// Reads from the given file. Returns false in case of error.
// If swap is true, assumes a big/little-endian swap is needed.
bool TrainingSample::DeSerialize(bool swap, FILE* fp) {
if (fread(&class_id_, sizeof(class_id_), 1, fp) != 1) return false;
if (fread(&font_id_, sizeof(font_id_), 1, fp) != 1) return false;
if (fread(&page_num_, sizeof(page_num_), 1, fp) != 1) return false;
if (!bounding_box_.DeSerialize(swap, fp)) return false;
if (fread(&num_features_, sizeof(num_features_), 1, fp) != 1) return false;
if (fread(&num_micro_features_, sizeof(num_micro_features_), 1, fp) != 1)
return false;
if (fread(&outline_length_, sizeof(outline_length_), 1, fp) != 1)
return false;
if (swap) {
ReverseN(&class_id_, sizeof(class_id_));
ReverseN(&num_features_, sizeof(num_features_));
ReverseN(&num_micro_features_, sizeof(num_micro_features_));
ReverseN(&outline_length_, sizeof(outline_length_));
}
// Arbitrarily limit the number of elements to protect against bad data.
if (num_features_ > UINT16_MAX) return false;
if (num_micro_features_ > UINT16_MAX) return false;
delete [] features_;
features_ = new INT_FEATURE_STRUCT[num_features_];
if (fread(features_, sizeof(*features_), num_features_, fp)
!= num_features_)
return false;
delete [] micro_features_;
micro_features_ = new MicroFeature[num_micro_features_];
if (fread(micro_features_, sizeof(*micro_features_), num_micro_features_,
fp) != num_micro_features_)
return false;
if (fread(cn_feature_, sizeof(*cn_feature_), kNumCNParams, fp) !=
kNumCNParams) return false;
if (fread(geo_feature_, sizeof(*geo_feature_), GeoCount, fp) != GeoCount)
return false;
return true;
}
bool read_spacing_info(FILE *f, FontInfo* fi, bool swap) {
inT32 vec_size, kern_size;
if (fread(&vec_size, sizeof(vec_size), 1, f) != 1) return false;
if (swap) Reverse32(&vec_size);
ASSERT_HOST(vec_size >= 0);
if (vec_size == 0) return true;
fi->init_spacing(vec_size);
for (int i = 0; i < vec_size; ++i) {
FontSpacingInfo *fs = new FontSpacingInfo();
if (fread(&fs->x_gap_before, sizeof(fs->x_gap_before), 1, f) != 1 ||
fread(&fs->x_gap_after, sizeof(fs->x_gap_after), 1, f) != 1 ||
fread(&kern_size, sizeof(kern_size), 1, f) != 1) {
delete fs;
return false;
}
if (swap) {
ReverseN(&(fs->x_gap_before), sizeof(fs->x_gap_before));
ReverseN(&(fs->x_gap_after), sizeof(fs->x_gap_after));
Reverse32(&kern_size);
}
if (kern_size < 0) { // indication of a NULL entry in fi->spacing_vec
delete fs;
continue;
}
if (kern_size > 0 && (!fs->kerned_unichar_ids.DeSerialize(swap, f) ||
!fs->kerned_x_gaps.DeSerialize(swap, f))) {
delete fs;
return false;
}
fi->add_spacing(i, fs);
}
return true;
}
bool TessdataManager::Init(const char *data_file_name, int debug_level) {
int i;
debug_level_ = debug_level;
data_file_name_ = data_file_name;
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_) {
ReverseN(&actual_tessdata_num_entries_,
sizeof(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) {
ReverseN(&offset_table_[i], sizeof(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;
}
// Reads from the given file. Returns false in case of error.
// If swap is true, assumes a big/little-endian swap is needed.
bool TrainingSample::DeSerialize(bool swap, FILE* fp) {
if (fread(&class_id_, sizeof(class_id_), 1, fp) != 1) return false;
if (fread(&font_id_, sizeof(font_id_), 1, fp) != 1) return false;
if (fread(&page_num_, sizeof(page_num_), 1, fp) != 1) return false;
if (!bounding_box_.DeSerialize(swap, fp)) return false;
if (fread(&num_features_, sizeof(num_features_), 1, fp) != 1) return false;
if (fread(&num_micro_features_, sizeof(num_micro_features_), 1, fp) != 1)
return false;
if (swap) {
ReverseN(&class_id_, sizeof(class_id_));
ReverseN(&num_features_, sizeof(num_features_));
ReverseN(&num_micro_features_, sizeof(num_micro_features_));
}
delete [] features_;
features_ = new INT_FEATURE_STRUCT[num_features_];
if (fread(features_, sizeof(*features_), num_features_, fp) != num_features_)
return false;
delete [] micro_features_;
micro_features_ = new MicroFeature[num_micro_features_];
if (fread(micro_features_, sizeof(*micro_features_), num_micro_features_,
fp) != num_micro_features_)
return false;
if (fread(cn_feature_, sizeof(*cn_feature_), kNumCNParams, fp) !=
kNumCNParams) return false;
if (fread(geo_feature_, sizeof(*geo_feature_), GeoCount, fp) != GeoCount)
return false;
return true;
}
// Reads from the given file. Returns false in case of error.
