/******** lookup related functions *****/
elem tree_lookup(tree* T, elem x, bst B)
{
REQUIRES(is_tree(T, B));
if (T == NULL) return NULL;
int r = B->elem_compare(x, T->data);
if (r == 0) {
return T->data;
} else if (r < 0) {
return tree_lookup(T->left, x, B);
} else {
return tree_lookup(T->right, x, B);
}
}
elem bst_lookup(bst B, elem x)
{
REQUIRES(is_bst(B) && x != NULL);
elem r = tree_lookup(B->root, x, B);
ENSURES(r == NULL || B->elem_compare(r, x) == 0);
return r;
}
static int tree_lookup(node_t *node, char target)
{
if (!node)
return 1;
if (node->data == target)
return 0;
return tree_lookup(target < node->data ? node->left : node->right, target);
}
/**
* Constructor.
*
* Compressed suffix tree is a self-index so text is not needed after construction.
* Text can be deleted during construction to save some construction space.
*
* Parameter <filename> is a prefix for filenames used for IO operation.
* Filename suffixes are ".bwt", ".lcp" and ".bp".
*
* @param text a pointer to the text.
* @param n number of symbols in the text.
* @param deletetext delete text as soon as possible.
* @param samplerate sample rate for Compressed suffix array.
* @param action IO action for <filename>. Defaults to no operation.
* @param filename Prefix for the filenames used.
*/
SSTree::SSTree(uchar *text, ulong n, bool deletetext, unsigned samplerate, io_action IOaction, const char *filename)
{
#ifdef SSTREE_HEAPPROFILE
ulong heapCon;
#endif
this->n = n;
unsigned floorLog2n = Tools::FloorLog2(n);
if (floorLog2n < 4)
floorLog2n = 4;
unsigned rmqSampleRate = floorLog2n / 2;
if (samplerate != 0)
floorLog2n = samplerate; // Samplerate override, affects only CSA
#ifdef SSTREE_TIMER
#ifdef SSTREE_HEAPPROFILE
std::cout << "--> HeapProfiler: " << HeapProfiler::GetHeapConsumption() << ", " << HeapProfiler::GetMaxHeapConsumption() << std::endl;
heapCon = HeapProfiler::GetHeapConsumption();
#endif
printf("Creating CSA with samplerate %u\n", floorLog2n);
fflush(stdout);
Tools::StartTimer();
#endif
if (IOaction == load_from && filename != 0)
sa = new CSA(text, n, floorLog2n, (string(filename)+".csa").c_str());
else if (IOaction == save_to && filename != 0)
sa = new CSA(text, n, floorLog2n, 0, (string(filename)+".csa").c_str());
else // No IO operation
sa = new CSA(text, n, floorLog2n);
#ifdef SSTREE_TIMER
printf("CSA created in %.0f seconds.\n", Tools::GetTime());
#ifdef SSTREE_HEAPPROFILE
std::cout << "--> HeapProfiler: " << HeapProfiler::GetHeapConsumption() - heapCon << ", " << HeapProfiler::GetMaxHeapConsumption() << std::endl;
heapCon = HeapProfiler::GetHeapConsumption();
#endif
printf("Creating CHgtArray\n");
fflush(stdout);
Tools::StartTimer();
#endif
if (IOaction == load_from && filename != 0)
hgt = new CHgtArray(sa, (string(filename)+".lcp").c_str());
else
hgt = new CHgtArray(sa, text, n);
if (IOaction == save_to && filename != 0)
hgt->SaveToFile((string(filename)+".lcp").c_str());
#ifdef SSTREE_TIMER
printf("CHgtArray created in %.0f seconds.\n", Tools::GetTime());
#ifdef SSTREE_HEAPPROFILE
std::cout << "--> HeapProfiler: " << HeapProfiler::GetHeapConsumption() - heapCon << ", " << HeapProfiler::GetMaxHeapConsumption() << std::endl;
heapCon = HeapProfiler::GetHeapConsumption();
#endif
printf("Creating parentheses sequence (LcpToParentheses)\n");
fflush(stdout);
Tools::StartTimer();
#endif
if (deletetext)
delete [] text;
#ifdef SSTREE_HEAPPROFILE
heapCon = HeapProfiler::GetHeapConsumption();
#endif
ulong bitsInP;
if (IOaction == load_from && filename != 0)
P = LcpToParentheses::GetBalancedParentheses((string(filename)+".bp").c_str(), bitsInP);
else
P = LcpToParentheses::GetBalancedParentheses(hgt, n, bitsInP);
if (IOaction == save_to && filename != 0)
LcpToParentheses::SaveToFile((string(filename)+".bp").c_str(), P, bitsInP);
