void increase_right(Annealing *search, const unsigned int wm_index, const unsigned int total, const unsigned int length) {
PSSM wm = WM(wm_index);
new_columns(search, wm_index, wm->length, length, total);
wm->last_col = length;
wm->length += wm->last_col;
}
MutableColumnPtr ColumnTuple::cloneEmpty() const
{
const size_t tuple_size = columns.size();
Columns new_columns(tuple_size);
for (size_t i = 0; i < tuple_size; ++i)
new_columns[i] = columns[i]->cloneEmpty();
return ColumnTuple::create(new_columns);
}
MutableColumnPtr ColumnTuple::replicate(const Offsets & offsets) const
{
const size_t tuple_size = columns.size();
Columns new_columns(tuple_size);
for (size_t i = 0; i < tuple_size; ++i)
new_columns[i] = columns[i]->replicate(offsets);
return ColumnTuple::create(new_columns);
}
MutableColumnPtr ColumnTuple::permute(const Permutation & perm, size_t limit) const
{
const size_t tuple_size = columns.size();
Columns new_columns(tuple_size);
for (size_t i = 0; i < tuple_size; ++i)
new_columns[i] = columns[i]->permute(perm, limit);
return ColumnTuple::create(new_columns);
}
MutableColumnPtr ColumnTuple::filter(const Filter & filt, ssize_t result_size_hint) const
{
const size_t tuple_size = columns.size();
Columns new_columns(tuple_size);
for (size_t i = 0; i < tuple_size; ++i)
new_columns[i] = columns[i]->filter(filt, result_size_hint);
return ColumnTuple::create(new_columns);
}
MutableColumns ColumnTuple::scatter(ColumnIndex num_columns, const Selector & selector) const
{
const size_t tuple_size = columns.size();
std::vector<MutableColumns> scattered_tuple_elements(tuple_size);
for (size_t tuple_element_idx = 0; tuple_element_idx < tuple_size; ++tuple_element_idx)
scattered_tuple_elements[tuple_element_idx] = columns[tuple_element_idx]->scatter(num_columns, selector);
MutableColumns res(num_columns);
for (size_t scattered_idx = 0; scattered_idx < num_columns; ++scattered_idx)
{
Columns new_columns(tuple_size);
for (size_t tuple_element_idx = 0; tuple_element_idx < tuple_size; ++tuple_element_idx)
new_columns[tuple_element_idx] = std::move(scattered_tuple_elements[tuple_element_idx][scattered_idx]);
res[scattered_idx] = ColumnTuple::create(new_columns);
}
return res;
}
void increase_left(Annealing *search, const unsigned int wm_index, const unsigned int total, const unsigned int length) {
SWITCH_WM(wm_index);
PSSM wm = WM(wm_index);
PSSM nwm = NWM(wm_index);
int cell, old_cell;
wm->last_col = length;
wm->length = nwm->length + length;
old_cell = nwm->offsets[nwm->length];
cell = wm->offsets[nwm->length + length];
while (old_cell--) {
cell--;
wm->counts[cell] = nwm->counts[old_cell];
wm->scores[cell] = nwm->scores[old_cell];
}
new_columns(search, wm_index, 0, wm->last_col, total);
}
struct table *parse_table(unsigned char *html, unsigned char *eof, unsigned char **end, struct rgb *bgcolor, int sh, struct s_e **bad_html, int *bhp)
{
int qqq;
struct table *t;
struct table_cell *cell;
