int
PinyinPhraseLib::find_phrases (PhraseVector &vec,
const char *keys,
bool noshorter,
bool nolonger)
{
PinyinParsedKeyVector pykeys;
PinyinDefaultParser parser;
parser.parse (*m_validator, pykeys, keys);
return find_phrases (vec, pykeys, noshorter, nolonger);
}
int main( int argc, char * argv[]){
PinyinCustomSettings custom;
PinyinLargeTable largetable(&custom);
FacadePhraseIndex phrase_index;
FILE * gbfile = fopen("../../data/gb_char.table", "r");
if ( gbfile == NULL ) {
fprintf(stderr, "open gb_char.table failed!\n");
exit(ENOENT);
}
largetable.load_text(gbfile);
fseek(gbfile, 0L, SEEK_SET);
phrase_index.load_text(1, gbfile);
fclose(gbfile);
FILE * gbkfile = fopen("../../data/gbk_char.table","r");
if ( gbkfile == NULL ) {
fprintf(stderr, "open gb_char.table failed!\n");
exit(ENOENT);
}
largetable.load_text(gbkfile);
fseek(gbkfile, 0L, SEEK_SET);
phrase_index.load_text(2, gbkfile);
fclose(gbkfile);
MemoryChunk* new_chunk = new MemoryChunk;
largetable.store(new_chunk);
largetable.load(new_chunk);
char* linebuf = NULL;
size_t size = 0;
while( getline(&linebuf, &size, stdin) ){
linebuf[strlen(linebuf)-1] = '\0';
if ( strcmp ( linebuf, "quit" ) == 0)
break;
PinyinDefaultParser parser;
NullPinyinValidator validator;
PinyinKeyVector keys;
PinyinKeyPosVector poses;
keys = g_array_new(FALSE, FALSE, sizeof( PinyinKey));
poses = g_array_new(FALSE, FALSE, sizeof( PinyinKeyPos));
parser.parse(validator, keys, poses, linebuf);
guint32 start = record_time();
PhraseIndexRanges ranges;
for( size_t i = 0 ; i < PHRASE_INDEX_LIBRARY_COUNT ; ++i){
ranges[i] = g_array_new(FALSE, FALSE, sizeof (PhraseIndexRange));
}
for ( size_t i = 0 ; i < bench_times; ++i){
largetable.search(keys->len, (PinyinKey *)keys->data, ranges);
}
for( size_t i = 0 ; i < PHRASE_INDEX_LIBRARY_COUNT ; ++i){
GArray * range = ranges[i];
g_array_set_size( range, 0);
}
print_time(start, bench_times);
largetable.search(keys->len, (PinyinKey *)keys->data, ranges);
for( size_t i = 0 ; i < PHRASE_INDEX_LIBRARY_COUNT ; ++i){
GArray * range = ranges[i];
if ( range ){
for (size_t k = 0; k < range->len; ++k){
PhraseIndexRange* onerange = &g_array_index(range, PhraseIndexRange, k);
printf("start:%d\tend:%d\n", onerange->m_range_begin, onerange->m_range_end);
PhraseItem item;
for ( phrase_token_t token = onerange->m_range_begin; token != onerange->m_range_end; ++token){
phrase_index.get_phrase_item( token, item);
gunichar2 bufstr[1024];
item.get_phrase_string(bufstr);
char * string = g_utf16_to_utf8
( bufstr, item.get_phrase_length(),
NULL, NULL, NULL);
printf("%s\t", string);
g_free(string);
PinyinKey pinyin_buffer[1024];
size_t npron = item.get_n_pronunciation();
guint32 freq;
for ( size_t n = 0; n < npron; ++n){
item.get_nth_pronunciation(n, pinyin_buffer, freq);
for ( size_t o = 0; o < item.get_phrase_length(); ++o){
printf("%s'", pinyin_buffer[o].get_key_string());
}
printf("\b\t%d\t", freq);
}
printf("\n");
}
}
if ( range->len)
printf("range items number:%d\n", range->len);
}
g_array_set_size( range, 0);
}
g_array_free(keys, TRUE);
g_array_free(poses, TRUE);
}
if (linebuf)
free(linebuf);
return 0;
}
void feed_line (const char * phrase, const char * pinyin, const guint32 freq){
phrase_item * new_phrase_ptr = (phrase_item *)
malloc( sizeof(phrase_item));
new_phrase_ptr->length = g_utf8_strlen(phrase, -1);
/* FIXME: modify ">" to ">=" according to pinyin_large_table.cpp
* where is the code which I don't want to touch. :-)
*/
if (new_phrase_ptr->length >= MAX_PHRASE_LENGTH ) {
fprintf(stderr, "too long phrase:%s\t%s\t%d\n", phrase,
