void textord_page( //make rows & words
ICOORD page_tr, //top right
BLOCK_LIST *blocks, //block list
TO_BLOCK_LIST *land_blocks, //rotated for landscape
TO_BLOCK_LIST *port_blocks //output list
) {
float gradient; //global skew
set_global_loc_code(LOC_TEXT_ORD_ROWS);
gradient = make_rows (page_tr, blocks, land_blocks, port_blocks);
if (global_monitor != NULL) {
global_monitor->ocr_alive = TRUE;
global_monitor->progress = 20;
}
set_global_loc_code(LOC_TEXT_ORD_WORDS);
make_words(page_tr, gradient, blocks, land_blocks, port_blocks);
if (global_monitor != NULL) {
global_monitor->ocr_alive = TRUE;
global_monitor->progress = 30;
}
cleanup_blocks(blocks); //remove empties
#ifndef GRAPHICS_DISABLED
close_to_win();
#endif
if (textord_exit_after && !interactive_mode)
exit (0);
}
// Make the textlines and words inside each block.
void Textord::TextordPage(PageSegMode pageseg_mode, const FCOORD &reskew,
int width, int height, Pix *binary_pix,
Pix *thresholds_pix, Pix *grey_pix,
bool use_box_bottoms,
BLOCK_LIST *blocks, TO_BLOCK_LIST *to_blocks) {
page_tr_.set_x(width);
page_tr_.set_y(height);
if (to_blocks->empty()) {
// AutoPageSeg was not used, so we need to find_components first.
find_components(binary_pix, blocks, to_blocks);
TO_BLOCK_IT it(to_blocks);
for (it.mark_cycle_pt(); !it.cycled_list(); it.forward()) {
TO_BLOCK *to_block = it.data();
// Compute the edge offsets whether or not there is a grey_pix.
// We have by-passed auto page seg, so we have to run it here.
// By page segmentation mode there is no non-text to avoid running on.
to_block->ComputeEdgeOffsets(thresholds_pix, grey_pix);
}
} else if (!PSM_SPARSE(pageseg_mode)) {
// AutoPageSeg does not need to find_components as it did that already.
// Filter_blobs sets up the TO_BLOCKs the same as find_components does.
filter_blobs(page_tr_, to_blocks, true);
}
ASSERT_HOST(!to_blocks->empty());
if (pageseg_mode == PSM_SINGLE_BLOCK_VERT_TEXT) {
const FCOORD anticlockwise90(0.0f, 1.0f);
const FCOORD clockwise90(0.0f, -1.0f);
TO_BLOCK_IT it(to_blocks);
for (it.mark_cycle_pt(); !it.cycled_list(); it.forward()) {
TO_BLOCK *to_block = it.data();
BLOCK *block = to_block->block;
// Create a fake poly_block in block from its bounding box.
block->set_poly_block(new POLY_BLOCK(block->bounding_box(),
PT_VERTICAL_TEXT));
// Rotate the to_block along with its contained block and blobnbox lists.
to_block->rotate(anticlockwise90);
// Set the block's rotation values to obey the convention followed in
// layout analysis for vertical text.
block->set_re_rotation(clockwise90);
block->set_classify_rotation(clockwise90);
}
}
TO_BLOCK_IT to_block_it(to_blocks);
TO_BLOCK *to_block = to_block_it.data();
// Make the rows in the block.
float gradient = 0;
// Do it the old fashioned way.
if (PSM_LINE_FIND_ENABLED(pageseg_mode)) {
gradient = make_rows(page_tr_, to_blocks);
} else if (!PSM_SPARSE(pageseg_mode)) {
// RAW_LINE, SINGLE_LINE, SINGLE_WORD and SINGLE_CHAR all need a single row.
gradient = make_single_row(page_tr_, pageseg_mode != PSM_RAW_LINE,
to_block, to_blocks);
}
BaselineDetect baseline_detector(textord_baseline_debug,
reskew, to_blocks);
baseline_detector.ComputeStraightBaselines(use_box_bottoms);
baseline_detector.ComputeBaselineSplinesAndXheights(page_tr_, true,
textord_heavy_nr,
textord_show_final_rows,
this);
// Now make the words in the lines.
if (PSM_WORD_FIND_ENABLED(pageseg_mode)) {
// SINGLE_LINE uses the old word maker on the single line.
make_words(this, page_tr_, gradient, blocks, to_blocks);
} else {
// SINGLE_WORD and SINGLE_CHAR cram all the blobs into a
// single word, and in SINGLE_CHAR mode, all the outlines
// go in a single blob.
