3 * Library to deal with pinyin.
5 * Copyright (C) 2011 Peng Wu <alexepico@gmail.com>
7 * This program is free software; you can redistribute it and/or modify
8 * it under the terms of the GNU General Public License as published by
9 * the Free Software Foundation; either version 2 of the License, or
10 * (at your option) any later version.
12 * This program is distributed in the hope that it will be useful,
13 * but WITHOUT ANY WARRANTY; without even the implied warranty of
14 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
15 * GNU General Public License for more details.
17 * You should have received a copy of the GNU General Public License
18 * along with this program; if not, write to the Free Software
19 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
27 #include "pinyin_parser2.h"
28 #include "pinyin_phrase2.h"
29 #include "pinyin_custom2.h"
30 #include "chewing_key.h"
31 #include "pinyin_parser_table.h"
32 #include "double_pinyin_table.h"
33 #include "chewing_table.h"
36 using namespace pinyin;
38 static bool check_pinyin_options(guint32 options, const pinyin_index_item_t * item) {
39 guint32 flags = item->m_flags;
40 assert (flags & IS_PINYIN);
42 /* handle incomplete pinyin. */
43 if (flags & PINYIN_INCOMPLETE) {
44 if (!(options & PINYIN_INCOMPLETE))
48 /* handle correct pinyin, currently only one flag per item. */
49 flags &= PINYIN_CORRECT_ALL;
50 options &= PINYIN_CORRECT_ALL;
53 if ((flags & options) != flags)
60 static bool check_chewing_options(guint32 options, const chewing_index_item_t * item) {
61 guint32 flags = item->m_flags;
62 assert (flags & IS_CHEWING);
64 /* handle incomplete chewing. */
65 if (flags & CHEWING_INCOMPLETE) {
66 if (!(options & CHEWING_INCOMPLETE))
74 /* methods for Chewing Keys to access pinyin parser table. */
75 const char * ChewingKeyRest::get_pinyin_string(){
76 if (m_table_index == 0)
79 /* check end boundary. */
80 assert(m_table_index < G_N_ELEMENTS(content_table));
81 return content_table[m_table_index].m_pinyin_str;
84 const char * ChewingKeyRest::get_chewing_string(){
85 if (m_table_index == 0)
88 /* check end boundary. */
89 assert(m_table_index < G_N_ELEMENTS(content_table));
90 return content_table[m_table_index].m_chewing_str;
96 /* internal information for pinyin parsers. */
99 ChewingKeyRest m_key_rest;
113 /* Full Pinyin Parser */
114 FullPinyinParser2::FullPinyinParser2 (){
115 m_parse_steps = g_array_new(TRUE, FALSE, sizeof(parse_value_t));
118 const guint16 max_full_pinyin_length = 7; /* include tone. */
120 static bool compare_less_than(const pinyin_index_item_t & lhs,
121 const pinyin_index_item_t & rhs){
122 return 0 > strcmp(lhs.m_pinyin_input, rhs.m_pinyin_input);
125 bool FullPinyinParser2::parse_one_key (guint32 options, ChewingKey & key,
126 ChewingKeyRest & key_rest,
127 const char * pinyin, int len) const {
128 /* "'" are not accepted in parse_one_key. */
129 assert(NULL == strchr(pinyin, '\''));
130 gchar * input = g_strndup(pinyin, len);
132 guint16 tone = CHEWING_ZERO_TONE; guint16 tone_pos = 0;
133 guint16 parsed_len = len;
134 key = ChewingKey(); key_rest = ChewingKeyRest();
136 if (options & USE_TONE) {
137 /* find the tone in the last character. */
138 char chr = input[parsed_len - 1];
139 if ( '0' < chr && chr <= '5' ) {
