bool TessdataManager::Init(const char *data_file_name, int debug_level) {
int i;
debug_level_ = debug_level;
data_file_ = fopen(data_file_name, "rb");
if (data_file_ == NULL) {
tprintf("Error opening data file %s\n", data_file_name);
tprintf("Please make sure the TESSDATA_PREFIX environment variable is set "
"to the parent directory of your \"tessdata\" directory.\n");
return false;
}
fread(&actual_tessdata_num_entries_, sizeof(inT32), 1, data_file_);
swap_ = (actual_tessdata_num_entries_ > kMaxNumTessdataEntries);
if (swap_) {
actual_tessdata_num_entries_ = reverse32(actual_tessdata_num_entries_);
}
ASSERT_HOST(actual_tessdata_num_entries_ <= TESSDATA_NUM_ENTRIES);
fread(offset_table_, sizeof(inT64),
actual_tessdata_num_entries_, data_file_);
if (swap_) {
for (i = 0 ; i < actual_tessdata_num_entries_; ++i) {
offset_table_[i] = reverse64(offset_table_[i]);
}
}
if (debug_level_) {
tprintf("TessdataManager loaded %d types of tesseract data files.\n",
actual_tessdata_num_entries_);
for (i = 0; i < actual_tessdata_num_entries_; ++i) {
tprintf("Offset for type %d is %lld\n", i, offset_table_[i]);
}
}
return true;
}
bool TessdataManager::Init(const char *data_file_name, int debug_level) {
int i;
debug_level_ = debug_level;
data_file_ = fopen(data_file_name, "rb");
if (data_file_ == NULL) {
// WILLUS MOD: Insert "if (debug_level_)" line below to avoid printing error msg.
if (debug_level_)
tprintf("Error opening data file %s\n", data_file_name);
return false;
}
fread(&actual_tessdata_num_entries_, sizeof(inT32), 1, data_file_);
bool swap = (actual_tessdata_num_entries_ > kMaxNumTessdataEntries);
if (swap) {
actual_tessdata_num_entries_ = reverse32(actual_tessdata_num_entries_);
}
ASSERT_HOST(actual_tessdata_num_entries_ <= TESSDATA_NUM_ENTRIES);
fread(offset_table_, sizeof(inT64),
actual_tessdata_num_entries_, data_file_);
if (swap) {
for (i = 0 ; i < actual_tessdata_num_entries_; ++i) {
offset_table_[i] = reverse64(offset_table_[i]);
}
}
if (debug_level_) {
tprintf("TessdataManager loaded %d types of tesseract data files.\n",
actual_tessdata_num_entries_);
for (i = 0; i < actual_tessdata_num_entries_; ++i) {
tprintf("Offset for type %d is %lld\n", i, offset_table_[i]);
}
}
return true;
}
void MessageBlock::reverseBlock()
{
#if BYTE_ORDER == LITTLE_ENDIAN
unsigned int i;
for (i=0; i<16; ++i)
{
reverse64(words[i], words[i]);
}
#endif
}
bool PMSource::getNextMessageBlock(MessageBlock& mBlock)
{
std::streamsize bytesRead = 0;
switch (m_Status)
{
case psUnpadded:
m_BufStream.read((char*) &(mBlock.words[0]), 64);
bytesRead = m_BufStream.gcount();
if (bytesRead==64)
{
m_BitsRead += 512;
mBlock.reverseBlock();
return true;
}
//not full block read
m_BitsRead += (bytesRead * 8);
//close file, we are done here with reading from file anyway
//m_BufStream.close(); // no close, since this will take care of itself in destructor
//add 1-bit (start of message padding)
((uint8_t*) &(mBlock.words[0]))[bytesRead] = 0x80;
//zero out rest of message block
memset(&(((uint8_t*) &(mBlock.words[0]))[bytesRead+1]), 0, 64 - (bytesRead+1));
//pad size value in there, too, if possible
if (bytesRead+1<=56)
{
#if BYTE_ORDER == LITTLE_ENDIAN
reverse64(m_BitsRead, m_BitsRead);
#endif
memcpy(&(mBlock.words[14]), &m_BitsRead, 8);
m_Status = psPaddedAndAllRead;
mBlock.reverseBlock();
return true;
}
m_Status = psPadded1024And512Read;
mBlock.reverseBlock();
return true;
break;
case psPadded1024And512Read:
//fill all with zeros (padding)
memset(&(mBlock.words[0]), 0, 64);
//pad size value in there, too
#if BYTE_ORDER == LITTLE_ENDIAN
reverse64(m_BitsRead, m_BitsRead);
#endif
memcpy(&(mBlock.words[14]), &m_BitsRead, 8);
m_Status = psPaddedAndAllRead;
mBlock.reverseBlock();
return true;
break;
case psPaddedAndAllRead:
return false;
break;
case psPadded512:
case psPadded1024:
//We should never get to this point!
throw std::logic_error("PMSource::getNextMessageBlock(): Code execution should never get to this point!");
return false;
break;
}//swi
//We should never get to this point either!
throw std::logic_error("PMSource::getNextMessageBlock(): Code execution should never get to this point either!");
return false;
}
