classdef plotStatistics < handle
% Main class to plot any one-dimensional statistic. Taking into account projects, experiments, labels, groups, etc
%
%   Copyright (C) 2016-2018, Javier G. Orlandi
  
  properties
    figureHandle;
    axisHandle;
    fullStatisticsData;
    fullStatisticsDataFull;
    statisticsName;
    labelList;
    mainGroup;
    mode;
    params; % Set at init
    guiHandle; % Set at init
    figName;
    figVisible;
    groupList;
    plotHandles;
    maxGroups;
    experiementNames;
    groupLabels;
    fullGroupList;
    exportFolder;
    figFolder;
    loadFields;
    plotDataFull;
    multiStatistic;
  end
  
  methods
    % Use this function to setup the statistics data
    %----------------------------------------------------------------------
    function success = getData(obj, optionsObject, projexp, varargin)
      success = false;
      if(isempty(obj.params.styleOptions.figureTitle))
        obj.figName = [obj.params.statistic ' - ' projexp.name obj.params.saveOptions.saveFigureTag];
      else
        obj.figName = obj.params.styleOptions.figureTitle;
      end
      obj.figName = strrep(obj.figName, '_', '-');
      switch obj.mode
        case 'project'
          success = getDataProject(obj, optionsObject, projexp, varargin{:});
        case 'experiment'
          success = getDataExperiment(obj, optionsObject, projexp, varargin{:});
      end
    end
    
    %----------------------------------------------------------------------
    function success = getDataExperiment(obj, optionsObject, experiment, varargin)
      success = true;
      obj.fullStatisticsData = {};
      
      % Get ALL subgroups in case of parents
      if(strcmpi(obj.mainGroup, 'all') || strcmpi(obj.mainGroup, 'ask'))
        obj.groupList = getExperimentGroupsNames(experiment);
      else
        obj.groupList = getExperimentGroupsNames(experiment, obj.mainGroup);
      end
      % If ask, open the popup
      if(strcmpi(obj.mainGroup, 'ask'))
        [selection, ok] = listdlg('PromptString', 'Select groups to use', 'ListString', obj.groupList, 'SelectionMode', 'multiple');
        if(~ok)
          success = false;
          return;
        end
        obj.groupList = obj.groupList(selection);
      end
      
      plotData = cell(1);
      plotData{1} = cell(length(obj.groupList), 1);
      if(isempty(obj.plotDataFull))
        obj.plotDataFull = cell(length(varargin{1}), 1);
        for it = 1:length(obj.plotDataFull)
          obj.plotDataFull{it} = cell(1);
          obj.plotDataFull{it}{1} = cell(length(obj.groupList), 1);
        end
      end
      
      
      for git = 1:length(obj.groupList)
        % Again, for compatibility reasons
        if(strcmpi(obj.groupList{git}, 'none'))
          obj.groupList{git} = 'everything';
        end
        % Here is where we obtain the data
        %obj.fullStatisticsData{git} = feval(funcHandle, experiment, obj.groupList{git}, varargin{:});

        try
          switch obj.multiStatistic
             case 'none'
               plotData{1}{git} = optionsObject.getData(experiment, obj.groupList{git}, varargin{:});
            case 'init'
              tmpFull = optionsObject.getData(experiment, obj.groupList{git}, varargin{:});
              for it = 1:length(obj.plotDataFull)
                obj.plotDataFull{it}{1}{git} = tmpFull{it};
              end
              plotData{1}{git} = obj.plotDataFull{1}{1}{git};
            case 'present'
              plotData{1}{git} = obj.plotDataFull{1}{git};
          end
        catch ME
          logMsg(strrep(getReport(ME),  sprintf('\n'), '<br/>'), 'w'); %#ok<SPRINTFN>
          logMsg(sprintf('Error getting data from experiment %s. Setting it to NaN', experiment.name), 'w');
          plotData{1}{git} = NaN;
        end
        % The important assignment
        obj.fullStatisticsData{git} = plotData{1}{git};