TO_BLOCK *to_block = to_block_it.data();
make_single_word(pageseg_mode == PSM_SINGLE_CHAR,
to_block->get_rows(), to_block->block->row_list());
}
cleanup_blocks(PSM_WORD_FIND_ENABLED(pageseg_mode), blocks);
// Remove empties.
// Compute the margins for each row in the block, to be used later for
// paragraph detection.
BLOCK_IT b_it(blocks);
for (b_it.mark_cycle_pt(); !b_it.cycled_list(); b_it.forward()) {
b_it.data()->compute_row_margins();
}
#ifndef GRAPHICS_DISABLED
close_to_win();
#endif
}
// Segment the page according to the current value of tessedit_pageseg_mode.
// If the pix_binary_ member is not NULL, it is used as the source image,
// and copied to image, otherwise it just uses image as the input.
// On return the blocks list owns all the constructed page layout.
int Tesseract::SegmentPage(const STRING* input_file,
IMAGE* image, BLOCK_LIST* blocks) {
int width = image->get_xsize();
int height = image->get_ysize();
int resolution = image->get_res();
#ifdef HAVE_LIBLEPT
if (pix_binary_ != NULL) {
width = pixGetWidth(pix_binary_);
height = pixGetHeight(pix_binary_);
resolution = pixGetXRes(pix_binary_);
}
#endif
// Zero resolution messes up the algorithms, so make sure it is credible.
if (resolution < kMinCredibleResolution)
resolution = kDefaultResolution;
// Get page segmentation mode.
PageSegMode pageseg_mode = static_cast<PageSegMode>(
static_cast<int>(tessedit_pageseg_mode));
// If a UNLV zone file can be found, use that instead of segmentation.
if (pageseg_mode != tesseract::PSM_AUTO &&
input_file != NULL && input_file->length() > 0) {
STRING name = *input_file;
const char* lastdot = strrchr(name.string(), '.');
if (lastdot != NULL)
name[lastdot - name.string()] = '\0';
read_unlv_file(name, width, height, blocks);
}
bool single_column = pageseg_mode > PSM_AUTO;
if (blocks->empty()) {
// No UNLV file present. Work according to the PageSegMode.
// First make a single block covering the whole image.
BLOCK_IT block_it(blocks);
BLOCK* block = new BLOCK("", TRUE, 0, 0, 0, 0, width, height);
block_it.add_to_end(block);
} else {
// UNLV file present. Use PSM_SINGLE_COLUMN.
pageseg_mode = PSM_SINGLE_COLUMN;
}
TO_BLOCK_LIST land_blocks, port_blocks;
TBOX page_box;
if (pageseg_mode <= PSM_SINGLE_COLUMN) {
if (AutoPageSeg(width, height, resolution, single_column,
image, blocks, &port_blocks) < 0) {
return -1;
}
// To create blobs from the image region bounds uncomment this line:
// port_blocks.clear(); // Uncomment to go back to the old mode.
} else {
#if HAVE_LIBLEPT
image->FromPix(pix_binary_);
#endif
deskew_ = FCOORD(1.0f, 0.0f);
reskew_ = FCOORD(1.0f, 0.0f);
}
if (blocks->empty()) {
tprintf("Empty page\n");
return 0; // AutoPageSeg found an empty page.
