break;
  1316. case L('u', 'k'): // Ukrainian
  1317. {
  1318. SetCyrillicLetters(tr);
  1319. tr->langopts.param[LOPT_UNPRONOUNCABLE] = 0x432; // [v] don't count this character at start of word
  1320. }
  1321. break;
  1322. case L('u', 'r'): // Urdu
  1323. case L('s', 'd'): // Sindhi
  1324. {
  1325. tr->letter_bits_offset = OFFSET_ARABIC;
  1326. tr->langopts.param[LOPT_UNPRONOUNCABLE] = 1; // disable check for unpronouncable words
  1327. tr->langopts.numbers = NUM_SWAP_TENS;
  1328. tr->langopts.break_numbers = 0x52a8; // for languages which have numbers for 100,000 and 100,00,000, eg Hindi
  1329. }
  1330. break;
  1331. case L('v', 'i'): // Vietnamese
  1332. {
  1333. static const short stress_lengths_vi[8] = { 150, 150, 180, 180, 210, 230, 230, 240 };
  1334. static const unsigned char stress_amps_vi[] = { 16, 16, 16, 16, 22, 22, 22, 22 };
  1335. static wchar_t vowels_vi[] = {
  1336. 0x61, 0xe0, 0xe1, 0x1ea3, 0xe3, 0x1ea1, // a
  1337. 0x103, 0x1eb1, 0x1eaf, 0x1eb3, 0x1eb5, 0x1eb7, // ă
  1338. 0xe2, 0x1ea7, 0x1ea5, 0x1ea9, 0x1eab, 0x1ead, // â
  1339. 0x65, 0xe8, 0xe9, 0x1ebb, 0x1ebd, 0x1eb9, // e
  1340. 0xea, 0x1ec1, 0x1ebf, 0x1ec3, 0x1ec5, 0x1ec7, // i
  1341. 0x69, 0xec, 0xed, 0x1ec9, 0x129, 0x1ecb, // i
  1342. 0x6f, 0xf2, 0xf3, 0x1ecf, 0xf5, 0x1ecd, // o
  1343. 0xf4, 0x1ed3, 0x1ed1, 0x1ed5, 0x1ed7, 0x1ed9, // ô
  1344. 0x1a1, 0x1edd, 0x1edb, 0x1edf, 0x1ee1, 0x1ee3, // ơ
  1345. 0x75, 0xf9, 0xfa, 0x1ee7, 0x169, 0x1ee5, // u
  1346. 0x1b0, 0x1eeb, 0x1ee9, 0x1eed, 0x1eef, 0x1ef1, // ư
  1347. 0x79, 0x1ef3, 0xfd, 0x1ef7, 0x1ef9, 0x1ef5, // y
  1348. 0
  1349. };
  1350. SetupTranslator(tr, stress_lengths_vi, stress_amps_vi);
  1351. tr->langopts.length_mods0 = tr->langopts.length_mods; // don't lengthen vowels in the last syllable
  1352. tr->langopts.stress_rule = STRESSPOSN_1L;
  1353. tr->langopts.word_gap = 0x21; // length of a final vowel is less dependent on the next consonant, don't merge consonant with next word
  1354. tr->letter_groups[0] = tr->letter_groups[7] = vowels_vi;
  1355. tr->langopts.tone_language = 1; // Tone language, use CalcPitches_Tone() rather than CalcPitches()
  1356. tr->langopts.unstressed_wd1 = 2;
  1357. tr->langopts.numbers = NUM_DECIMAL_COMMA | NUM_HUNDRED_AND_DIGIT | NUM_DFRACTION_4 | NUM_ZERO_HUNDRED;
  1358. }
  1359. break;
  1360. case L('w', 'o'):
  1361. tr->langopts.stress_rule = STRESSPOSN_1L;
  1362. tr->langopts.numbers = NUM_AND_UNITS | NUM_HUNDRED_AND | NUM_OMIT_1_HUNDRED | NUM_OMIT_1_THOUSAND | NUM_SINGLE_STRESS;
  1363. break;
  1364. case L_cmn: // no break, just go to 'zh' case
  1365. case L_zhy: // just go to 'zh' case
  1366. case L('z','h'):
  1367. {
  1368. static const short stress_lengths_zh[8] = { 230, 150, 230, 230, 230, 0, 240, 250 }; // 1=tone5. end-of-sentence, 6=tone 1&4, 7=tone 2&3
  1369. static const unsigned char stress_amps_zh[] = { 22, 16, 22, 22, 22, 22, 22, 22 };
