69 lines
2.3 KiB
C++
69 lines
2.3 KiB
C++
#include "concordia/hash_generator.hpp"
|
|
#include "concordia/common/utils.hpp"
|
|
#include "concordia/token_annotation.hpp"
|
|
|
|
#include <boost/filesystem.hpp>
|
|
#include <boost/archive/binary_oarchive.hpp>
|
|
#include <boost/archive/binary_iarchive.hpp>
|
|
#include <boost/algorithm/string.hpp>
|
|
#include <boost/foreach.hpp>
|
|
|
|
#include <fstream>
|
|
|
|
HashGenerator::HashGenerator(boost::shared_ptr<ConcordiaConfig> config)
|
|
throw(ConcordiaException) :
|
|
_wordMapFilePath(config->getWordMapFilePath()),
|
|
_wordMap(boost::shared_ptr<WordMap>(new WordMap)),
|
|
_sentenceTokenizer(boost::shared_ptr<SentenceTokenizer>(
|
|
new SentenceTokenizer(config))) {
|
|
if (boost::filesystem::exists(_wordMapFilePath)) {
|
|
std::ifstream ifs(_wordMapFilePath.c_str(), std::ios::binary);
|
|
boost::archive::binary_iarchive ia(ifs);
|
|
boost::shared_ptr<WordMap> restoredWordMap(new WordMap);
|
|
ia >> *_wordMap;
|
|
}
|
|
}
|
|
|
|
HashGenerator::~HashGenerator() {
|
|
}
|
|
|
|
std::vector<INDEX_CHARACTER_TYPE> HashGenerator::generateHash(
|
|
const std::string & sentence) throw(ConcordiaException) {
|
|
std::vector<INDEX_CHARACTER_TYPE> result;
|
|
|
|
boost::shared_ptr<TokenizedSentence> ts = _sentenceTokenizer->tokenize(sentence);
|
|
ts->generateHash(_wordMap);
|
|
|
|
if (ts->getTokens().size() > Utils::maxSentenceSize) {
|
|
throw ConcordiaException("Trying to add too long sentence.");
|
|
}
|
|
|
|
return ts->getCodes();
|
|
}
|
|
|
|
std::vector<std::string> HashGenerator::generateTokenVector(
|
|
const std::string & sentence) {
|
|
boost::shared_ptr<TokenizedSentence> ts = _sentenceTokenizer->tokenize(sentence);
|
|
std::vector<std::string> tokenTexts;
|
|
BOOST_FOREACH(TokenAnnotation annotation, ts->getAnnotations()) {
|
|
if (annotation.getType() == TokenAnnotation::WORD ||
|
|
annotation.getType() == TokenAnnotation::NE) {
|
|
tokenTexts.push_back(annotation.getValue());
|
|
}
|
|
}
|
|
return tokenTexts;
|
|
}
|
|
|
|
|
|
void HashGenerator::serializeWordMap() {
|
|
std::ofstream ofs(_wordMapFilePath.c_str(), std::ios::binary);
|
|
boost::archive::binary_oarchive oa(ofs);
|
|
oa << *_wordMap;
|
|
}
|
|
|
|
void HashGenerator::clearWordMap() {
|
|
_wordMap = boost::shared_ptr<WordMap>(new WordMap);
|
|
boost::filesystem::remove(_wordMapFilePath);
|
|
}
|
|
|