void Misc::removeStopwords(std::string& text) {
if (stopwords.empty())
loadStopwords();
// Voy a recorrer palabra por palabra del string text
std::string word;
std::string::iterator it = text.begin();
while (it != text.end()) {
int itPos = std::distance(text.begin(), it);
int nextSpacePos = text.find(' ', itPos);
word = text.substr(itPos, nextSpacePos - itPos);
std::vector<std::string>::iterator stopwordIt =
std::find(stopwords.begin(), stopwords.end(), word);
if (stopwordIt != stopwords.end()) {
// Se encontro la stopword
text.erase(itPos, nextSpacePos - itPos + 1);
} else {
it += word.size();
if (it != text.end())
++it; // Me salteo el espacio (si hay)
}
}
}
int main(int argc, char* argv[])
{
if (argc != 6) {
cerr << "[usage] <sentencesText.buf> <patterns.csv> <stopwords.txt> <stopwordsFromText.txt> <final.csv>" << endl;
return -1;
}
selftest();
loadSentences(argv[1]);
loadPattern(argv[2]);
loadStopwords(argv[3], argv[4]);
int corpusTokensN = 0;
for (size_t sentenceID = 0; sentenceID < sentences.size(); ++ sentenceID) {
const string &text = sentences[sentenceID];
string alpha = text;
for (size_t i = 0; i < alpha.size(); ++ i) {
if (isalpha(alpha[i])) {
alpha[i] = tolower(alpha[i]);
} else {
if (alpha[i] != '\'') {
alpha[i] = ' ';
}
}
}
corpusTokensN += splitBy(alpha, ' ').size();
string outsideText = alpha;
if (sentenceID > 0) {
outsideText += " " + sentences[sentenceID - 1];
}
if (sentenceID + 1 < sentences.size()) {
outsideText += " " + sentences[sentenceID + 1];
}
for (size_t i = 0; i < outsideText.size(); ++ i) {
if (isalpha(outsideText[i])) {
outsideText[i] = tolower(outsideText[i]);
} else {
outsideText[i] = ' ';
}
}
vector<string> outside = splitBy(outsideText, ' ');
unordered_map<string, int> outsideCnt;
FOR (token, outside) {
++ outsideCnt[*token];
}
vector< pair<int, int> > positions;
tree.search(" " + alpha + " ", positions);
unordered_map<string, int> patternCnt;
FOR (pos, positions) {
int st = pos->first;
int ed = pos->second - 2;
string pattern = alpha.substr(st, ed - st);
++ patternCnt[pattern];
}
FOR (pos, positions) {
int st = pos->first;
int ed = pos->second - 2;
string pattern = alpha.substr(st, ed - st);
vector<string> tokens = splitBy(pattern, ' ');
unordered_map<string, int> tokenCnt;
int delta = patternCnt[pattern];
for (size_t i = 0; i < tokens.size(); ++ i) {
tokenCnt[tokens[i]] += delta;
}
for (size_t i = 0; i < tokens.size(); ++ i) {
if (outsideCnt[tokens[i]] > tokenCnt[tokens[i]]) {
f[pattern][i] += 1;
sumOutside[pattern][i] += outsideCnt[tokens[i]] - tokenCnt[tokens[i]];
}
}
total[pattern] += 1;
if (st > 0 && ed < (int)text.size()) {
if (text[st - 1] == '(' && text[ed] == ')') {
parenthesis[pattern] += 1;
}
if (text[st - 1] == '"' && text[ed] == '"') {
quote[pattern] += 1;
}
}
bool found = false;
for (int i = st; i < ed && !found; ++ i) {
found |= text[i] == '-';
}
dash[pattern] += found;
bool valid = true;
for (int i = st; i < ed && valid; ++ i) {
if (isalpha(alpha[i]) && (i == st || alpha[i - 1] == ' ')) {
if (text[i] < 'A' && text[i] > 'Z') {
valid = false;
}
}
}
capital[pattern] += valid;
}