}
if (port_blocks.empty()) {
// AutoPageSeg was not used, so we need to find_components first.
find_components(blocks, &land_blocks, &port_blocks, &page_box);
} else {
// AutoPageSeg does not need to find_components as it did that already.
page_box.set_left(0);
page_box.set_bottom(0);
page_box.set_right(width);
page_box.set_top(height);
// Filter_blobs sets up the TO_BLOCKs the same as find_components does.
filter_blobs(page_box.topright(), &port_blocks, true);
}
TO_BLOCK_IT to_block_it(&port_blocks);
ASSERT_HOST(!port_blocks.empty());
TO_BLOCK* to_block = to_block_it.data();
if (pageseg_mode <= PSM_SINGLE_BLOCK ||
to_block->line_size < 2) {
// For now, AUTO, SINGLE_COLUMN and SINGLE_BLOCK all map to the old
// textord. The difference is the number of blocks and how the are made.
textord_page(page_box.topright(), blocks, &land_blocks, &port_blocks,
this);
} else {
// SINGLE_LINE, SINGLE_WORD and SINGLE_CHAR all need a single row.
float gradient = make_single_row(page_box.topright(),
to_block, &port_blocks, this);
if (pageseg_mode == PSM_SINGLE_LINE) {
// SINGLE_LINE uses the old word maker on the single line.
make_words(page_box.topright(), gradient, blocks,
&land_blocks, &port_blocks, this);
} else {
// SINGLE_WORD and SINGLE_CHAR cram all the blobs into a
// single word, and in SINGLE_CHAR mode, all the outlines
// go in a single blob.
make_single_word(pageseg_mode == PSM_SINGLE_CHAR,
to_block->get_rows(), to_block->block->row_list());
}
}
return 0;
}
//--------------------------------------------------------------------------
int idaapi accept_file(linput_t *li, char fileformatname[MAX_FILE_FORMAT_NAME], int n)
{
char line[MAXSTR];
char *words[MAXSTR];
if ( n )
return 0;
// We try to interpret the input file as a text
// file with a dump format, i.e. all lines should look like
//00000020: 59 69 74 54-55 B6 3E F7-D6 B9 C9 B9-45 E6 A4 52 YitTU¦>?O?E?E??R
//0020: 59 69 74 54 55 B6 3E F7 D6 B9 C9 B9 45 E6 A4 52 "YitTU¦>?O?E?E??R"
//1000: 12 23 34 56 78
//0100: 31 C7 1D AF 32 04 1E 32 05 1E 3C 32 07 1E 21 D9
//12 23 34 56 78
// and similar lines
// We allow non-ascii characters at the end of the line
// We skip empty lines
ssize_t p0len = -1; // length of the first word's hex part
char w0sep[10]; // separator after the first word
w0sep[0] = '\0';
int nl = 0;
int nontrivial_line_count = 0;
bool no_more_lines = false;
uint32 adr, oldadr=0;
while ( qlgets(line, sizeof(line), li) )
{
nl++;
strrpl(line, '-', ' ');
int nw = make_words(line, words, qnumber(words));
if ( line[0] == ';' || line[0] == '#' )
continue;
if ( nw == 0 )
continue;
nontrivial_line_count++;
if ( no_more_lines )
FAILED;
// od -x format may contain '*' lines which mean repetition
if ( strcmp(words[0], "*") == 0 && nw == 1 )
continue;
// the first word must be a number (more than one digit)
char *ptr = words[0];
adr = hex(ptr);
ssize_t p0 = ptr - words[0];
if ( p0 <= 1 )
FAILED;
if ( nontrivial_line_count > 1 && p0 < p0len )
FAILED;
p0len = p0;
// take the separator from the first line
if ( nontrivial_line_count == 1 )
{
qstrncpy(w0sep, ptr, sizeof(w0sep));
while ( *ptr )
if ( strchr(":>-.", *ptr++) == NULL )
FAILED;
}
else
{
if ( strcmp(w0sep, ptr) != 0 )
