main_final.m

function [] = main_final(beh_type,distinguish_str,beh)
    % This function runs all of the modeling, followed by analysis code.
    % It takes parameters
    % beh_type: 'AGG', 'SIB', or 'BOTH' strings
    % distinugish_str: how to distinguish this run's folder name
    % beh: include behavior as a lagged variable (true) or not (false)
    
    % create a directory to save results in
    dirstring = strcat(beh_type, distinguish_str);
    mkdir(dirstring);
    numCores = feature('numcores');
    fprintf('* Creating parallel pool with %d workers...\n', numCores);
    parallelTime = tic;
    % Parallelization code is from the RPI IDEA repository.
    try
        if (~isempty(gcp('nocreate')))
            currentPool = gcp;
            if currentPool.NumWorkers < numCores
                delete(gcp('nocreate'));
                parpool(numCores);
            end
        else
            parpool(numCores);
        end
    catch E
        fprintf('* Unable to create parallel pool.');
        diary off;
        clearvars;
        rethrow(E);
    end

    fprintf('*Parallel pool successfully created after %.2f seconds.\n',...
        toc(parallelTime));

    %%%% GENERATE MODELS %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

    % Select variables to use for the model by prefix or by specific name
    pats =["Weather_TMAX", "Moon Phase", "Allergen_Total",...
        "Sleep", "BM", "Menses"];
    % set number of lagged days
    lag =  1;
    
    % Read in the data
    [Xc, yc, pc, ynan, p, text_vars] = read_data(beh_type, lag, beh,...
        dirstring, pats);
    
    % Standardize to the population
    Xc_uns = Xc;
    save('Xc_uns_new.mat','Xc_uns');
    [Xc, xmeans, xstds] = zscore(Xc);
    save('Xc_test_new.mat','Xc');
    save('pc_test_new.mat','pc');
    %error()
    % Log output
    dirstring_slash = strcat(dirstring,'/');
    log_output(dirstring_slash, strcat('xmeans', string(xmeans)))
    log_output(dirstring_slash, strcat('xstds', string(xstds)))
    
    % Random seeds to run over
    rngs = 1:30;
    % People in the cohort of 136
    people = 1:136;
    
    % Set model inputs
    linear_or_logistic = 'logistic';
    kernel_type = 'yes';
    
    % Put all in an outfile for logging
    log_output(dirstring_slash,strcat('random numbers ', string(rngs)))
    log_output(dirstring_slash,strcat('people ', string(people)))
    log_output(dirstring_slash,linear_or_logistic)
    log_output(dirstring_slash,kernel_type)
    
    % Make a directory to save individuals' data in
    mkdir(strcat(dirstring,'/individuals'));
    filepath = strcat(dirstring,'/individuals/');
    
    % Start timer
    tic
    
    % Run each person's models in parallel
    parfor person=people
        
        file_saving_string = strcat(dirstring, '/individuals/',...
            beh_type, distinguish_str, '_', linear_or_logistic,...
            '_', kernel_type, '_', string(person), '.mat');
        
        % Skip person if their models are already run
        if isfile(file_saving_string)
            disp('file found alIBready')
            continue
        else 
            Mod_RNGs = cell(max(size(rngs)),1)
            for RNG=rngs
                Mod = RunCode(RNG, linear_or_logistic, kernel_type,...
                    person, Xc, pc, yc)
                Mod_RNGs{RNG} = Mod;
                if Mod.Best == -1
                    break % Don't run additional models if the first one
                    % doesn't work
                end
            end
            Mods = Mod_RNGs;
            mySave(file_saving_string,Mods)
        end
    end
    
    toc
    
    % Get the current time as the save string file
    c = string(clock);
    c = strjoin(c(1:5),'_'); %convert to string

    % Plot according to the BCR parameter
    plotting_param = 'BCR';

    
    
    females = load('females.csv');


    vars = text_vars;
    
    % file extension for reading/saving models
    file_ext = strcat(distinguish_str, '_', linear_or_logistic,...
        '_', kernel_type, '_');
    
    % Plot behavior and model performance plots
    imbal_plots(dirstring, beh_type, plotting_param, filepath, ...
        file_ext, Xc, yc, ynan, p, pc, c);


    % Run the Pareto analysis
    Run_Pareto(dirstring, beh_type, filepath, file_ext, ...
    linear_or_logistic, kernel_type)

    % Plot the Pareto analysis results
    plot_important_feature(dirstring, females, beh_type, vars, c)

    
    % Run hierarchical clustering
    variable_hclust(dirstring, text_vars, c, Xc, pc, beh_type)
    
    % Display, then close all figures
    pause(10)
    close all;

 end