// If swap is true, assumes a big/little-endian swap is needed.
// Should be overridden by subclasses, but NOT called by their DeSerialize.
bool Network::DeSerialize(bool swap, TFile* fp) {
inT8 data = 0;
if (fp->FRead(&data, sizeof(data), 1) != 1) return false;
if (data == NT_NONE) {
STRING type_name;
if (!type_name.DeSerialize(swap, fp)) return false;
for (data = 0; data < NT_COUNT && type_name != kTypeNames[data]; ++data) {
}
if (data == NT_COUNT) {
tprintf("Invalid network layer type:%s\n", type_name.string());
return false;
}
}
type_ = static_cast<NetworkType>(data);
if (fp->FRead(&data, sizeof(data), 1) != 1) return false;
training_ = data == TS_ENABLED ? TS_ENABLED : TS_DISABLED;
if (fp->FRead(&data, sizeof(data), 1) != 1) return false;
needs_to_backprop_ = data != 0;
if (fp->FRead(&network_flags_, sizeof(network_flags_), 1) != 1) return false;
if (fp->FRead(&ni_, sizeof(ni_), 1) != 1) return false;
if (fp->FRead(&no_, sizeof(no_), 1) != 1) return false;
if (fp->FRead(&num_weights_, sizeof(num_weights_), 1) != 1) return false;
if (!name_.DeSerialize(swap, fp)) return false;
if (swap) {
ReverseN(&network_flags_, sizeof(network_flags_));
ReverseN(&ni_, sizeof(ni_));
ReverseN(&no_, sizeof(no_));
ReverseN(&num_weights_, sizeof(num_weights_));
}
return true;
}
// Reads from the given file. Returns false in case of error.
// If swap is true, assumes a big/little-endian swap is needed.
bool ICOORD::DeSerialize(bool swap, FILE* fp) {
if (fread(&xcoord, sizeof(xcoord), 1, fp) != 1) return false;
if (fread(&ycoord, sizeof(ycoord), 1, fp) != 1) return false;
if (swap) {
ReverseN(&xcoord, sizeof(xcoord));
ReverseN(&ycoord, sizeof(ycoord));
}
return true;
}
// Reads from the given file. Returns false in case of error.
// If swap is true, assumes a big/little-endian swap is needed.
bool TrainingSampleSet::FontClassInfo::DeSerialize(bool swap, FILE *fp) {
if (fread(&num_raw_samples, sizeof(num_raw_samples), 1, fp) != 1)
return false;
if (fread(&canonical_sample, sizeof(canonical_sample), 1, fp) != 1)
return false;
if (fread(&canonical_dist, sizeof(canonical_dist), 1, fp) != 1) return false;
if (!samples.DeSerialize(swap, fp)) return false;
if (swap) {
ReverseN(&num_raw_samples, sizeof(num_raw_samples));
ReverseN(&canonical_sample, sizeof(canonical_sample));
ReverseN(&canonical_dist, sizeof(canonical_dist));
}
return true;
}
// Reads from the given file. Returns false in case of error.
// If swap is true, assumes a big/little-endian swap is needed.
bool BitVector::DeSerialize(bool swap, FILE* fp) {
uinT32 new_bit_size;
if (fread(&new_bit_size, sizeof(new_bit_size), 1, fp) != 1) return false;
if (swap) {
ReverseN(&new_bit_size, sizeof(new_bit_size));
}
Alloc(new_bit_size);
int wordlen = WordLength();
if (fread(array_, sizeof(*array_), wordlen, fp) != wordlen) return false;
if (swap) {
for (int i = 0; i < wordlen; ++i)
ReverseN(&array_[i], sizeof(array_[i]));
}
return true;
}
// Reads from the given file. Returns false in case of error.
// If swap is true, assumes a big/little-endian swap is needed.
bool WordFeature::DeSerialize(bool swap, FILE* fp) {
if (fread(&x_, sizeof(x_), 1, fp) != 1) return false;
if (swap) ReverseN(&x_, sizeof(x_));
if (fread(&y_, sizeof(y_), 1, fp) != 1) return false;
if (fread(&dir_, sizeof(dir_), 1, fp) != 1) return false;
return true;
}
// Reads from the given file. Returns false in case of error.
