IndexPtr GetValidIndexPointer(char *word, int pos)
{
IndexPtr idx;
char *morphword;
idx = getindex(word, pos);
if (idx == NULL) {
if ((morphword = morphstr(word, pos)) != NULL)
while (morphword) {
if ((idx = getindex(morphword, pos)) != NULL) break;
morphword = morphstr(NULL, pos);
}
}
return (idx);
}
void
morphMe(char *str) {
char *temp;
if(morphinit()<0) printf("\nError, cant init morphy");
//wninit();
printf("\n%s", morphword(str, NOUN));
while (temp != NULL) {
temp=morphstr(str, NOUN);
printf("\n%s", temp);
}
re_morphinit();
}
vector<PopWord> TopTensSearch::parsePage(string buffer, string word)
{
vector<PopWord> results;
set<PopWord*, DereferenceLess> unique;
// boost::smatch what;
// boost::regex expr(regex);
//parse the XML document
rapidxml::xml_document<> doc;
doc.parse<0>((char*)buffer.c_str());
rapidxml::xml_node<> *node = doc.first_node();
//Filter the titles
string exp(".*st "+word+"s?( (?!(in|about|from)).*|$)|.(?!\\w+s)* "+word+"s( (?!.*(in|about|from)).*|$)|.* "+word+"s? ((?!color)|(\\w+))s( (?!.*(in|about|from)).*|$)");
boost::regex e(exp, boost::regex::icase);
wninit();
for (rapidxml::xml_node<> *list = node->first_node(); list; list = list->next_sibling())
{
//The list has:
// * title
// * link
// * items
// printf("Title: %s\n", list->first_node("title")->value());
const string test = list->first_node("title")->value();
//Determine if the title is any good
if(boost::regex_match(test, e))
{
int count = 0;
for (rapidxml::xml_node<> *attr = list->first_node("items")->first_node(); attr && count < 5; attr = attr->next_sibling())
{
std::string item = (std::string) attr->value();
boost::regex re(" (in |from |of |- |\\(|:)", boost::regbase::icase);
boost::sregex_token_iterator i(item.begin(), item.end(), re, -1);
boost::sregex_token_iterator j;
std::string found = *i++;
if (i!=j)
{
boost::replace_all(found, " ", "_");
//PopWord* pop = new PopWord(found, test, *i);
set<PopWord*>::iterator it = unique.insert(unique.begin(), new PopWord(found, test, *i));
// if (it != unique.end())
// {
// cout << (*it)->getCount();
(*it)->incrementCount();
// cout << "\t1 |" << (*it)->getStem() << "|\t" << *i << " " << (*it)->getCount() << endl;
// }
// else
// unique.insert(pop);
}
else
{
boost::replace_all(item, " ", "_");
bool inWN = false;
for (int j=1; j <= 4; j++)
{
//If the word is defined for the POS
if (is_defined((char*)item.c_str(), j) != 0)
{
set<PopWord*>::iterator it = unique.insert(unique.begin(), new PopWord(item, test));
(*it)->incrementCount();
inWN=true;
// cout << "\t\t2 " << item << endl;
break;
}
else
{
char* morph = morphstr((char*)item.c_str(),j);
if (morph != NULL)
{
set<PopWord*>::iterator it = unique.insert(unique.begin(), new PopWord(item, test));
(*it)->incrementCount();
inWN=true;
// cout << "\t\t3 " << item << endl;
break;
}
}
}
// if(!inWN)
// std::cout << "Not used\t\t4 "<<item << std::endl;
}
}
}
}
//Iterate through the unique set. Check how apparent the concept is in wikipedia
int maxCount = 0;
set<PopWord*>::iterator it;
for (it=unique.begin(); it !=unique.end(); it++)
{
PopWord ttr = *(*it);
// cout << ttr.getWord() << endl;
//int count = -1;
if (false)
{
string buffer = requestURL("http://en.wikipedia.org/w/api.php?format=json&action=parse&page="+ttr.getStem()+"&prop=links&redirects");
boost::regex re("[\"\\s(:]" + word + "s?[)\"\\s]", boost::regbase::icase);
boost::sregex_token_iterator i(buffer.begin(), buffer.end(), re, -1);
boost::sregex_token_iterator j;
// cout << "-->\t" << buffer << endl;
while (i != j)
{
// cout << "\t" << *i++ << endl;
ttr.incrementCount();
}
// cout << count << endl << endl;
}
// ttr.setCount(count);
if (ttr.getCount() > maxCount)
maxCount=ttr.getCount();
results.push_back(ttr);
}
for (int i =0; i < results.size(); i++)
{
results[i].setCount(results[i].getCount()/maxCount);
}
return results;
}