142 tone_pos = parsed_len;
146 /* parse pinyin core staff here. */
147 pinyin_index_item_t item;
148 memset(&item, 0, sizeof(item));
150 /* Note: optimize here? */
151 for (; parsed_len >= len - 1; --parsed_len) {
152 input[parsed_len] = '\0';
153 item.m_pinyin_input = input;
154 std_lite::pair<const pinyin_index_item_t *,
155 const pinyin_index_item_t *> range;
156 range = std_lite::equal_range
157 (pinyin_index, pinyin_index + G_N_ELEMENTS(pinyin_index),
158 item, compare_less_than);
160 guint16 range_len = range.second - range.first;
161 assert (range_len <= 1);
162 if ( range_len == 1 ) {
163 const pinyin_index_item_t * index = range.first;
165 if (!check_pinyin_options(options, index))
168 key_rest.m_table_index = index->m_table_index;
169 key = content_table[key_rest.m_table_index].m_chewing_key;
174 if (options & USE_TONE) {
175 /* post processing tone. */
176 if ( parsed_len == tone_pos ) {
177 if (tone != CHEWING_ZERO_TONE) {
184 key_rest.m_raw_begin = 0; key_rest.m_raw_end = parsed_len;
186 return parsed_len == len;
190 int FullPinyinParser2::parse (guint32 options, ChewingKeyVector & keys,
191 ChewingKeyRestVector & key_rests,
192 const char *str, int len) const {
195 g_array_set_size(keys, 0);
196 g_array_set_size(key_rests, 0);
198 /* init m_parse_steps, and prepare dynamic programming. */
199 size_t step_len = len + 1;
200 g_array_set_size(m_parse_steps, 0);
202 for (i = 0; i < step_len; ++i) {
203 g_array_append_val(m_parse_steps, value);
206 size_t str_len = len; size_t next_sep = 0;
207 gchar * input = g_strndup(str, len);
208 parse_value_t * curstep = NULL, * nextstep = NULL;
210 for (i = 0; i < len; ) {
211 if (input[i] == '\'') {
212 curstep = &g_array_index(m_parse_steps, parse_value_t, i);
213 nextstep = &g_array_index(m_parse_steps, parse_value_t, i + 1);
215 /* propagate current step into next step. */
216 nextstep->m_key = ChewingKey();
217 nextstep->m_key_rest = ChewingKeyRest();
218 nextstep->m_num_keys = curstep->m_num_keys;
219 nextstep->m_parsed_len = curstep->m_parsed_len + 1;
220 nextstep->m_last_step = i;
225 /* forward to next "'" */
226 if ( 0 == next_sep ) {
228 for (k = i; k < len; ++k) {
229 if (input[k] == '\'')
236 /* dynamic programming here. */
237 for (size_t m = i; m < next_sep; ++m) {
238 curstep = &g_array_index(m_parse_steps, parse_value_t, m);
239 size_t try_len = std_lite::min
240 (m + max_full_pinyin_length, next_sep);
241 for (size_t n = m + 1; n < try_len + 1; ++n) {
242 nextstep = &g_array_index(m_parse_steps, parse_value_t, n);
245 const char * onepinyin = input + m;
246 gint16 onepinyinlen = n - m;
247 value = parse_value_t();
249 ChewingKey key; ChewingKeyRest rest;
250 bool parsed = parse_one_key
251 (options, key, rest, onepinyin, onepinyinlen);
252 rest.m_raw_begin = m; rest.m_raw_end = n;
255 value.m_key = key; value.m_key_rest = rest;
256 value.m_num_keys = curstep->m_num_keys + 1;
257 value.m_parsed_len = curstep->m_parsed_len + onepinyinlen;
258 value.m_last_step = m;
261 if (-1 == nextstep->m_last_step)
263 if (value.m_parsed_len > nextstep->m_parsed_len)
265 if (value.m_parsed_len == nextstep->m_parsed_len &&
266 value.m_num_keys < nextstep->m_num_keys)
272 /* final step for back tracing. */
273 gint16 parsed_len = final_step(step_len, keys, key_rests);
