function [mappingData] = kelvasa2015_calibratemapping(varargin)
%KELVASA2015_CALIBRATEMAPPING  Produces necessary mappings for localization model
%   Usage:[mappingData] = kelvasa2015_calibratemapping(varargin)
%
%   Input parameters:
%       varargin      : structure with all parameters required for model. If
%                       this is not included, default paramters are loaded.
%
%   The output structure "mappingData" has the following fields:
%        calibHRTFsig                 : 
%                                       NxMxS matrix of signal levels in 
%                                       which N is the range of azimuthal 
%                                       angles overwhich the signal was 
%                                       computed, M is the number of time 
%                                       samples,and S are audio channels.
% 
%        calSpikeDiffPerNeuronPerAzi  : 
%                                       NxMxS matrix of chan2 - chan1 spike
%                                       rate differences in spikes/sec in 
%                                       which N is the range of azimuthal
%                                       angles overwhich the signal was 
%                                       computed, M is the number of 
%                                       simulated AN fibers, and S is the 
%                                       number of time bins.
%                                       
%        calSpikeRatePerNeuronPerLevel: 
%                                       NxM matrix of spike rates in 
%                                       spikes/sec in which N is a range of
%                                       signal levels in dB SPL and M is 
%                                       the number of simulated AN fibers
%
%        calParameters                : 
%                                       structure of model paramters used in
%                                       processing calibration stimulus
% 
% 
%   kelvasa2015_calibratemapping(varargin) processes a user specified 
%   calibration wavfile and extracts the necessary data required to map 
%   simulated bilateral neural outputs onto a predicted azimuthal angle. 
%   This function computes data required by all three localization models 
%   described in (Kelvasa & Dietz(2015)) and can therefore take several
%   hours to process. 
%
%   References:
%     D. Kelvasa and M. Dietz. Auditory model-based sound direction
%     estimation with bilateral cochlear implants. Trends in Hearing,
%     19:2331216515616378, 2015.
%     
%
%   Url: http://amtoolbox.org/amt-1.1.0/doc/modelstages/kelvasa2015_calibratemapping.php

% Copyright (C) 2009-2021 Piotr Majdak, Clara Hollomey, and the AMT team.
% This file is part of Auditory Modeling Toolbox (AMT) version 1.1.0
%
% 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
% GNU General Public License for more details.
%
% You should have received a copy of the GNU General Public License
% along with this program.  If not, see <http://www.gnu.org/licenses/>.

%
%   Authors: 
%            Daryl Kelvasa (daryl.kelvasa@uni-oldenburg.de) 2016
%            Mathias Dietz (mdietz@uwo.ca) 2016
% 
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

definput.import={'kelvasa2015'};
[flags,kv]  = ltfatarghelper({},definput,varargin);


%% Load HRTF data
HRTF = SOFAload(fullfile((SOFAdbPath),...
                    'kelvasa2015',kv.HRTFfile));
[~,ind_elev] = min(abs(HRTF.SourcePosition(:,2)-kv.HRTFelevation));
[~,ind_dist] = min(abs(HRTF.SourcePosition(:,3)-kv.HRTFsourceDistance));                
ind = find(sum([HRTF.SourcePosition(:,2) == HRTF.SourcePosition(ind_elev,2),...
                HRTF.SourcePosition(:,3) == HRTF.SourcePosition(ind_dist,3)],2)...
                ==2);           
HRTFnew.SourcePosition = HRTF.SourcePosition(ind,:);
HRTFnew.Data.IR = HRTF.Data.IR(ind,kv.HRTFchannels,:);   
HRTFnew.Data.SamplingRate = HRTF.Data.SamplingRate;
HRTF = HRTFnew;

%% Set dB SPL offset
dboffset=71.778;

%% Main Code
%Initialize variables
              
[signal, fs] = amt_load('kelvasa2015',kv.localizationModelCalibWav);
signal = signal(1:6*fs,:);
sigLengthSec = size(signal,1)/fs;
signal = resample(signal,kv.FS_ACE,fs);
numWindows = sigLengthSec/kv.timeWindowSec; 
numNeurons = kv.N_nervecells;
spikeRatePerNeuron = zeros(2,numNeurons,numWindows);
spikeDiffPerNeuronPerAzi = zeros(numel(kv.azis),numNeurons,numWindows);