        % Eliminate NaNs
        if(obj.params.zeroToNan)
          obj.fullStatisticsData{git} = obj.fullStatisticsData{git}(~isnan(obj.fullStatisticsData{git}));
        end
        if(isempty(obj.fullStatisticsData))
          logMsg(sprintf('Found no data for group %s on experiment %s', obj.groupList{git}, experiment.name), obj.gui, 'w');
          continue;
        end
      end
    end
    
    %----------------------------------------------------------------------
    function success = getDataProject(obj, optionsObject, project, varargin)
      success = true;
      obj.fullStatisticsData = {};
      checkedExperiments = find(project.checkedExperiments);
      
      plotData = cell(length(checkedExperiments), 1);
      if(isempty(obj.plotDataFull))
        obj.plotDataFull = cell(length(varargin{1}), 1);
        for it = 1:length(obj.plotDataFull)
          obj.plotDataFull{it} = cell(length(checkedExperiments), 1);
          for it2 = 1:length(obj.plotDataFull{it})
            obj.plotDataFull{it}{it2} = cell(length(obj.fullGroupList), 1);
          end
        end
      end
      if(obj.params.pbar > 0)
        ncbar.setCurrentBarName(sprintf('Gathering data (%d/%d)', 0, length(checkedExperiments))');
        ncbar.update(0);
      end
      
      obj.maxGroups = 0;
      %%% First pass to gather all groups names
      for i = 1:length(checkedExperiments)
        experimentName = project.experiments{checkedExperiments(i)};
        experimentFile = [project.folderFiles experimentName '.exp'];
        experiment = load(experimentFile, '-mat', 'traceGroups', 'traceGroupsNames');
        % Get ALL subgroups in case of parents
        if(strcmpi(obj.mainGroup, 'all') || strcmpi(obj.mainGroup, 'ask'))
          obj.groupList = getExperimentGroupsNames(experiment);
        else
          obj.groupList = getExperimentGroupsNames(experiment, obj.mainGroup);
        end
        obj.fullGroupList = unique([obj.fullGroupList(:); obj.groupList(:)]);
      end
      % If ask, open the popup
      if(strcmpi(obj.mainGroup, 'ask'))
        [selection, ok] = listdlg('PromptString', 'Select groups to use', 'ListString', obj.fullGroupList, 'SelectionMode', 'multiple');
        if(~ok)
          success = false;
          return;
        end
        obj.fullGroupList = obj.fullGroupList(selection);
      end
      %%% Gather the data
      for i = 1:length(checkedExperiments)
        experimentName = project.experiments{checkedExperiments(i)};
        switch obj.multiStatistic
          case {'none', 'init'}
            experimentFile = [project.folderFiles experimentName '.exp'];
            if(~isempty(obj.params.loadFields))
              warning('off', 'MATLAB:load:variableNotFound');
              experiment = load(experimentFile, '-mat', 'traceGroups', 'traceGroupsNames', 'name', 'folder', 'ROI', obj.params.loadFields{:});
              warning('on', 'MATLAB:load:variableNotFound');
            else
              experiment = loadExperiment(experimentFile, 'verbose', false, 'project', project);
            end

            % Get ALL subgroups in case of parents
            if(strcmpi(obj.mainGroup, 'all')  || strcmpi(obj.mainGroup, 'ask'))
              obj.groupList = getExperimentGroupsNames(experiment);
            else
              obj.groupList = getExperimentGroupsNames(experiment, obj.mainGroup);
            end
        end        
        plotData{i} = cell(length(obj.fullGroupList), 1);
        for git = 1:length(obj.fullGroupList)
          groupIdx = find(strcmp(obj.fullGroupList{git}, obj.groupList));
          if(isempty(groupIdx))
            continue;
          end
          % Again, for compatibility reasons
          if(strcmpi(obj.groupList{groupIdx}, 'none'))
            obj.groupList{groupIdx} = 'everything';
          end
          %plotData{i}{git} = getData(experiment, obj.groupList{git}, obj.params.statistic);
          try
            switch obj.multiStatistic
               case 'none'
                 plotData{i}{git} = optionsObject.getData(experiment, obj.groupList{groupIdx}, varargin{:});
              case 'init'
                tmpFull = optionsObject.getData(experiment, obj.groupList{groupIdx}, varargin{:});
                for it = 1:length(obj.plotDataFull)
                  obj.plotDataFull{it}{i}{git} = tmpFull{it};
                end
                plotData{i}{git} = obj.plotDataFull{1}{i}{git};
              case 'present'
                plotData{i}{git} = obj.plotDataFull{i}{git};
            end
          catch ME
            logMsg(strrep(getReport(ME),  sprintf('\n'), '<br/>'), 'w'); %#ok<SPRINTFN>
            logMsg(sprintf('Error getting data from experiment %s. Setting it to NaN', experimentName), 'w');
            plotData{i}{git} = NaN;
          end
          if(obj.params.zeroToNan)
            plotData{i}{git}(plotData{i}{git} == 0) = NaN;
          end
        end
        obj.maxGroups = max(obj.maxGroups, length(plotData{i}));
        if(obj.params.pbar > 0)
          ncbar.setCurrentBarName(sprintf('Gathering data (%d/%d)', i, length(checkedExperiments))');
          ncbar.update(i/length(checkedExperiments));
        end
      end

