// Common initialization for a single language.
// arg0 is the datapath for the tessdata directory, which could be the
// path of the tessdata directory with no trailing /, or (if tessdata
// lives in the same directory as the executable, the path of the executable,
// hence the name arg0.
// textbase is an optional output file basename (used only for training)
// language is the language code to load.
// oem controls which engine(s) will operate on the image
// configs (argv) is an array of config filenames to load variables from.
// May be NULL.
// configs_size (argc) is the number of elements in configs.
// vars_vec is an optional vector of variables to set.
// vars_values is an optional corresponding vector of values for the variables
// in vars_vec.
// If set_only_init_params is true, then only the initialization variables
// will be set.
int Tesseract::init_tesseract_internal(
const char *arg0, const char *textbase, const char *language,
OcrEngineMode oem, char **configs, int configs_size,
const GenericVector<STRING> *vars_vec,
const GenericVector<STRING> *vars_values,
bool set_only_non_debug_params) {
if (!init_tesseract_lang_data(arg0, textbase, language, oem, configs,
configs_size, vars_vec, vars_values,
set_only_non_debug_params)) {
return -1;
}
if (tessedit_init_config_only) {
tessdata_manager.End();
return 0;
}
// If only Cube will be used, skip loading Tesseract classifier's
// pre-trained templates.
bool init_tesseract_classifier =
(tessedit_ocr_engine_mode == OEM_TESSERACT_ONLY ||
tessedit_ocr_engine_mode == OEM_TESSERACT_CUBE_COMBINED);
// If only Cube will be used and if it has its own Unicharset,
// skip initializing permuter and loading Tesseract Dawgs.
bool init_dict =
!(tessedit_ocr_engine_mode == OEM_CUBE_ONLY &&
tessdata_manager.SeekToStart(TESSDATA_CUBE_UNICHARSET));
program_editup(textbase, init_tesseract_classifier, init_dict);
tessdata_manager.End();
return 0; //Normal exit
}
// init the LM component
int Tesseract::init_tesseract_lm(const char *arg0,
const char *textbase,
const char *language) {
init_tesseract_lang_data(arg0, textbase, language, NULL, 0, false);
getDict().init_permute();
tessdata_manager.End();
return 0;
}
// init the LM component
int Tesseract::init_tesseract_lm(const char *arg0,
const char *textbase,
const char *language) {
if (!init_tesseract_lang_data(arg0, textbase, language, OEM_TESSERACT_ONLY,
NULL, 0, NULL, NULL, false))
return -1;
getDict().Load(Dict::GlobalDawgCache());
tessdata_manager.End();
return 0;
}
// init the LM component
int Tesseract::init_tesseract_lm(const char *arg0, const char *textbase,
const char *language, TessdataManager *mgr) {
if (!init_tesseract_lang_data(arg0, textbase, language, OEM_TESSERACT_ONLY,
NULL, 0, NULL, NULL, false, mgr))
return -1;
getDict().SetupForLoad(Dict::GlobalDawgCache());
getDict().Load(lang, mgr);
getDict().FinishLoad();
return 0;
}
int Tesseract::init_tesseract(
const char *arg0, const char *textbase, const char *language,
char **configs, int configs_size, bool configs_global_only) {
if (!init_tesseract_lang_data(arg0, textbase, language, configs,
configs_size, configs_global_only)) {
return -1;
}
start_recog(textbase);
tessdata_manager.End();
return 0; //Normal exit
}
// Init everything except the language model
int Tesseract::init_tesseract_classifier(
const char *arg0, const char *textbase, const char *language,
char **configs, int configs_size, bool configs_global_only) {
if (!init_tesseract_lang_data (arg0, textbase, language, configs,
configs_size, configs_global_only)) {
return -1;
}
// Dont initialize the permuter.
program_editup(textbase, false);
tessdata_manager.End();
return 0;
}
// init the LM component
int init_tesseract_lm(const char *arg0,
const char *textbase,
const char *language,
const char *configfile,
int configc,
const char *const *configv) {
init_tesseract_lang_data (arg0, textbase, language,
configfile, configc, configv);
init_permute();
return 0; //Normal exit
}
int init_tesseract(const char *arg0,
const char *textbase,
const char *language,
const char *configfile,
int configc,
const char *const *configv) {
init_tesseract_lang_data (arg0, textbase, language,
configfile, configc, configv);
start_recog(configfile, textbase);
set_tess_tweak_vars();
if (tessedit_use_nn) //phils nn stuff
init_net();
return 0; //Normal exit
}
// Common initialization for a single language.
// arg0 is the datapath for the tessdata directory, which could be the
// path of the tessdata directory with no trailing /, or (if tessdata
// lives in the same directory as the executable, the path of the executable,
// hence the name arg0.
// textbase is an optional output file basename (used only for training)
// language is the language code to load.
// oem controls which engine(s) will operate on the image
// configs (argv) is an array of config filenames to load variables from.
// May be NULL.
// configs_size (argc) is the number of elements in configs.
// vars_vec is an optional vector of variables to set.
// vars_values is an optional corresponding vector of values for the variables
// in vars_vec.
// If set_only_init_params is true, then only the initialization variables
// will be set.
int Tesseract::init_tesseract_internal(const char *arg0, const char *textbase,
const char *language, OcrEngineMode oem,
char **configs, int configs_size,
const GenericVector<STRING> *vars_vec,
const GenericVector<STRING> *vars_values,
bool set_only_non_debug_params,
TessdataManager *mgr) {
if (!init_tesseract_lang_data(arg0, textbase, language, oem, configs,
configs_size, vars_vec, vars_values,
set_only_non_debug_params, mgr)) {
return -1;
}
if (tessedit_init_config_only) {
return 0;
}
// If only LSTM will be used, skip loading Tesseract classifier's
// pre-trained templates and dictionary.
bool init_tesseract = tessedit_ocr_engine_mode != OEM_LSTM_ONLY;
program_editup(textbase, init_tesseract ? mgr : nullptr,
init_tesseract ? mgr : nullptr);
return 0; //Normal exit
}