pinyin, freq);
free(new_phrase_ptr);
return;
}
new_phrase_ptr->uniphrase = g_utf8_to_ucs4(phrase, -1, NULL, NULL, NULL);
PinyinDefaultParser parser;
NullPinyinValidator validator;
PinyinKeyVector keys;
PinyinKeyPosVector poses;
keys = g_array_new(FALSE, FALSE, sizeof( PinyinKey));
poses = g_array_new(FALSE, FALSE, sizeof( PinyinKeyPos));
parser.parse(validator, keys, poses, pinyin);
GArray * array = (GArray *)g_tree_lookup(g_pinyin_tree, new_phrase_ptr);
pinyin_and_freq_item value_item;
value_item.pinyin = keys;
value_item.freq = freq;
if(new_phrase_ptr->length != value_item.pinyin->len){
fprintf(stderr, "error:phrase:%s\tpinyin:%s\n", phrase, pinyin);
return;
}
if ( array == NULL){
array = g_array_new(FALSE, TRUE, sizeof(pinyin_and_freq_item));
g_array_append_val(array, value_item);
g_tree_insert(g_pinyin_tree, new_phrase_ptr, array);
return;
}
bool found = false;
for ( size_t i = 0; i < array->len ; ++i){
pinyin_and_freq_item * old_value_item = &g_array_index(array, pinyin_and_freq_item, i);
int result = pinyin_exact_compare((PinyinKey *)value_item.pinyin->data,
(PinyinKey *)old_value_item->pinyin->data , value_item.pinyin->len);
if ( result == 0 ){
printf("Duplicate item: phrase:%s\tpinyin:%s\tfreq:%u\n",
phrase, pinyin, freq);
old_value_item->freq += freq;
found = true;
}
}
g_array_free(poses, TRUE);
if ( !found ){
g_array_append_val(array, value_item);
g_tree_insert(g_pinyin_tree, new_phrase_ptr, array);
}else
g_array_free(keys, TRUE);
free(new_phrase_ptr);
//g_array_free(keys, TRUE);
}
int main( int argc, char * argv[]){
PinyinCustomSettings custom;
PinyinLargeTable largetable(&custom);
FILE * gbfile = fopen("../../data/gb_char.table", "r");
if ( gbfile == NULL ) {
fprintf(stderr, "open gb_char.table failed!\n");
exit(ENOENT);
}
largetable.load_text(gbfile);
fclose(gbfile);
FILE * gbkfile = fopen("../../data/gbk_char.table","r");
if ( gbkfile == NULL ) {
fprintf(stderr, "open gbk_char.table failed!\n");
exit(ENOENT);
}
largetable.load_text(gbkfile);
fclose(gbkfile);
MemoryChunk* new_chunk = new MemoryChunk;
largetable.store(new_chunk);
largetable.load(new_chunk);
char* linebuf = NULL;
size_t size = 0;
while( getline(&linebuf, &size, stdin) ){
linebuf[strlen(linebuf)-1] = '\0';
if ( strcmp ( linebuf, "quit" ) == 0)
break;
PinyinDefaultParser parser;
NullPinyinValidator validator;
PinyinKeyVector keys;
PinyinKeyPosVector poses;
keys = g_array_new(FALSE, FALSE, sizeof( PinyinKey));
poses = g_array_new(FALSE, FALSE, sizeof( PinyinKeyPos));
parser.parse(validator, keys, poses, linebuf);
guint32 start = record_time();
PhraseIndexRanges ranges;
for( size_t i = 0 ; i < PHRASE_INDEX_LIBRARY_COUNT ; ++i){
ranges[i] = g_array_new(FALSE, FALSE, sizeof (PhraseIndexRange));
}
for ( size_t i = 0 ; i < bench_times; ++i){
largetable.search(keys->len, (PinyinKey *)keys->data, ranges);
}
for( size_t i = 0 ; i < PHRASE_INDEX_LIBRARY_COUNT ; ++i){
GArray * range = ranges[i];
g_array_set_size( range, 0);
}
print_time(start, bench_times);
largetable.search(keys->len, (PinyinKey *)keys->data, ranges);
for( size_t i = 0 ; i < PHRASE_INDEX_LIBRARY_COUNT ; ++i){
GArray * range = ranges[i];
if (range) {
if (range->len)
printf("range items number:%d\n", range->len);
for (size_t k = 0; k < range->len; ++k) {
PhraseIndexRange * onerange =
&g_array_index(range, PhraseIndexRange, k);
printf("start:%d\tend:%d\n", onerange->m_range_begin,
onerange->m_range_end);
}
}
g_array_set_size(range, 0);
}
g_array_free(keys, TRUE);
g_array_free(poses, TRUE);
}
if (linebuf)
free(linebuf);
return 0;
}