// Parse a string of the form [~]<lang>[+[~]<lang>]*.
// Langs with no prefix get appended to to_load, provided they
// are not in there already.
// Langs with ~ prefix get appended to not_to_load, provided they are not in
// there already.
void Tesseract::ParseLanguageString(const char* lang_str,
GenericVector<STRING>* to_load,
GenericVector<STRING>* not_to_load) {
STRING remains(lang_str);
while (remains.length() > 0) {
// Find the start of the lang code and which vector to add to.
const char* start = remains.string();
while (*start == '+')
++start;
GenericVector<STRING>* target = to_load;
if (*start == '~') {
target = not_to_load;
++start;
}
// Find the index of the end of the lang code in string start.
int end = strlen(start);
const char* plus = strchr(start, '+');
if (plus != NULL && plus - start < end)
end = plus - start;
STRING lang_code(start);
lang_code.truncate_at(end);
STRING next(start + end);
remains = next;
// Check whether lang_code is already in the target vector and add.
if (!IsStrInList(lang_code, *target)) {
if (tessdata_manager_debug_level)
tprintf("Adding language '%s' to list\n", lang_code.string());
target->push_back(lang_code);
}
}
}
std::string getOSXSystemLang()
{
// Get the user's language list (in order of preference)
CFArrayRef langs = CFLocaleCopyPreferredLanguages();
if( CFArrayGetCount( langs ) == 0 ) {
return "en_US";
}
const char *lang_code_raw = CFStringGetCStringPtr(
( CFStringRef )CFArrayGetValueAtIndex( langs, 0 ),
kCFStringEncodingUTF8 );
if( !lang_code_raw ) {
return "en_US";
}
// Convert to the underscore format expected by gettext
std::string lang_code( lang_code_raw );
std::replace( lang_code.begin(), lang_code.end(), '-', '_' );
/**
* Handle special case for simplified/traditional Chinese. Simplified/Traditional
* is actually denoted by the region code in older iterations of the
* language codes, whereas now (at least on OS X) region is distinct.
* That is, CDDA expects 'zh_CN' but OS X might give 'zh-Hans-CN'.
*/
if( string_starts_with( lang_code, "zh_Hans" ) ) {
return "zh_CN";
} else if( string_starts_with( lang_code, "zh_Hant" ) ) {
return "zh_TW";
}
return isValidLanguage( lang_code ) ? lang_code : "en_US";
}
int main(int argc, char **argv)
{
const glist *seg = NULL;
u32 key = 0;
Pvoid_t entries = NULL;
uint dc = 0;
uint ic = 0;
const char *stats = NULL;
dub_init();
stats = pparm_common_name("istats");
open_serializer();
while ((seg = pull_head(&key))){
uint i;
u32 lang = lang_code(seg);
do{
for (i = 0; i < seg->len; i++){
Word_t *e = NULL;
Word_t idx = seg->lst[i];
struct istat_entry *ent = NULL;
JLI(e, entries, idx);
if (!*e){
*e = (Word_t)xmalloc(
sizeof(struct istat_entry));
ent = (struct istat_entry*)*e;
memset(ent, 0,
sizeof(struct istat_entry));
ent->xid = idx;
}
ent = (struct istat_entry*)*e;
++ent->freq;
if (!ent->lang_code)
ent->lang_code = lang;
}
}while ((seg = pull_segment()));
++dc;
}
ic = write_istats(stats, entries);
dub_msg("Number of documents: %u\n", dc);
dub_msg("Number of ixemes: %u\n", ic);
return 0;
}