unsigned char *t_name, *t_attr, *en;
int t_namelen;
int x = 0, y = -1;
int p = 0;
unsigned char *lbhp = NULL;
int l_al = AL_LEFT;
int l_val = VAL_MIDDLE;
int csp, rsp;
int group = 0;
int i, j, k;
struct rgb l_col;
int c_al = AL_TR, c_val = VAL_TR, c_width = W_AUTO, c_span = 0;
memcpy(&l_col, bgcolor, sizeof(struct rgb));
*end = html;
if (bad_html) {
*bad_html = DUMMY;
*bhp = 0;
}
if (!(t = new_table())) return NULL;
se:
en = html;
see:
html = en;
if (bad_html && !p && !lbhp) {
if (!(*bhp & (ALLOC_GR-1))) {
if ((unsigned)*bhp > MAXINT / sizeof(struct s_e) - ALLOC_GR) overalloc();
*bad_html = mem_realloc(*bad_html, (*bhp + ALLOC_GR) * sizeof(struct s_e));
}
lbhp = (*bad_html)[(*bhp)++].s = html;
}
while (html < eof && *html != '<') html++;
if (html >= eof) {
if (p) CELL(t, x, y)->end = html;
if (lbhp) (*bad_html)[*bhp-1].e = html;
goto scan_done;
}
if (html + 2 <= eof && (html[1] == '!' || html[1] == '?')) {
html = skip_comment(html, eof);
goto se;
}
if (parse_element(html, eof, &t_name, &t_namelen, &t_attr, &en)) {
html++;
goto se;
}
if (t_namelen == 5 && !casecmp(t_name, "TABLE", 5)) {
en = skip_table(en, eof);
goto see;
}
if (t_namelen == 6 && !casecmp(t_name, "/TABLE", 6)) {
if (c_span) new_columns(t, c_span, c_width, c_al, c_val, 1);
if (p) CELL(t, x, y)->end = html;
if (lbhp) (*bad_html)[*bhp-1].e = html;
goto scan_done;
}
if (t_namelen == 8 && !casecmp(t_name, "COLGROUP", 8)) {
if (c_span) new_columns(t, c_span, c_width, c_al, c_val, 1);
if (lbhp) (*bad_html)[*bhp-1].e = html, lbhp = NULL;
c_al = AL_TR;
c_val = VAL_TR;
c_width = W_AUTO;
get_align(t_attr, &c_al);
get_valign(t_attr, &c_val);
get_c_width(t_attr, &c_width, sh);
if ((c_span = get_num(t_attr, "span")) == -1) c_span = 1;
goto see;
}
if (t_namelen == 9 && !casecmp(t_name, "/COLGROUP", 9)) {
if (c_span) new_columns(t, c_span, c_width, c_al, c_val, 1);
if (lbhp) (*bad_html)[*bhp-1].e = html, lbhp = NULL;
c_span = 0;
c_al = AL_TR;
c_val = VAL_TR;
c_width = W_AUTO;
goto see;
}
if (t_namelen == 3 && !casecmp(t_name, "COL", 3)) {
int sp, wi, al, val;
if (lbhp) (*bad_html)[*bhp-1].e = html, lbhp = NULL;
if ((sp = get_num(t_attr, "span")) == -1) sp = 1;
wi = c_width;
al = c_al;
val = c_val;
get_align(t_attr, &al);
get_valign(t_attr, &val);
get_c_width(t_attr, &wi, sh);
new_columns(t, sp, wi, al, val, !!c_span);
c_span = 0;
goto see;
}
if (t_namelen == 3 && (!casecmp(t_name, "/TR", 3) || !casecmp(t_name, "/TD", 3) || !casecmp(t_name, "/TH", 3))) {
if (c_span) new_columns(t, c_span, c_width, c_al, c_val, 1);
if (p) CELL(t, x, y)->end = html, p = 0;
if (lbhp) (*bad_html)[*bhp-1].e = html, lbhp = NULL;
}
if (t_namelen == 2 && !casecmp(t_name, "TR", 2)) {
if (c_span) new_columns(t, c_span, c_width, c_al, c_val, 1);
if (p) CELL(t, x, y)->end = html, p = 0;
if (lbhp) (*bad_html)[*bhp-1].e = html, lbhp = NULL;
if (group) group--;
l_al = AL_LEFT;
l_val = VAL_MIDDLE;
memcpy(&l_col, bgcolor, sizeof(struct rgb));
get_align(t_attr, &l_al);
get_valign(t_attr, &l_val);
