void split_to_pair(const string& str, vector< pair<string, string> >& vecPair)
{
assert(vecPair.empty());
vector<string> vec;
split_bychar(str, vec);
convert_to_pair(vec, vecPair);
}
bool DepSRL::IsPosPattern(
int intBegin,
int intEnd,
const vector<string>& vecPos,
const string& strPattern) const
{
vector<string> vecItem;
split_bychar(strPattern, vecItem, C_PATTERN_SEP);
for (int index = intBegin; index < intEnd; index++)
{
if ( find(vecItem.begin(), vecItem.end(), vecPos.at(index)) == vecItem.end() )
{
return 0;
}
}
return 1;
}
void Labeler::readWordEmbeddings(const string& inFile, NRMat<dtype>& wordEmb) {
static ifstream inf;
if (inf.is_open()) {
inf.close();
inf.clear();
}
inf.open(inFile.c_str());
static string strLine, curWord;
static int wordId;
//find the first line, decide the wordDim;
while (1) {
if (!my_getline(inf, strLine)) {
break;
}
if (!strLine.empty())
break;
}
int unknownId = m_wordAlphabet.from_string(unknownkey);
static vector<string> vecInfo;
split_bychar(strLine, vecInfo, ' ');
int wordDim = vecInfo.size() - 1;
std::cout << "word embedding dim is " << wordDim << std::endl;
m_options.wordEmbSize = wordDim;
wordEmb.resize(m_wordAlphabet.size(), wordDim);
wordEmb = 0.0;
curWord = normalize_to_lowerwithdigit(vecInfo[0]);
wordId = m_wordAlphabet.from_string(curWord);
hash_set<int> indexers;
dtype sum[wordDim];
int count = 0;
bool bHasUnknown = false;
if (wordId >= 0) {
count++;
if (unknownId == wordId)
bHasUnknown = true;
indexers.insert(wordId);
for (int idx = 0; idx < wordDim; idx++) {
dtype curValue = atof(vecInfo[idx + 1].c_str());
sum[idx] = curValue;
wordEmb[wordId][idx] = curValue;
}
} else {
for (int idx = 0; idx < wordDim; idx++) {
sum[idx] = 0.0;
}
}
while (1) {
if (!my_getline(inf, strLine)) {
break;
}
if (strLine.empty())
continue;
split_bychar(strLine, vecInfo, ' ');
if (vecInfo.size() != wordDim + 1) {
std::cout << "error embedding file" << std::endl;
}
curWord = normalize_to_lowerwithdigit(vecInfo[0]);
wordId = m_wordAlphabet.from_string(curWord);
if (wordId >= 0) {
count++;
if (unknownId == wordId)
bHasUnknown = true;
indexers.insert(wordId);
for (int idx = 0; idx < wordDim; idx++) {
dtype curValue = atof(vecInfo[idx + 1].c_str());
sum[idx] += curValue;
wordEmb[wordId][idx] += curValue;
}
}
}
if (!bHasUnknown) {
for (int idx = 0; idx < wordDim; idx++) {
wordEmb[unknownId][idx] = sum[idx] / count;
}
count++;
std::cout << unknownkey << " not found, using averaged value to initialize." << std::endl;
}
int oovWords = 0;
int totalWords = 0;
for (int id = 0; id < m_wordAlphabet.size(); id++) {
if (indexers.find(id) == indexers.end()) {
oovWords++;
for (int idx = 0; idx < wordDim; idx++) {
wordEmb[id][idx] = wordEmb[unknownId][idx];
}
}
totalWords++;
}
std::cout << "OOV num is " << oovWords << ", total num is " << m_wordAlphabet.size() << ", embedding oov ratio is " << oovWords * 1.0 / m_wordAlphabet.size()
<< std::endl;
}
int Segmentor::allWordAlphaEmb(const string& inFile, NRMat<dtype>& emb) {
cout << "All word alphabet and emb creating..." << endl;
hash_map<string, int> word_stat;
static ifstream inf;
if (inf.is_open()) {
inf.close();
inf.clear();
}
inf.open(inFile.c_str());
static string strLine, curWord;
static int wordId;
static vector<string> vecInfo;
vector<string> allLines;
int wordDim = 0;
while (1) {
if (!my_getline(inf, strLine)) {
break;
}
if (!strLine.empty()){
split_bychar(strLine, vecInfo, ' ');
if(wordDim == 0){
wordDim = vecInfo.size() - 1;
std::cout << "allword embedding dim is " << wordDim << std::endl;
}
curWord = normalize_to_lowerwithdigit(vecInfo[0]);
word_stat[curWord]++;
allLines.push_back(strLine);
}
}
m_classifier.addToAllWordAlphabet(word_stat);
cout << "Remain all word num: " << m_classifier.fe._allwordAlphabet.size() << endl;
emb.resize(m_classifier.fe._allwordAlphabet.size(), wordDim);
emb = 0.0;
int unknownId = m_classifier.fe._allwordAlphabet.from_string(m_classifier.fe.unknownkey);
dtype sum[wordDim];
int count = 0;
bool bHasUnknown = false;
for (int idx = 0; idx < wordDim; idx++) {
sum[idx] = 0.0;
}
for(int idx = 0; idx < allLines.size(); idx++){
split_bychar(allLines[idx], vecInfo, ' ');
if (vecInfo.size() != wordDim + 1) {
std::cout << "error embedding file" << std::endl;
}
curWord = normalize_to_lowerwithdigit(vecInfo[0]);
wordId = m_classifier.fe._allwordAlphabet.from_string(curWord);
if (wordId >= 0) {
count++;
if (unknownId == wordId)
bHasUnknown = true;
for (int idx = 0; idx < wordDim; idx++) {
dtype curValue = atof(vecInfo[idx + 1].c_str());
sum[idx] += curValue;
emb[wordId][idx] += curValue;
}
}
else{
std::cout << "read all word embedding strange...." << std::endl;
}
}
if (!bHasUnknown) {
for (int idx = 0; idx < wordDim; idx++) {
emb[unknownId][idx] = sum[idx] / count;
}
count++;
std::cout << unknownkey << " not found, using averaged value to initialize." << std::endl;
}
return 0;
}
