bopt_state.cpp

/*
-------------------------------------------------------------------------
   This file is part of BayesOpt, an efficient C++ library for 
   Bayesian optimization.

   Copyright (C) 2011-2015 Ruben Martinez-Cantin <rmcantin@unizar.es>
 
   BayesOpt is free software: you can redistribute it and/or modify it 
   under the terms of the GNU Affero General Public License as published by
   the Free Software Foundation, either version 3 of the License, or
   (at your option) any later version.

   BayesOpt 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
   GNU Affero General Public License for more details.

   You should have received a copy of the GNU Affero General Public License
   along with BayesOpt.  If not, see <http://www.gnu.org/licenses/>.
------------------------------------------------------------------------
*/

#include "bopt_state.hpp"
#include "fileparser.hpp"

#include <fstream>
#include <iostream>

namespace bayesopt
{       
    BOptState::BOptState(): mParameters(){}
    
    void BOptState::saveToFile(std::string filename){
        utils::FileParser fp(filename);
        fp.openOutput();
        loadOrSave(fp);
    }
    
    bool BOptState::loadFromFile(std::string filename, Parameters &program_params){
        utils::FileParser fp(filename);
        if(!fp.fileExists()){
            return false;
        }
        fp.openInput();
        loadOrSave(fp);
        
        // Non-stored fields: verbose_level, log_filename, load_save_flag, load_filename, 
        //                    save_filename, random_seed.
        
        // Fill non-stored values with program_params values:
        mParameters.verbose_level = program_params.verbose_level;
        mParameters.log_filename = program_params.log_filename;
        mParameters.load_save_flag = program_params.load_save_flag;
        mParameters.load_filename = program_params.load_filename;
        mParameters.save_filename = program_params.save_filename;
        mParameters.random_seed = program_params.random_seed;
        
        // n_iterations can be taken from program_params if value is greater than current iteration
        if(program_params.n_iterations > mCurrentIter){
            mParameters.n_iterations = program_params.n_iterations;
        }
        
        return true;
    }
    
    void BOptState::loadOrSave(utils::FileParser &fp){    
        fp.readOrWrite("mCurrentIter", mCurrentIter);
        fp.readOrWrite("mCounterStuck", mCounterStuck);
        fp.readOrWrite("mYPrev", mYPrev);
        
        fp.readOrWrite("mParameters.n_iterations", mParameters.n_iterations);
        fp.readOrWrite("mParameters.n_inner_iterations", mParameters.n_inner_iterations);
        fp.readOrWrite("mParameters.n_init_samples", mParameters.n_init_samples);
        fp.readOrWrite("mParameters.n_iter_relearn", mParameters.n_iter_relearn);
        fp.readOrWrite("mParameters.init_method", mParameters.init_method);

        fp.readOrWrite("mParameters.surr_name", mParameters.surr_name);
        fp.readOrWrite("mParameters.sigma_s", mParameters.sigma_s);
        fp.readOrWrite("mParameters.noise", mParameters.noise);
        fp.readOrWrite("mParameters.alpha", mParameters.alpha);
        fp.readOrWrite("mParameters.beta", mParameters.beta);
        
        // Enums
        if(fp.isReading()){
            mParameters.sc_type = str2score(fp.read("mParameters.sc_type").c_str());
            mParameters.l_type = str2learn(fp.read("mParameters.l_type").c_str());
        }
        else if(fp.isWriting()){
            fp.write("mParameters.sc_type",score2str(mParameters.sc_type));
            fp.write("mParameters.l_type", learn2str(mParameters.l_type));
        }
        
        fp.readOrWrite("mParameters.l_all", mParameters.l_all);
        fp.readOrWrite("mParameters.epsilon", mParameters.epsilon);
        fp.readOrWrite("mParameters.force_jump", mParameters.force_jump);
        
        fp.readOrWrite("mParameters.kernel.name", mParameters.kernel.name);
        fp.readOrWrite("mParameters.kernel.hp_mean", mParameters.kernel.hp_mean);
        fp.readOrWrite("mParameters.kernel.hp_std", mParameters.kernel.hp_std);
        
        fp.readOrWrite("mParameters.mean.name", mParameters.mean.name);
        fp.readOrWrite("mParameters.mean.coef_mean", mParameters.mean.coef_mean);
        fp.readOrWrite("mParameters.mean.coef_std", mParameters.mean.coef_std);
        
        fp.readOrWrite("mParameters.crit_name", mParameters.crit_name);
        fp.readOrWrite("mParameters.crit_params", mParameters.crit_params);
        
        fp.readOrWrite("mY", mY);
        fp.readOrWrite("mX", mX);
    } 
} //namespace bayesopt