get_bgcolor(t_attr, &l_col);
y++, x = 0;
goto see;
}
if (t_namelen == 5 && ((!casecmp(t_name, "THEAD", 5)) || (!casecmp(t_name, "TBODY", 5)) || (!casecmp(t_name, "TFOOT", 5)))) {
if (c_span) new_columns(t, c_span, c_width, c_al, c_val, 1);
if (lbhp) (*bad_html)[*bhp-1].e = html, lbhp = NULL;
group = 2;
}
if (t_namelen != 2 || (casecmp(t_name, "TD", 2) && casecmp(t_name, "TH", 2))) goto see;
if (c_span) new_columns(t, c_span, c_width, c_al, c_val, 1);
if (lbhp) (*bad_html)[*bhp-1].e = html, lbhp = NULL;
if (p) CELL(t, x, y)->end = html, p = 0;
if (y == -1) y = 0, x = 0;
nc:
cell = new_cell(t, x, y);
if (cell->used) {
if (cell->colspan == -1) goto see;
x++;
goto nc;
}
cell->mx = x;
cell->my = y;
cell->used = 1;
cell->start = en;
p = 1;
cell->align = l_al;
cell->valign = l_val;
if ((cell->b = upcase(t_name[1]) == 'H')) cell->align = AL_CENTER;
if (group == 1) cell->group = 1;
if (x < t->c) {
if (t->cols[x].align != AL_TR) cell->align = t->cols[x].align;
if (t->cols[x].valign != VAL_TR) cell->valign = t->cols[x].valign;
}
memcpy(&cell->bgcolor, &l_col, sizeof(struct rgb));
get_align(t_attr, &cell->align);
get_valign(t_attr, &cell->valign);
get_bgcolor(t_attr, &cell->bgcolor);
if ((csp = get_num(t_attr, "colspan")) == -1) csp = 1;
if (!csp) csp = -1;
if ((rsp = get_num(t_attr, "rowspan")) == -1) rsp = 1;
if (!rsp) rsp = -1;
if (csp >= 0 && rsp >= 0 && csp * rsp > 100000) {
if (csp > 10) csp = -1;
if (rsp > 10) rsp = -1;
}
cell->colspan = csp;
cell->rowspan = rsp;
if (csp == 1) {
int w = W_AUTO;
get_c_width(t_attr, &w, sh);
if (w != W_AUTO) set_td_width(t, x, w, 0);
}
qqq = t->x;
for (i = 1; csp != -1 ? i < csp : x + i < qqq; i++) {
struct table_cell *sc = new_cell(t, x + i, y);
if (sc->used) {
csp = i;
for (k = 0; k < i; k++) CELL(t, x + k, y)->colspan = csp;
break;
}
sc->used = sc->spanned = 1;
sc->rowspan = rsp;
sc->colspan = csp;
sc->mx = x;
sc->my = y;
}
qqq = t->y;
for (j = 1; rsp != -1 ? j < rsp : y + j < qqq; j++) {
for (k = 0; k < i; k++) {
struct table_cell *sc = new_cell(t, x + k, y + j);
if (sc->used) {
int l, m;
if (sc->mx == x && sc->my == y) continue;
/*internal("boo");*/
for (l = 0; l < k; l++) memset(CELL(t, x + l, y + j), 0, sizeof(struct table_cell));
rsp = j;
for (l = 0; l < i; l++) for (m = 0; m < j; m++) CELL(t, x + l, y + m)->rowspan = j;
goto brk;
}
sc->used = sc->spanned = 1;
sc->rowspan = rsp;
sc->colspan = csp;
sc->mx = x;
sc->my = y;
}
}
brk:
goto see;
scan_done:
*end = html;
for (x = 0; x < t->x; x++) for (y = 0; y < t->y; y++) {
struct table_cell *c = CELL(t, x, y);
if (!c->spanned) {
if (c->colspan == -1) c->colspan = t->x - x;
if (c->rowspan == -1) c->rowspan = t->y - y;
}
}
if ((unsigned)t->y > MAXINT / sizeof(int)) overalloc();
t->r_heights = mem_alloc(t->y * sizeof(int));
memset(t->r_heights, 0, t->y * sizeof(int));
for (x = 0; x < t->c; x++) if (t->cols[x].width != W_AUTO) set_td_width(t, x, t->cols[x].width, 1);
set_td_width(t, t->x, W_AUTO, 0);
return t;
}