275 /* post processing for re-split table. */
276 if (options & USE_RESPLIT_TABLE) {
277 post_process(options, keys, key_rests);
284 int FullPinyinParser2::final_step(size_t step_len, ChewingKeyVector & keys,
285 ChewingKeyRestVector & key_rests) const{
287 gint16 parsed_len = 0;
288 parse_value_t * curstep = NULL;
290 /* find longest match, which starts from the beginning of input. */
291 for (i = step_len - 1; i >= 0; --i) {
292 curstep = &g_array_index(m_parse_steps, parse_value_t, i);
293 if (i == curstep->m_parsed_len)
296 /* prepare saving. */
297 parsed_len = curstep->m_parsed_len;
298 gint16 num_keys = curstep->m_num_keys;
299 g_array_set_size(keys, num_keys);
300 g_array_set_size(key_rests, num_keys);
302 /* save the match. */
303 while (curstep->m_last_step != -1) {
304 gint16 pos = curstep->m_num_keys - 1;
307 if (0 != curstep->m_key_rest.m_table_index) {
308 ChewingKey * key = &g_array_index(keys, ChewingKey, pos);
309 ChewingKeyRest * rest = &g_array_index
310 (key_rests, ChewingKeyRest, pos);
311 *key = curstep->m_key; *rest = curstep->m_key_rest;
315 curstep = &g_array_index(m_parse_steps, parse_value_t,
316 curstep->m_last_step);
322 bool FullPinyinParser2::post_process(guint32 options,
323 ChewingKeyVector & keys,
324 ChewingKeyRestVector & key_rests) const {
326 assert(keys->len == key_rests->len);
327 gint16 num_keys = keys->len;
329 ChewingKey * cur_key = NULL, * next_key = NULL;
330 ChewingKeyRest * cur_rest = NULL, * next_rest = NULL;
331 guint16 cur_tone = CHEWING_ZERO_TONE, next_tone = CHEWING_ZERO_TONE;
333 for (i = 0; i < num_keys - 1; ++i) {
334 cur_rest = &g_array_index(key_rests, ChewingKeyRest, i);
335 next_rest = &g_array_index(key_rests, ChewingKeyRest, i + 1);
338 if (cur_rest->m_raw_end != next_rest->m_raw_begin)
341 cur_key = &g_array_index(keys, ChewingKey, i);
342 next_key = &g_array_index(keys, ChewingKey, i + 1);
344 if (options & USE_TONE) {
345 cur_tone = cur_key->m_tone;
346 next_tone = next_key->m_tone;
347 cur_key->m_tone = next_key->m_tone = CHEWING_ZERO_TONE;
350 /* lookup re-split table */
352 const resplit_table_item_t * item = NULL;
353 for (k = 0; k < G_N_ELEMENTS(resplit_table); ++k) {
354 item = resplit_table + k;
356 if (item->m_orig_freq >= item->m_new_freq)
359 /* use pinyin_exact_compare2 here. */
360 if (0 == pinyin_exact_compare2(item->m_orig_keys,
367 if (k < G_N_ELEMENTS(resplit_table)) {
369 item = resplit_table + k;
370 *cur_key = item->m_new_keys[0];
371 *next_key = item->m_new_keys[1];
372 /* assumes only moved one char in gen_all_resplit script. */
373 cur_rest->m_raw_end --;
374 next_rest->m_raw_begin --;
377 /* save back tones */
378 if (options & USE_TONE) {
379 cur_key->m_tone = cur_tone;
380 next_key->m_tone = next_tone;
388 bool DoublePinyinParser2::parse_one_key (guint32 options, ChewingKey & key,
389 ChewingKeyRest & key_rest,
390 const char *str, int len) const{
392 if (!(options & PINYIN_INCOMPLETE))
397 options &= ~(PINYIN_CORRECT_ALL|PINYIN_AMB_ALL);
399 if (2 == len || 3 == len) {
400 /* parse shengmu and yunmu here. */
405 if (!(options & USE_TONE))
414 int DoublePinyinParser2::parse (guint32 options, ChewingKeyVector & keys,
415 ChewingKeyRestVector & key_rests,
416 const char *str, int len) const{