      %%% Get the labels we need
      labelsToUse = obj.params.pipelineProject.labelGroups;
      try
        [labelList, uniqueLabels, labelsCombinations, labelsCombinationsNames, experimentsPerCombinedLabel] = getLabelList(project, find(project.checkedExperiments), labelsToUse);
        %[labelList, uniqueLabels, labelsCombinations, labelsCombinationsNames, experimentsPerCombinedLabel] = getLabelList(project, find(project.checkedExperiments));
      catch
      end

      labelsToUse = obj.params.pipelineProject.labelGroups;
      labelsToUseJoined = cell(length(labelsToUse), 1);
      try
        if(isempty([obj.params.pipelineProject.labelGroups{:}]))
          labelsToUseJoined = uniqueLabels;
        else
          for it = 1:length(obj.params.pipelineProject.labelGroups)
            labelsToUseJoined{it} = strjoin(sort(strtrim(strsplit(obj.params.pipelineProject.labelGroups{it}, ','))), ', ');
          end
        end
      catch
        labelsToUseJoined = uniqueLabels;
      end
      validCombinations = cellfun(@(x)find(strcmp(labelsCombinationsNames, x)), labelsToUseJoined, 'UniformOutput', false);
      valid = find(cellfun(@(x)~isempty(x), validCombinations));
      if(length(valid) ~= length(validCombinations))
        logMsg('Some label sets had no representative experiments', 'w');
      end
      validCombinations = cell2mat(validCombinations(valid));
      labelsToUseJoined = labelsToUseJoined(valid);
      
      %%% Average the data as needed
      expPos = 0;
      switch obj.params.pipelineProject.groupingOrder
        case 'none'
          % Do nothing
          plotDataAveragedFull = cell(length(validCombinations), 1);
          obj.groupLabels = project.experiments(checkedExperiments);
          obj.fullStatisticsDataFull = plotDataAveragedFull;
        case 'label'
          plotDataAveraged = cell(length(validCombinations), 1);
          plotDataAveragedFull = cell(length(validCombinations), 1);
          maxExpPerLabel = -inf;
          maxDataPoints = -inf;
          for it = 1:length(validCombinations)
            plotDataAveraged{it} = cell(obj.maxGroups, 1);
            maxExpPerLabel = max(maxExpPerLabel,length(experimentsPerCombinedLabel{validCombinations(it)}));
            valid = [experimentsPerCombinedLabel{validCombinations(it)}{:}];
            for git = 1:obj.maxGroups
              plotDataAveraged{it}{git} = [];
              for k = 1:length(valid)
                if(length(valid) >= k && length(plotData) >= valid(k) && length(plotData{valid(k)}) >= git)
                  maxDataPoints = max(maxDataPoints, length(plotData{valid(k)}{git}));
                end
              end
            end
          end
          plotDataAveraged = nan(maxDataPoints, length(validCombinations), maxExpPerLabel);

          for it = 1:length(validCombinations)
            valid = [experimentsPerCombinedLabel{validCombinations(it)}{:}];
            expPos = [expPos; length(valid)];
            for git = 1:obj.maxGroups
              for k = 1:length(valid)
                if(length(plotData) >= valid(k) && length(plotData{valid(k)}) >= git)
                  plotDataAveraged(1:length(plotData{valid(k)}{git}), it, k) = plotData{valid(k)}{git};
                end
              end
            end
          end
          plotData = plotDataAveraged;
          