%% Calibration of the AN Linear Rate Difference and Max Likelihood model 
    
for ang = 1 : numel(kv.azis)
tic

    %HRTF filter signal and choose microphone channels       
            [~,ind_ang] = min(abs(HRTF.SourcePosition(:,1)-kv.azis(ang)));

            HRIR = resample(squeeze(HRTF.Data.IR(ind_ang,:,:))',...
                    kv.FS_ACE,HRTF.Data.SamplingRate);        
          
            HRTFchan1 = ifft(fft(signal).*fft(HRIR(:,1),numel(signal)));
            HRTFchan2 = ifft(fft(signal).*fft(HRIR(:,2),numel(signal)));
            HRIR = [HRTFchan1,HRTFchan2];
           
           if kv.azis(ang) == 0
                  temp = HRIR(:,1)./rms(HRIR(:,1));
                  scalor = scaletodbspl(kv.localizationModelCalibStimulusLevelDB,[],dboffset);
                  scalor = rms(temp.*scalor)/rms(HRIR(:,1));
           end
            
            HRIR = HRIR .* scalor;                          

            mappingData.calibHRTFsig(ang,:,:) = HRIR; 
 
%Process signal with CI and AN models for right and left channels
    for chan = 1 : 2
 
        %ACE CI processing strategy
        [electrodogram, vTime] = ...
                                kelvasa2015_ciprocessing(HRIR(:,chan),...
                                kv.FS_ACE,'argimport',flags,kv);
        
        %Fredelake Hohmann CI/AN model
        [APvec] = ...
                                kelvasa2015_anprocessing(electrodogram,...
                                vTime, 'argimport',flags,kv);
                            
        [spikeRatePerNeuron(chan,:,:), ~] = ...
                                kelvasa2015_anbinning(APvec,...
                                sigLengthSec,'argimport',flags,kv);
                            
    end
    
    spikeDiffPerNeuronPerAzi(ang,:,:) = squeeze(spikeRatePerNeuron(2,:,:) - ...
                                        spikeRatePerNeuron(1,:,:));  

a = toc;
timeLeft = round((a*(numel(kv.azis)- ang))/60);
amt_disp(['calibrating with ',kv.localizationModelCalibWav,'.wav at ',...
                num2str(kv.localizationModelCalibStimulusLevelDB),...
                    ' dB Time left (min):', num2str(timeLeft)],'volatile');
 end

amt_disp();

%% Calibration of the AN Rate Level localization model
% %Initialize variables
spikeRatePerNeuronPerLevel = zeros(numel(kv.dBRange), numNeurons);
 
signal =  squeeze(mappingData.calibHRTFsig(1,:,1))';

for level = 1 : numel(kv.dBRange)
tic
                                 
%Adjust signal to level over which to compute rate level slopes (???)
                  temp = signal./rms(signal);
                  scalor = scaletodbspl(kv.dBRange(level),[],dboffset);
                  scalor = rms(temp.*scalor)/rms(signal(:,1));
                  HRTFsig = scalor.*signal;

%Process signal with CI and AN models for right and left channels
        [electrodogram, vTime] = ...
                                kelvasa2015_ciprocessing(HRTFsig,...
                                kv.FS_ACE,'argimport',flags,kv);
                           
        [APvec] = ...
                                kelvasa2015_anprocessing(electrodogram,...
                                vTime,'argimport',flags,kv);
                            
        [spikeRatePerNeuron, ~] = ...
                                kelvasa2015_anbinning(APvec,...
                                sigLengthSec, 'argimport',flags,kv);  
         
        %Compute mean spike rate over all time windows                    
        spikeRatePerNeuronPerLevel(level,:) = mean(spikeRatePerNeuron,2);

        
        a = toc; timeLeft = round((a*(numel(kv.dBRange)- level))/60);
        amt_disp(['Calibration with ',...
                 kv.localizationModelCalibWav,'.wav at ',...
                 num2str(kv.localizationModelCalibStimulusLevelDB),...
                 ' dB Time left (min):', num2str(timeLeft)],'volatile');                        
end
amt_disp();

mappingData.calSpikeDiffPerNeuronPerAzi = spikeDiffPerNeuronPerAzi;
mappingData.calSpikeRatePerNeuronPerLevel = spikeRatePerNeuronPerLevel;
mappingData.calParameters = kv;