bool TessdataManager::TessdataTypeFromFileName(
const char *filename, TessdataType *type, bool *text_file) {
// Get the file suffix (extension)
const char *suffix = strrchr(filename, '.');
if (suffix == NULL || *(++suffix) == '\0') return false;
return TessdataTypeFromFileSuffix(suffix, type, text_file);
}
bool TessdataManager::CombineDataFiles(
const char *language_data_path_prefix,
const char *output_filename) {
// Load individual tessdata components from files.
for (int i = 0; i < TESSDATA_NUM_ENTRIES; ++i) {
TessdataType type;
ASSERT_HOST(TessdataTypeFromFileSuffix(kTessdataFileSuffixes[i], &type));
STRING filename = language_data_path_prefix;
filename += kTessdataFileSuffixes[i];
FILE *fp = fopen(filename.string(), "rb");
if (fp != nullptr) {
fclose(fp);
if (!LoadDataFromFile(filename, &entries_[type])) {
tprintf("Load of file %s failed!\n", filename.string());
return false;
}
}
}
is_loaded_ = true;
// Make sure that the required components are present.
if (!IsBaseAvailable() && !IsLSTMAvailable()) {
tprintf(
"Error: traineddata file must contain at least (a unicharset file"
"and inttemp) OR an lstm file.\n");
return false;
}
// Write updated data to the output traineddata file.
return SaveFile(output_filename, nullptr);
}
bool TessdataManager::CombineDataFiles(
const char *language_data_path_prefix,
const char *output_filename) {
int i;
inT64 offset_table[TESSDATA_NUM_ENTRIES];
for (i = 0; i < TESSDATA_NUM_ENTRIES; ++i) offset_table[i] = -1;
FILE *output_file = fopen(output_filename, "wb");
if (output_file == NULL) {
tprintf("Error opening %s for writing\n", output_filename);
return false;
}
// Leave some space for recording the offset_table.
if (fseek(output_file,
sizeof(inT32) + sizeof(inT64) * TESSDATA_NUM_ENTRIES, SEEK_SET)) {
tprintf("Error seeking %s\n", output_filename);
fclose(output_file);
return false;
}
TessdataType type = TESSDATA_NUM_ENTRIES;
bool text_file = false;
FILE *file_ptr[TESSDATA_NUM_ENTRIES];
// Load individual tessdata components from files.
for (i = 0; i < TESSDATA_NUM_ENTRIES; ++i) {
ASSERT_HOST(TessdataTypeFromFileSuffix(
kTessdataFileSuffixes[i], &type, &text_file));
STRING filename = language_data_path_prefix;
filename += kTessdataFileSuffixes[i];
file_ptr[i] = fopen(filename.string(), "rb");
if (file_ptr[i] != NULL) {
offset_table[type] = ftell(output_file);
CopyFile(file_ptr[i], output_file, text_file, -1);
fclose(file_ptr[i]);
}
}
// Make sure that the required components are present.
if (file_ptr[TESSDATA_UNICHARSET] == NULL) {
tprintf("Error opening %sunicharset file\n", language_data_path_prefix);
fclose(output_file);
return false;
}
if (file_ptr[TESSDATA_INTTEMP] != NULL &&
(file_ptr[TESSDATA_PFFMTABLE] == NULL ||
file_ptr[TESSDATA_NORMPROTO] == NULL)) {
tprintf("Error opening %spffmtable and/or %snormproto files"
" while %sinttemp file was present\n", language_data_path_prefix,
language_data_path_prefix, language_data_path_prefix);
fclose(output_file);
return false;
}
return WriteMetadata(offset_table, language_data_path_prefix, output_file);
}