          expPos = cumsum(expPos);
          expPos = expPos(1:end-1);
          obj.groupLabels = labelsToUseJoined;
          obj.fullStatisticsDataFull = plotDataAveragedFull;
        case 'label average'
          % Average statistics for each experiment, and group them by label
          plotDataAveraged = cell(length(validCombinations), 1);
          plotDataAveragedFull = cell(length(validCombinations), 1);
          switch obj.params.pipelineProject.factor
            case 'experiment'
              % Label change
              switch obj.params.pipelineProject.factorAverageFunction
                case 'mean'
                case 'median'
                  obj.figName = [obj.figName ' median factor'];
                case 'std'
                  obj.figName = [obj.figName ' std factor'];
                case 'var'
                  obj.figName = [obj.figName ' var factor'];
                case 'skewness'
                  obj.figName = [obj.figName ' ske factor'];
                case 'cv'
                  obj.figName = [obj.figName ' cv factor'];
              end
              for it = 1:length(validCombinations)
                plotDataAveraged{it} = cell(obj.maxGroups, 1);
                valid = [experimentsPerCombinedLabel{validCombinations(it)}{:}];
                valid2 = arrayfun(@(x)find(x==checkedExperiments), valid, 'UniformOutput', false);
                valid2 = cell2mat(valid2);
                valid = valid2;
                for git = 1:obj.maxGroups
                  plotDataAveraged{it}{git} = zeros(length(valid), 1);
                  for k = 1:length(valid)
                    try
                      switch obj.params.pipelineProject.factorAverageFunction
                        case 'mean'
                          plotDataAveraged{it}{git}(k) = nanmean(plotData{valid(k)}{git});
                        case 'median'
                          plotDataAveraged{it}{git}(k) = nanmedian(plotData{valid(k)}{git});
                        case 'std'
                          plotDataAveraged{it}{git}(k) = nanstd(plotData{valid(k)}{git});
                        case 'var'
                          plotDataAveraged{it}{git}(k) = nanvar(plotData{valid(k)}{git});
                        case 'skewness'
                          plotDataAveraged{it}{git}(k) = skewness(plotData{valid(k)}{git});
                        case 'cv'
                          plotDataAveraged{it}{git}(k) = nanstd(plotData{valid(k)}{git})./nanmean(plotData{valid(k)}{git});
                      end
                    catch
                    end
                  end
                end
              end
            case 'event'
              for it = 1:length(validCombinations)
                plotDataAveraged{it} = cell(obj.maxGroups, 1);
                valid = [experimentsPerCombinedLabel{validCombinations(it)}{:}];
                valid2 = arrayfun(@(x)find(x==checkedExperiments), valid, 'UniformOutput', false);
                valid2 = cell2mat(valid2);
                valid = valid2;
                for git = 1:obj.maxGroups
                  plotDataAveraged{it}{git} = [];
                  for k = 1:length(valid)
                    if(length(plotData) >= valid(k) && length(plotData{valid(k)}) >= git)
                      plotDataAveraged{it}{git} = [plotDataAveraged{it}{git}; plotData{valid(k)}{git}(:)];
                    end
                  end
                end
              end
            case 'mixed'
              for it = 1:length(validCombinations)
                plotDataAveraged{it} = cell(obj.maxGroups, 1);
                plotDataAveragedFull{it} = cell(obj.maxGroups, 1);
                valid = [experimentsPerCombinedLabel{validCombinations(it)}{:}];
                valid2 = arrayfun(@(x)find(x==checkedExperiments), valid, 'UniformOutput', false);
                valid2 = cell2mat(valid2);
                valid = valid2;
                for git = 1:obj.maxGroups
                  plotDataAveraged{it}{git} = [];
                  plotDataAveragedFull{it}{git} = cell(length(valid), 1);
                  for k = 1:length(valid)
                    if(length(plotData) >= valid(k) && length(plotData{valid(k)}) >= git)
                      plotDataAveraged{it}{git} = [plotDataAveraged{it}{git}; plotData{valid(k)}{git}(:)];
                      plotDataAveragedFull{it}{git}{k} = plotData{valid(k)}{git}(:);
                    end
                  end
                end
              end
          end
          plotData = plotDataAveraged;
          obj.fullStatisticsDataFull = plotDataAveragedFull;
          obj.groupLabels = labelsToUseJoined;
      end

