void getNB(HashGraph* G, int v, int distance, hash_set<int>& result, vector<bool>& mark)
{
assert(distance==1 || distance==2);
EdgeMap* p_neighbors=G->getNeighbors(v);
EdgeMap::iterator pnb;
for (pnb=p_neighbors->begin(); pnb!=p_neighbors->end(); pnb++)
{
UINT w=pnb->first;
if(!mark[w])
result.insert(w);
}
if(distance==1)
return;
hash_set<int>::iterator p;
vector<int> temp;
for(p=result.begin(); p!=result.end(); p++)
temp.push_back(*p);
int imnb_size=result.size();
//result.clear();
for(int i=0; i<imnb_size; i++)
{
p_neighbors=G->getNeighbors(temp[i]);
for (pnb=p_neighbors->begin(); pnb!=p_neighbors->end(); pnb++)
{
UINT w=pnb->first;
if(!mark[w])
result.insert(w); //automatically handel duplication.
}
}
for(int i=0; i<imnb_size; i++)
result.erase(temp[i]);
}
void print(hash_set &S) {
iterator_t p = S.begin();
while(p!=S.end()) {
cout << S.retrieve(p) << " ";
p = S.next(p);
}
cout << endl;
}
void LinearSNNClassifier::checkgradRowSparse(const vector<Example>& examples, mat& Wd, const mat& gradWd, const string& mark, int iter,
const hash_set<int>& sparseRowIndexes, const mat& ft) {
//Random randWdRowcheck = new Random(iter + "Row".hashCode() + hash));
int charseed = mark.length();
for (int i = 0; i < mark.length(); i++) {
charseed = (int) (mark[i]) * 5 + charseed;
}
srand(iter + charseed);
std::vector<int> idRows, idCols;
idRows.clear();
idCols.clear();
if (sparseRowIndexes.empty()) {
for (int i = 0; i < Wd.n_rows; ++i)
idRows.push_back(i);
} else {
hash_set<int>::iterator it;
for (it = sparseRowIndexes.begin(); it != sparseRowIndexes.end(); ++it)
idRows.push_back(*it);
}
for (int idx = 0; idx < Wd.n_cols; idx++)
idCols.push_back(idx);
random_shuffle(idRows.begin(), idRows.end());
random_shuffle(idCols.begin(), idCols.end());
int check_i = idRows[0], check_j = idCols[0];
double orginValue = Wd(check_i, check_j);
Wd(check_i, check_j) = orginValue + 0.001;
double lossAdd = 0.0;
for (int i = 0; i < examples.size(); i++) {
Example oneExam = examples[i];
lossAdd += computeScore(oneExam);
}
Wd(check_i, check_j) = orginValue - 0.001;
double lossPlus = 0.0;
for (int i = 0; i < examples.size(); i++) {
Example oneExam = examples[i];
lossPlus += computeScore(oneExam);
}
double mockGrad = (lossAdd - lossPlus) / (0.002 * ft(check_i, check_j));
mockGrad = mockGrad / examples.size();
double computeGrad = gradWd(check_i, check_j);
printf("Iteration %d, Checking gradient for %s[%d][%d]:\t", iter, mark.c_str(), check_i, check_j);
printf("mock grad = %.18f, computed grad = %.18f\n", mockGrad, computeGrad);
Wd(check_i, check_j) = orginValue;
}
int main(){
for(int i=0;i<100;i++){
tab.insert(2*i + 1);
}
for(hash_set<int>::iterator iter = tab.begin();
iter!=tab.end();
iter++){
cout<<(*iter)<<" ";
}
cout<<endl;
return 0;
}
void Query::mapNodes(vector<Neighborhood>& nbs, HashGraph* Gdb, OrthologInfoList* pOrthinfolist_db, hash_set<int>& Sq, hash_set<int>& Sdb, GraphMatch* gm, Queue& Q, hash_set<int>& nodesInQ, vector<bool>& mark, vector<bool>& dbmark)
{
hash_set<int>::iterator p;
hash_map< int, vector<int>, inthash > lmap_q, lmap_db;
for(p=Sq.begin(); p!=Sq.end(); p++)
{
for(unsigned int i=0; i<(*pOrthinfolist)[*p].size(); i++)
{
if((*pOrthinfolist)[*p][i]==0)
continue;
lmap_q[(*pOrthinfolist)[*p][i]].push_back(*p);
}
}
for(p=Sdb.begin(); p!=Sdb.end(); p++)
{
if(!dbmark[*p])
{
for(unsigned int i=0; i<(*pOrthinfolist_db)[*p].size(); i++)
{
if((*pOrthinfolist_db)[*p][i]==0)
continue;
lmap_db[(*pOrthinfolist_db)[*p][i]].push_back(*p);
}
}
}
hash_map<int, vector<int>, inthash>::iterator iter, iter2;
for(iter=lmap_q.begin(); iter!=lmap_q.end(); iter++)
{
iter2=lmap_db.find(iter->first);
if(iter2==lmap_db.end())
continue;
else
mapNodesHelp(nbs, Gdb, pOrthinfolist_db, iter->second, iter2->second, gm, Q, nodesInQ, mark, dbmark);
}
}
int trim_entropy_filter(vector<string>* keep_words, hash_set<string>& cad_words_set, WordInfoMap& wordinfo_map)
{
keep_words->reserve(cad_words_set.size());
for (hash_set<string>::iterator it = cad_words_set.begin(); it != cad_words_set.end(); ++it) {
WordInfoMap::iterator it_map = wordinfo_map.find(*it);
if (it_map == wordinfo_map.end()) {
fprintf(stderr, "WARNING, word[%s] in cad_word, not in word_info", it->c_str());
continue;
}
if (it_map->first.size() <= WORD_LEN - 4 && it_map->second.calc_is_keep()) {
keep_words->push_back(*it);
}
}
return 0;
}