Pckbuckettable *gt_pckbuckettable_new(const FMindex *fmindex,
unsigned int numofchars,
unsigned long totallength,
unsigned int maxdepth)
{
GtArrayPckbck_Boundsatdepth stack;
Pckbck_Boundsatdepth parent, child;
unsigned long rangesize, idx, *rangeOccs;
Pckbuckettable *pckbt;
Mbtab *tmpmbtab;
GT_INITARRAY(&stack,Pckbck_Boundsatdepth);
child.lowerbound = 0;
child.upperbound = totallength+1;
child.depth = 0;
child.code = (GtCodetype) 0;
GT_STOREINARRAY(&stack,Pckbck_Boundsatdepth,128,child);
rangeOccs = gt_malloc(sizeof (*rangeOccs) * GT_MULT2(numofchars));
tmpmbtab = gt_malloc(sizeof (*tmpmbtab) * numofchars);
pckbt = pckbuckettable_allocandinittable(numofchars,maxdepth,true);
while (stack.nextfreePckbck_Boundsatdepth > 0)
{
parent
= stack.spacePckbck_Boundsatdepth[--stack.nextfreePckbck_Boundsatdepth];
gt_assert(parent.lowerbound < parent.upperbound);
rangesize = gt_bwtrangesplitallwithoutspecial(tmpmbtab,
rangeOccs,
fmindex,
parent.lowerbound,
parent.upperbound);
gt_assert(rangesize <= (unsigned long) numofchars);
for (idx = 0; idx < rangesize; idx++)
{
child.lowerbound = tmpmbtab[idx].lowerbound;
child.upperbound = tmpmbtab[idx].upperbound;
child.depth = parent.depth + 1;
gt_assert(child.depth <= maxdepth);
child.code = parent.code * numofchars + idx;
pckbuckettable_storeBoundsatdepth(pckbt,&child);
if (child.depth < maxdepth)
{
if (child.lowerbound + 1 < child.upperbound)
{
GT_STOREINARRAY(&stack,Pckbck_Boundsatdepth,128,child);
} else
{
pckbuckettable_followleafedge(pckbt,fmindex,&child);
}
}
}
}
GT_FREEARRAY(&stack,Pckbck_Boundsatdepth);
gt_free(rangeOccs);
gt_free(tmpmbtab);
printf("filled: %lu (%.2f)\n",pckbt->numofvalues,
(double) pckbt->numofvalues/pckbt->maxnumofvalues);
return pckbt;
}
static void visit_count_children(const FMindex *index,
Nodecount *parent,
GtStackNodecount *stack,
Mbtab *tmpmbtab,
GtUword *rangeOccs,
GT_UNUSED unsigned int numofchars)
{
GtUword rangesize, idx, num_special;
unsigned int offset;
Nodecount child;
gt_assert(parent->lower < parent->upper);
rangesize = gt_bwtrangesplitallwithoutspecial(tmpmbtab,
rangeOccs,
index,
parent->lower,
parent->upper);
gt_assert(rangesize <= (GtUword) numofchars);
offset = 0U;
num_special = parent->upper - parent->lower;
for (idx = 0; idx < rangesize; idx++)
{
child.lower = tmpmbtab[idx].lowerbound;
child.upper = tmpmbtab[idx].upperbound;
child.leaves = 0UL;
child.branching = 1UL;
child.parentOffset = offset + 1U;
child.visited = false;
child.on_branch = false;
/* check if child is part of a branch */
if (child.upper - child.lower == parent->upper - parent->lower)
{
/* TODO DW do sth with non branching nodes */
child.on_branch = true;
parent->branching--;
GT_STACK_PUSH(stack, child);
num_special = 0UL;
offset++;
}
else
{
/* we found a leave on parent*/
if (child.lower + 1 == child.upper)
{
parent->leaves++;
num_special--;
}
else
/* child is a branch of parent node */
{
if (child.lower == child.upper)
{
/* do nothing, this is no node, this is a missing char */
}
else
{
GT_STACK_PUSH(stack, child);
offset++;
num_special -= (child.upper - child.lower);
}
}
}
}
parent->leaves += num_special;
}