      obj.fullStatisticsData = plotData;
    end
    
    %----------------------------------------------------------------------
    function createFigure(obj)
      switch obj.mode
        case 'experiment'
          obj.createFigureExperiment();
        case 'project'
          obj.createFigureProject();
      end
    end
    
    %----------------------------------------------------------------------
    function createFigureExperiment(obj)
      obj.plotHandles = [];
      
      if(obj.params.saveOptions.onlySaveFigure)
        obj.figVisible = 'off';
      else
        obj.figVisible = 'on';
      end
      obj.figureHandle = figure('Name', obj.figName, 'NumberTitle', 'off', 'Visible', obj.figVisible, 'Tag', 'netcalPlot');
      %obj.figureHandle.Position = setFigurePosition(obj.guiHandle, 'width', obj.params.styleOptions.figureSize(1), 'height', obj.params.styleOptions.figureSize(2));
      obj.figureHandle.Position = setFigurePosition(obj.guiHandle, 'width', obj.params.styleOptions.figureSize(1), 'height', obj.params.styleOptions.figureSize(2));
      obj.axisHandle = axes;
      hold on;
    
      if(isempty(obj.params.styleOptions.xLabel))
        xlabel(obj.params.statistic);
      else
        xlabel(obj.params.styleOptions.xLabel);
      end
      if(isempty(obj.params.styleOptions.yLabel))
        ylabel('PDF');
      else
        ylabel(obj.params.styleOptions.yLabel);
      end
      
      if(isfield(obj.params.styleOptions, 'colormap') && ~isempty(obj.params.styleOptions.colormap))
        cmap = eval(sprintf('%s (%d)', obj.params.styleOptions.colormap, length(obj.groupList)));
      else
        cmap = lines(length(obj.groupList));
      end
      if(obj.params.styleOptions.invertColormap)
        if(size(cmap, 1) == 1)
          % If the colormap has a single entry we need to add another entry
          if(isfield(obj.params.styleOptions, 'colormap') && ~isempty(obj.params.styleOptions.colormap))
            cmap = eval(sprintf('%s (%d)', obj.params.styleOptions.colormap, length(obj.groupList)+1));
          else
            cmap = lines(length(obj.groupList)+1);
          end
          cmap = cmap(end:-1:1, :);
        else
          % Default behavior
          cmap = cmap(end:-1:1, :);
        end
      end
      if(length(obj.groupList) > 1)
        alpha = 0.5;
      else
        alpha = 1;
      end
      validGroups = [];
      for git = 1:length(obj.groupList)
        curData = obj.fullStatisticsData{git};
        if(isempty(curData))
          logMsg(sprintf('No data found for group: %s', obj.groupList{git}), obj.guiHandle, 'w');
          continue;
        else
          validGroups = [validGroups, git];
        end
        % Compute the distribution
        [f, xi] = obj.computeDistribution(curData);
        % The actual plot
        switch obj.params.distributionEstimation.mode
          case {'unbounded', 'positive', 'histogram line'}
             h = plot(xi, f, 'Color', cmap(git, :));
          case 'histogram'
            h = bar(xi, f, 'FaceColor', cmap(git, :), 'EdgeColor', cmap(git, :)*0.5, 'FaceAlpha', alpha, 'EdgeAlpha', alpha);
          case 'raw'
            h = bar(xi, f, 'FaceColor', cmap(git, :), 'EdgeColor', cmap(git, :)*0.5, 'FaceAlpha', alpha, 'EdgeAlpha', alpha);
            if(isempty(obj.params.styleOptions.yLabel))
              ylabel('count');
            else
              ylabel(obj.params.styleOptions.yLabel);
            end
          case 'raw line'
            h = plot(xi, f, 'Color', cmap(git, :));
            if(isempty(obj.params.styleOptions.yLabel))
              ylabel('count');
            else
              ylabel(obj.params.styleOptions.yLabel);
            end
        end
        obj.plotHandles = [obj.plotHandles; h];
      end
      l = legend(obj.groupList(validGroups));
      l.Location = 'best';
      l.Box = 'off';
      title(obj.axisHandle, obj.figName, 'interpreter', 'none');
      set(obj.axisHandle, 'XTickLabelRotation', obj.params.styleOptions.XTickLabelRotation);
      set(obj.axisHandle, 'YTickLabelRotation', obj.params.styleOptions.YTickLabelRotation);
      box on;
      set(obj.axisHandle,'Color','w');
      set(obj.figureHandle,'Color','w');
      ui = uimenu(obj.figureHandle, 'Label', 'Export');
      uimenu(ui, 'Label', 'Figure',  'Callback', {@exportFigCallback, {'*.pdf';'*.eps'; '*.tiff'; '*.png'}, strrep([obj.figFolder, obj.figName], ' - ', '_'), obj.params.saveOptions.saveFigureResolution});
      