// If swap is true, assumes a big/little-endian swap is needed.
bool UnicharAndFonts::DeSerialize(bool swap, FILE* fp) {
if (fread(&unichar_id, sizeof(unichar_id), 1, fp) != 1) return false;
if (swap)
ReverseN(&unichar_id, sizeof(unichar_id));
if (!font_ids.DeSerialize(swap, fp)) return false;
return true;
}
// Loads from the given memory buffer as if a file.
bool TessdataManager::LoadMemBuffer(const char *name, const char *data,
int size) {
Clear();
data_file_name_ = name;
TFile fp;
fp.Open(data, size);
uint32_t num_entries;
if (!fp.DeSerialize(&num_entries)) return false;
swap_ = num_entries > kMaxNumTessdataEntries;
fp.set_swap(swap_);
if (swap_) ReverseN(&num_entries, sizeof(num_entries));
if (num_entries > kMaxNumTessdataEntries) return false;
GenericVector<int64_t> offset_table;
offset_table.resize_no_init(num_entries);
if (!fp.DeSerialize(&offset_table[0], num_entries)) return false;
for (int i = 0; i < num_entries && i < TESSDATA_NUM_ENTRIES; ++i) {
if (offset_table[i] >= 0) {
int64_t entry_size = size - offset_table[i];
int j = i + 1;
while (j < num_entries && offset_table[j] == -1) ++j;
if (j < num_entries) entry_size = offset_table[j] - offset_table[i];
entries_[i].resize_no_init(entry_size);
if (!fp.DeSerialize(&entries_[i][0], entry_size)) return false;
}
}
if (entries_[TESSDATA_VERSION].empty()) {
SetVersionString("Pre-4.0.0");
}
is_loaded_ = true;
return true;
}
// Reads from the given file. Returns false in case of error.
// If swap is true, assumes a big/little-endian swap is needed.
bool STRING::DeSerialize(bool swap, FILE* fp) {
inT32 len;
if (fread(&len, sizeof(len), 1, fp) != 1) return false;
if (swap)
ReverseN(&len, sizeof(len));
truncate_at(len);
if (fread(GetCStr(), 1, len, fp) != len) return false;
return true;
}
int TFile::FReadEndian(void* buffer, size_t size, int count) {
int num_read = FRead(buffer, size, count);
if (swap_) {
char* char_buffer = static_cast<char*>(buffer);
for (int i = 0; i < num_read; ++i, char_buffer += size) {
ReverseN(char_buffer, size);
}
}
return num_read;
}
// Reads from the given file. Returns false in case of error.
// If swap is true, assumes a big/little-endian swap is needed.
bool ImageData::DeSerialize(bool swap, TFile* fp) {
if (!imagefilename_.DeSerialize(swap, fp)) return false;
if (fp->FRead(&page_number_, sizeof(page_number_), 1) != 1) return false;
if (swap) ReverseN(&page_number_, sizeof(page_number_));
if (!image_data_.DeSerialize(swap, fp)) return false;
if (!transcription_.DeSerialize(swap, fp)) return false;
// WARNING: Will not work across different endian machines.
if (!boxes_.DeSerialize(swap, fp)) return false;
if (!box_texts_.DeSerializeClasses(swap, fp)) return false;
inT8 vertical = 0;
if (fp->FRead(&vertical, sizeof(vertical), 1) != 1) return false;
vertical_text_ = vertical != 0;
return true;
}
// Reads from the given file. Returns false in case of error.
// If swap is true, assumes a big/little-endian swap is needed.
bool ImageData::DeSerialize(bool swap, FILE* fp) {
if (!imagefilename_.DeSerialize(swap, fp)) return false;
if (fread(&page_number_, sizeof(page_number_), 1, fp) != 1) return false;
if (swap) ReverseN(&page_number_, sizeof(page_number_));
if (!image_data_.DeSerialize(swap, fp)) return false;
if (!transcription_.DeSerialize(swap, fp)) return false;
// WARNING: Will not work across different endian machines.
if (!boxes_.DeSerialize(swap, fp)) return false;
if (!box_texts_.DeSerializeClasses(swap, fp)) return false;
if (!features_.DeSerializeClasses(swap, fp)) return false;
if (!side_data_.DeSerialize(swap, fp)) return false;
STRING box_str;
for (int i = 0; i < box_texts_.size(); ++i) {
box_str += box_texts_[i];
}
partial_boxes_ = !box_texts_.empty() && transcription_ != box_str;
return true;
}
