SEXP
R_ocr_alternatives(SEXP filename, SEXP r_vars, SEXP r_level)
{
SEXP ans = R_NilValue;
Pix *image = pixRead(CHAR(STRING_ELT(filename, 0)));
int i;
tesseract::TessBaseAPI *api = new tesseract::TessBaseAPI();
api->Init(NULL, "eng");
api->SetImage(image);
SEXP r_optNames = GET_NAMES(r_vars);
for(i = 0; i < Rf_length(r_vars); i++)
api->SetVariable(CHAR(STRING_ELT(r_optNames, i)), CHAR(STRING_ELT(r_vars, i)));
api->Recognize(0);
tesseract::ResultIterator* ri = api->GetIterator();
tesseract::PageIteratorLevel level = (tesseract::PageIteratorLevel) INTEGER(r_level)[0];
int n = 1;
while(ri->Next(level))
n++;
ri = api->GetIterator();
SEXP names;
PROTECT(names = NEW_CHARACTER(n));
PROTECT(ans = NEW_LIST(n));
i = 0;
do {
const char* word = ri->GetUTF8Text(level);
float conf = ri->Confidence(level);
SET_STRING_ELT(names, i, Rf_mkChar(word));
SET_VECTOR_ELT(ans, i, getAlternatives(ri, word, conf));
delete[] word;
i++;
} while (ri->Next(level));
SET_NAMES(ans, names);
UNPROTECT(2);
return(ans);
}
bool ChemicalValidity::optimizeAtomGroup(AtomRefs& data) const
{
logEnterFunction();
// here we know that data contains the references to the bad atoms
// we can do anything with them.
// HACK for "R", probably outdated
if (data.size() == 1 && data[0]->getPrintableForm() == "R")
{
return true;
}
std::vector<IterationRecord> bruteforce;
std::string molecule;
for (size_t u = 0; u < data.size(); u++)
{
std::string s = data[u]->getPrintableForm(false);
molecule += s;
for (size_t v = 0; v < data[u]->labels.size(); v++)
{
IterationRecord irec;
irec.atom = u;
irec.pos = v;
irec.counter = 0;
irec.alts = getAlternatives(data[u]->labels[v].selected_character, data[u]->labels[v].alternatives);
getLogExt().append("alternatives", irec.alts);
bruteforce.push_back(irec);
}
}
getLogExt().append("bad part", molecule);
if (bruteforce.empty())
return false;
while (true)
{
// increment counter
bruteforce[0].counter++;
size_t idx = 0;
while (bruteforce[idx].counter >= (int)bruteforce[idx].alts.size())
{
bruteforce[idx].counter = 0;
idx++;
if (idx >= bruteforce.size())
goto not_found;
bruteforce[idx].counter++;
}
// update atoms
for (size_t u = 0; u < bruteforce.size(); u++)
{
if (!bruteforce[u].alts.empty())
{
CharacterRecognitionEntry& cg = data[bruteforce[u].atom]->labels[bruteforce[u].pos];
cg.selected_character = bruteforce[u].alts[bruteforce[u].counter];
}
}
// assemble string
std::string test;
for (size_t u = 0; u < data.size(); u++)
{
test += data[u]->getPrintableForm(false);
}
getLogExt().append("check string", test);
if (calcSplitProbability(optimalSplit(test, elements.names)) > EPS)
{
getLogExt().append("passed!", test);
return true;
}
}
not_found: ;
return false;
}