      if(obj.params.saveOptions.saveFigure)
        export_fig([obj.figFolder, obj.figName, '.', obj.params.saveOptions.saveFigureType], ...
                    sprintf('-r%d', obj.params.saveOptions.saveFigureResolution), ...
                    sprintf('-q%d', obj.params.saveOptions.saveFigureQuality), obj.figureHandle);
      end
    end
    
    %----------------------------------------------------------------------
    function createFigureProject(obj)
      curData = obj.fullStatisticsData;
      if(obj.params.saveOptions.onlySaveFigure)
        obj.figVisible = 'off';
      else
        obj.figVisible = 'on';
      end

      gap = 1;
      rowData = [];
      groupData = [];

      % Let's start 
      for rt = 1:obj.maxGroups
        for kr = 1:size(curData, 1)
          if(isempty(curData{kr}) || length(curData{kr}) < rt || isempty(curData{kr}{rt}))
              rowData = [rowData; NaN];
              groupData = [groupData; kr];
          else
              rowData = [rowData; curData{kr}{rt}(:)];
              groupData = [groupData; ones(size(curData{kr}{rt}(:)))*((kr-1)*(obj.maxGroups+gap) + rt)];
          end
        end
      end
      % Turn into an array now
      maxData = length(rowData);

      bpData = nan(maxData, obj.maxGroups, size(curData, 1));
      for rt = 1:obj.maxGroups
        for kr = 1:size(curData, 1)
          if(length(curData{kr}) >= rt)
            bpData(1:length(curData{kr}{rt}), rt, kr) = curData{kr}{rt};
          end
        end
      end