FAILED;
}
bool haspref = p0len >= 4 || w0sep[0] != '\0';
if ( haspref )
{
// if the line contains only the address, then don't accept lines anymore
if ( nw == 1 )
{
if ( nontrivial_line_count == 1 )
FAILED;
no_more_lines = true;
if ( adr <= oldadr )
FAILED;
}
else
{
// the remaining words should be numbers with at least 1 position
// (at least the second word should be so)
ptr = words[1];
hex(ptr);
if ( ptr == words[1] )
FAILED;
}
}
oldadr = adr;
}
if ( nontrivial_line_count == 0 )
FAILED;
qstrncpy(fileformatname, "Dump file", MAX_FILE_FORMAT_NAME);
return 1;
}
//--------------------------------------------------------------------------
void idaapi load_file(linput_t *li, ushort _neflag, const char * /*fileformatname*/)
{
char line[MAXSTR];
char *words[MAXSTR];
neflag = _neflag;
iscode = (neflag & NEF_CODE) != 0;
sel = BADSEL;
sea = BADADDR;
ea_t ea = 0;
ea_t top= 0;
bool use32 = false;
bool octpref = false;
bool octnum = false;
size_t fill = 0;
// Since we made all the checks in accept_file,
// here we don't repeat them
ssize_t p0len = -1; // length of the first word's hex part
char w0sep[10]; // separator after the first word
w0sep[0] = '\0';
int nontrivial_line_count = 0;
while ( qlgets(line, sizeof(line), li) )
{
strrpl(line, '-', ' ');
if ( line[0] == ';' || line[0] == '#' )
continue;
int n = make_words(line, words, qnumber(words));
if ( n == 0 )
continue;
nontrivial_line_count++;
ssize_t bi;
// od -x format may contain '*' lines which mean repetition
if ( strcmp(words[0], "*") == 0 && n == 1 )
{
fill = size_t(top - ea);
octpref = true; // od -x have octal prefixes
continue;
}
// the first word must be a number (more than one digit)
char *ptr = words[0];
uint32 w0 = octpref ? oct(ptr) : hex(ptr);
p0len = ptr - words[0];
// take the separator from the first line
if ( nontrivial_line_count == 1 )
qstrncpy(w0sep, ptr, sizeof(w0sep));
// process '*' and fill the gap
if ( fill > 0 )
{
while ( top < w0 )
{
ea = top;
top = ea + fill;
copy(ea, top);
}
}
int idx = 0;
if ( w0sep[0] != '\0' || p0len >= 4 )
{
if ( nontrivial_line_count > 1 && !octpref && top != w0 )
{
// strange, the sequence is not contiguous
// check if the prefixes are octal (od -x)
ptr = words[0];
if ( oct(ptr) == top )
{
octpref = true;
ptr = words[0];
w0 = oct(ptr);
}
}
ea = w0;
idx = 1;
}
else
{
ea = top;
}
for ( bi=0; idx < n; idx++ ) //lint !e443
{
ptr = words[idx];
if ( nontrivial_line_count == 1 && !octnum && strlen(ptr) == 6 )
{
oct(ptr);
if ( ptr-words[idx] == 6 )
octnum = true;
ptr = words[idx];
// msg("ptr=%s octnum=%d\n", ptr, octnum);
}
uint32 b = octnum ? oct(ptr) : hex(ptr);
ssize_t nc = ptr - words[idx];
if ( nc < 2 )
{
// we tolerate one-letter separators between numbers
if ( words[idx][1] == '\0' && strchr("\xA6|-:", words[idx][0]) != NULL )
continue;
break;
}
nc /= octnum ? 3 : 2; // number of bytes
*(uint32 *)&bytes[bi] = b;
bi += nc;
}
top = ea + bi;
copy(ea, top);
}
if ( eea >= 0x10000 || p0len > 4 )
use32 = true;
if ( neflag & NEF_SEGS )
{
if ( use32 )
{
set_segm_addressing(getseg(sea), 1);
if ( ph.id == PLFM_386 ) inf.lflags |= LFLG_PC_FLAT;
}
set_default_dataseg(sel);
}
if ( (neflag & NEF_RELOAD) == 0 )
create_filename_cmt();
}