  1370. SetupTranslator(tr, stress_lengths_zh, stress_amps_zh);
  1371. tr->langopts.stress_rule = STRESSPOSN_1R; // stress on final syllable of a "word"
  1372. tr->langopts.stress_flags = S_NO_DIM; // don't automatically set diminished stress (may be set in the intonation module)
  1373. tr->langopts.vowel_pause = 0;
  1374. tr->langopts.tone_language = 1; // Tone language, use CalcPitches_Tone() rather than CalcPitches()
  1375. tr->langopts.length_mods0 = tr->langopts.length_mods; // don't lengthen vowels in the last syllable
  1376. tr->langopts.tone_numbers = 1; // a number after letters indicates a tone number (eg. pinyin or jyutping)
  1377. tr->langopts.ideographs = 1;
  1378. tr->langopts.our_alphabet = 0x3100;
  1379. tr->langopts.word_gap = 0x21; // length of a final vowel is less dependent on the next consonant, don't merge consonant with next word
  1380. if (name2 == L_zhy) {
  1381. tr->langopts.textmode = 1;
  1382. tr->langopts.listx = 1; // compile zh_listx after zh_list
  1383. tr->langopts.numbers = 1;
  1384. tr->langopts.numbers2 = NUM2_ZERO_TENS;
  1385. tr->langopts.break_numbers = 0x00018;
  1386. }
  1387. break;
  1388. }
  1389. default:
  1390. tr->langopts.param[LOPT_UNPRONOUNCABLE] = 1; // disable check for unpronouncable words
  1391. break;
  1392. }
  1393. tr->translator_name = name2;
  1394. ProcessLanguageOptions(&tr->langopts);
  1395. return tr;
  1396. }
  1397. void ProcessLanguageOptions(LANGUAGE_OPTIONS *langopts)
  1398. {
  1399. if (langopts->numbers & NUM_DECIMAL_COMMA) {
  1400. // use . and ; for thousands and decimal separators
  1401. langopts->thousands_sep = '.';
  1402. langopts->decimal_sep = ',';
  1403. }
  1404. if (langopts->numbers & NUM_THOUS_SPACE)
  1405. langopts->thousands_sep = 0; // don't allow thousands separator, except space
  1406. }
  1407. static void Translator_Russian(Translator *tr)
  1408. {
  1409. static const unsigned char stress_amps_ru[] = { 16, 16, 18, 18, 20, 24, 24, 22 };
  1410. static const short stress_lengths_ru[8] = { 150, 140, 220, 220, 0, 0, 260, 280 };
  1411. static const char ru_ivowels2[] = { 0x2c, 0x15, 0x18, 0x2e, 0x2f, 0 }; // add more vowels to letter group Y (iotated vowels & soft-sign)
  1412. SetupTranslator(tr, stress_lengths_ru, stress_amps_ru);
  1413. SetCyrillicLetters(tr);
  1414. SetLetterBits(tr, 6, ru_ivowels2);
  1415. tr->langopts.param[LOPT_UNPRONOUNCABLE] = 0x432; // [v] don't count this character at start of word
  1416. tr->langopts.param[LOPT_REGRESSIVE_VOICING] = 1;
  1417. tr->langopts.param[LOPT_REDUCE] = 2;
  1418. tr->langopts.stress_rule = 5;
  1419. tr->langopts.stress_flags = S_NO_AUTO_2;
  1420. tr->langopts.numbers = NUM_DECIMAL_COMMA | NUM_OMIT_1_HUNDRED;
  1421. tr->langopts.numbers2 = 0x2 + NUM2_THOUSANDS_VAR1; // variant numbers before thousands
  1422. tr->langopts.phoneme_change = 1;
  1423. tr->langopts.testing = 2;
  1424. }