hash_generator.hpp

#ifndef HASH_GENERATOR_HDR
#define HASH_GENERATOR_HDR

#include <string>
#include <map>
#include <vector>
#include <boost/shared_ptr.hpp>
#include <boost/algorithm/string/predicate.hpp>
#include "concordia/word_map.hpp"
#include "concordia/common/config.hpp"
#include "concordia/sentence_tokenizer.hpp"
#include "concordia/concordia_config.hpp"
#include "concordia/concordia_exception.hpp"


class HashGenerator {
public:
    explicit HashGenerator(std::string indexPath,
                           boost::shared_ptr<ConcordiaConfig> config)
                                            throw(ConcordiaException);

    virtual ~HashGenerator();

    TokenizedSentence generateHash(const std::string & sentence,
                                   bool byWhitespace = false)
                                       throw(ConcordiaException);

    TokenizedSentence generateTokens(const std::string & sentence,
                                     bool byWhitespace = false)
                                       throw(ConcordiaException);

    void serializeWordMap();

    void clearWordMap();

private:
    boost::shared_ptr<WordMap> _wordMap;

    boost::shared_ptr<SentenceTokenizer> _sentenceTokenizer;

    std::string _wordMapFilePath;
};

#endif