#ifdef SSTREE_TIMER
printf("Parentheses sequence created in %.0f seconds.\n", Tools::GetTime());
#ifdef SSTREE_HEAPPROFILE
std::cout << "--> HeapProfiler: " << HeapProfiler::GetHeapConsumption() - heapCon << ", " << HeapProfiler::GetMaxHeapConsumption() << std::endl;
heapCon = HeapProfiler::GetHeapConsumption();
#endif
//printf("Creating CSA with sample rate %d\n", floorLog2n);
//Tools::StartTimer();
#endif
//delete sa;
//sa = new CSA(text, n, floorLog2n);
//hgt->SetSA(sa); // Update SA pointer
#ifdef SSTREE_TIMER
//printf("CSA created in %f seconds.\n", Tools::GetTime());
printf("Creating BitRank\n");
fflush(stdout);
Tools::StartTimer();
#endif
br = new BitRank(P, bitsInP, false);
#ifdef SSTREE_TIMER
printf("BitRank created in %.0f seconds.\n", Tools::GetTime());
#ifdef SSTREE_HEAPPROFILE
std::cout << "--> HeapProfiler: " << HeapProfiler::GetHeapConsumption() - heapCon << ", " << HeapProfiler::GetMaxHeapConsumption() << std::endl;
heapCon = HeapProfiler::GetHeapConsumption();
#endif
printf("Creating Parentheses\n");
fflush(stdout);
Tools::StartTimer();
#endif
Pr = new Parentheses(P, bitsInP, true, br);
#ifdef SSTREE_TIMER
printf("Parentheses created in %.0f seconds.\n", Tools::GetTime());
#ifdef SSTREE_HEAPPROFILE
std::cout << "--> HeapProfiler: " << HeapProfiler::GetHeapConsumption() - heapCon << ", " << HeapProfiler::GetMaxHeapConsumption() << std::endl;
heapCon = HeapProfiler::GetHeapConsumption();
#endif
printf("Creating ReplacePatterns\n");
fflush(stdout);
Tools::StartTimer();
#endif
rpLeaf = new ReplacePattern(1, 8);
rpSibling = new ReplacePattern(0, 8);
#ifdef SSTREE_TIMER
printf("ReplacePatterns created in %.0f seconds.\n", Tools::GetTime());
#ifdef SSTREE_HEAPPROFILE
std::cout << "--> HeapProfiler: " << HeapProfiler::GetHeapConsumption() - heapCon << ", " << HeapProfiler::GetMaxHeapConsumption() << std::endl;
heapCon = HeapProfiler::GetHeapConsumption();
#endif
printf("Creating BitRanks\n");
fflush(stdout);
Tools::StartTimer();
#endif
brLeaf = new BitRank(P, bitsInP, false, rpLeaf); //for ()
brSibling = new BitRank(P, bitsInP, false, rpSibling); //for )(
if (rmqSampleRate < 4)
rmqSampleRate = 4;
#ifdef SSTREE_TIMER
printf("BitRanks created in %.0f seconds.\n", Tools::GetTime());
// printf("<enter>\n");
// std::cin.get();
#ifdef SSTREE_HEAPPROFILE
std::cout << "--> HeapProfiler: " << HeapProfiler::GetHeapConsumption() - heapCon << ", " << HeapProfiler::GetMaxHeapConsumption() << std::endl;
heapCon = HeapProfiler::GetHeapConsumption();
#endif
printf("Creating CRMQ with sample rates %d, %d and %d\n", rmqSampleRate * rmqSampleRate * rmqSampleRate, rmqSampleRate * rmqSampleRate, rmqSampleRate);
Tools::StartTimer();
fflush(stdout);
#endif
rmq = new CRMQ(br, P, bitsInP, rmqSampleRate * rmqSampleRate * rmqSampleRate, rmqSampleRate * rmqSampleRate, rmqSampleRate);
#ifdef SSTREE_TIMER
#ifdef SSTREE_HEAPPROFILE
std::cout << "--> HeapProfiler: " << HeapProfiler::GetHeapConsumption() - heapCon << ", " << HeapProfiler::GetMaxHeapConsumption() << std::endl;
#endif
printf("CRMQ created in %.0f seconds.\n", Tools::GetTime());
fflush(stdout);
#endif
/*
* adicionado por Jean para criar o vetor sa-1
* preparação para o LCE. A idéia é trocar alguns processamentos O(logn) por O(1)
*/
ulong len = numberofnodes(0);
select_br = new int[n+1];
_findclose = new int[len * 4];
//_depth = new int[len * 4];
_enclose = new int[len * 4];
ulong x, w;
for(int v = 0; v <= n; v++){
x = sa->inverse(v);
w = brLeaf->select(x + 1);
select_br[v] = w;
_findclose[w] = Pr->findclose(w);
}
tree_lookup(0, _enclose);
}
void SSTree::tree_lookup(ulong v, int *_enclose){
if (v != 0) add_enclose(v, _enclose);
if(!(isleaf(v))) tree_lookup(v + 1, _enclose);
v = sibling(v);
if (v != 0) tree_lookup(v, _enclose);
}
