symboldictionary.h

/*
    Copyright (C) 2010 Tomasz Śniatowski, Adam Radziszewski
    Part of the libcorpus2 project

    This program is free software; you can redistribute it and/or modify it
under the terms of the GNU General Public License as published by the Free
Software Foundation; either version 3 of the License, or (at your option)
any later version.

    This program is distributed in the hope that it will be useful, but
WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
or FITNESS FOR A PARTICULAR PURPOSE. 

    See the LICENSE and COPYING files for more details.
*/

#ifndef LIBCORPUS2_SYMBOLDICTIONARY_H
#define LIBCORPUS2_SYMBOLDICTIONARY_H

#include <libcorpus2/typedefs.h>
#include <vector>
#include <string>
#include <boost/range.hpp>
#include <map>
#include <iostream>

namespace Corpus2 {

/**
 * A template for string-index dictionaries offering lookups in both ways.
 *
 * A dictionary is created from a set of strings, which is then mapped to
 * indices which can be used.
 *
 * The class is templated on the index type, which should be some form of
 * an integer. The amount of used indices should be less than the maximum
 * value of the type. There should be no empty strings or duplicate
 * strings.
 */
template <typename IndexT>
class SymbolDictionary
{
public:
	/// Empty dictionary constructor
	SymbolDictionary();

	/// Load data into the dictionary
	void load_data(const std::vector<std::string>& data);

	/// Load data into the dictionary. The strings in the vector are
	/// assumed to be already sorted.
	void load_sorted_data(const std::vector<std::string>& data);

	/// Check if an index is valid in this dictionary
	bool is_id_valid(IndexT idx) const;

	/// Getter for the size of this dictionary
	size_t size() const;

	/**
	 * Get the index for a given string identifier, const char* version.
	 *
	 * Essentially a wrapper for the range overload, needed to avoid
	 * confusion when a static char array is passed there.
	 */
	IndexT get_id(const char* c) const;

	/**
	 * Get the index for a given string identifier, range version.
	 *