/**
* Calculate score of a given UYiV. The score is based on the formulas written in the inform.
* @param uyv
* @return score
*/
double Dencoder::score(shared_ptr<UYiV> uyv) const
{
if (uyv == NULL) return -1;
WordList* y = uyv->getYList();
//if the yList is NULL return score 0
assert(y != NULL);
if (y == NULL) return 0;
// if (y->getSize() == 0) return -1;
int p, q, k, l, pipes, I, encoding, total;
p = uyv->getWordsPosPairs()->size();
//This is the complete encoding... And I just to count the encoding of the new uyv...
encoding = 0;
q = y->getWordVec().size();
k = uyv->getU()->getSize();
l = uyv->getV()->getSize();
pipes = (q-1) ;
I = 1;
// the terms with the + are our gains, the - our penalties
// total = - getProdElectionSize() + p*k + p*l + y->sumWordsLengthWithoutRepeats() - y->sumWordsLengthWithRepeats() - p - 1 - k - l - 1 - 1 - (q - 1);
total = p*k + p*l + y->sumWordsLengthWithRepeats() -
y->sumWordsLengthWithoutRepeats() -
encoding -
p - 1 - k - l - 1 - I - pipes ;
assert(y->sumWordsLengthWithRepeats()>=y->sumWordsLengthWithoutRepeats());
return total;
}
/**
* replace the given yield and generate the productions needed
* @param uyv
*/
void Dencoder::replaceYield(shared_ptr<UYiV> uyv)
{
shared_ptr<Word> u = uyv->getU();
int * uInt = u->getSubSequence();
shared_ptr<Word> v = uyv->getV();
int * vInt = v->getSubSequence();
int uSize = u->getSize();
int vSize = v->getSize();
int ySize = 0;
int *yInt;
int currentPos = 0;
WordList* y = uyv->getYList();
assert(y!= NULL);
// set<shared_ptr<Word> > * yWords = y->getWordSet();
vector< pair<shared_ptr<Word>, int> > vecWord = y->getWordVec();
int lenghtInside = y->sumWordsLengthWithoutRepeats();
int I = symbol_->getNonTerminal();
int N = symbol_->getNonTerminal();
std::queue<int> initialQ;
prodElectionOrder_[I] = initialQ ;
replaceNonTerminal(uyv, N);
// U size + I NonTerminal + v Size
int * nRHS = new int[uSize+1+vSize];
// Rule for the insides + y->size() are the pipes
int * iRHS = new int[lenghtInside+vecWord.size()];
//Add u symbols
for(int i = 0; i < uSize; ++i)
{
nRHS[i] = uInt[i];
}
//Add I nonTerminal
nRHS[uSize] = I;
//Add v symbols
for(int i = 0; i < vSize; ++i)
{
nRHS[uSize+1+i] = vInt[i];
}
//Add y simbols
for(vector< pair<shared_ptr<Word>, int> >::iterator it = vecWord.begin(); it != vecWord.end(); ++it)
{
shared_ptr<Word> y = (*it).first;
yInt = y->getSubSequence();
ySize = y->getSize();
//cannot be 0
// assert(ySize!=0);
for (int i = 0; i < ySize; ++i)
{
iRHS[currentPos+i] = yInt[i];
}
iRHS[currentPos + ySize] = 2;
currentPos += ySize+1;
// cout << endl << *y << endl << (*it).second << endl;
}
// int sProdSize = getSeq()->getSize() - y->sumWordsLengthWithRepeats() - n*(uSize+vSize) + n;
//Create Productions:
shared_ptr<Production> p1 = make_shared<Production>(I, iRHS, lenghtInside+y->getWordSet()->size()-1);
shared_ptr<Production> p2 = make_shared<Production>(N, nRHS, uSize+1+vSize);
grammar_->addProd(p1);
grammar_->addProd(p2);
}