      switch obj.params.pipelineProject.barGroupingOrder
        case 'default'
          subData = permute(bpData,[1 3 2]);
          xList = obj.groupLabels;
          legendList = obj.fullGroupList;
          if(~iscell(legendList) || (iscell(legendList) && length(legendList) == 1))
            legendList = {legendList};
          end
        case 'group'
          subData = permute(bpData,[1 2 3]);
          legendList = obj.groupLabels;
          xList= obj.fullGroupList;
      end
      switch obj.params.pipelineProject.outlierRemoval
        case 'none'
        otherwise
          for git = 1:size(subData,3)
            for it = 1:size(subData, 2)
              outlierHit = find(isoutlier(subData(:, it, git), obj.params.pipelineProject.outlierRemoval));
              if(~isempty(outlierHit))
                logMsg(sprintf('Found %d outliers on %s %s (setting to NaN)', length(outlierHit), xList{it}, obj.fullGroupList{git}));
                subData(outlierHit, it, git) = NaN;
              end
            end
          end
      end
      if(isfield(obj.params.styleOptions, 'colormap') && ~isempty(obj.params.styleOptions.colormap))
        cmap = eval(sprintf('%s (%d)', obj.params.styleOptions.colormap, length(legendList)));
      else
        cmap = lines(length(legendList));
      end
      if(obj.params.styleOptions.invertColormap)
        if(size(cmap, 1) == 1)
          % If the colormap has a single entry we need to add another entry
          if(isfield(obj.params.styleOptions, 'colormap') && ~isempty(obj.params.styleOptions.colormap))
            cmap = eval(sprintf('%s (%d)', obj.params.styleOptions.colormap, 2));
          else
            cmap = lines(2);
          end
          cmap = cmap(2, :);
        else
          % Default behavior
          cmap = cmap(end:-1:1, :);
        end
      end
      switch obj.params.pipelineProject.showSignificance
        case 'none'
          obj.fullGroupList = {obj.fullGroupList};
        otherwise
          obj.fullGroupList = {obj.fullGroupList};
          grList = cell(length(obj.fullGroupList{1}), 1);
          pList = cell(length(obj.fullGroupList{1}), 1);
          intraPlist = cell(length(obj.fullGroupList{1}), 1);
          intraGrList = cell(length(obj.fullGroupList{1}), 1);
          nTests = cell(length(obj.fullGroupList{1}), 1);
          switch obj.params.pipelineProject.factor
            case 'mixed'
              for git = 1:length(obj.fullStatisticsDataFull{1})
                grList{git} = {};
                intraGrList{git} = {};
                nTests{git} = 0;
              end
            otherwise
              for git = 1:size(subData, 3)
                grList{git} = {};
                intraGrList{git} = {};
                nTests{git} = 0;
              end
          end
      end
      % Normality testing
      %logMsg(sprintf('%s vs %s . Group: %s . Mann-Whitney U test %s - Kolmogorov-Smirnov test %s - Ttest2 %s', xList{it}, xList{itt}, obj.fullGroupList{1}{git}, MWtext, KStext, TTtext));
      warning('off', 'stats:lillietest:OutOfRangePHigh');
      warning('off', 'stats:lillietest:OutOfRangePLow');
      for it = 1:size(subData, 2)
        for git = 1:size(subData, 3)
          try
            [h, p] = lillietest(subData(:, it, git));
            Ltext = sprintf('P=%.3g', p);
            if(p <= 0.05)
              Ltext = ['<b>' Ltext '</b>'];
            end
            logMsg(sprintf('%s. Group: %s . Lilliefors test %s ', xList{it}, obj.fullGroupList{1}{git}, Ltext));
          catch
          end
        end
      end
      warning('on', 'stats:lillietest:OutOfRangePHigh');
      warning('on', 'stats:lillietest:OutOfRangePLow');
      switch obj.params.pipelineProject.factor
        case 'mixed'
          fullData = obj.fullStatisticsDataFull;
          switch obj.params.pipelineProject.showSignificance
            case 'none'
            case 'partial'
              for it = 1:length(fullData)
                for itt = (it+1):length(fullData)
                  % Avoid cross-comparisons
                  if(obj.params.pipelineProject.avoidCrossComparisons)
                    if(~any(cellfun(@(x)any(strcmp(x, strtrim(strsplit(xList{itt}, ',')))), strtrim(strsplit(xList{it}, ',')))))
                      continue;
                    end
                  end
                  % Assuming the groups are always there
                  for git = 1:length(fullData{it})
                    nTests{git} = nTests{git} + 1;
                    pValueList = [];
                    for k = 1:length(fullData{it}{git})
                      for kk = 1:length(fullData{itt}{git})
                        if(all(isnan(fullData{it}{git}{k})) || all(isnan(fullData{itt}{git}{kk})))
                          pValueList = [pValueList; 1];
                          continue;
                        end
                        try
                          p = ranksum(fullData{it}{git}{k}, fullData{itt}{git}{kk});
                          [~, p2] = kstest2(fullData{it}{git}{k}, fullData{itt}{git}{kk});
                          [~, p3] = ttest2(fullData{it}{git}{k}, fullData{itt}{git}{kk});
                          %logMsg(sprintf('%s vs %s . Group: %s . Idx: %d vs %d Mann-Whitney U test P= %.3g - Kolmogorov-Smirnov test P= %.3g', xList{it}, xList{itt}, obj.fullGroupList{git}, k, kk, p, p2));
                          switch obj.params.pipelineProject.significanceTest
                            case 'Mann-Whitney'
                              pValueList= [pValueList; p];
                            case 'Kolmogorov-Smirnov'
                              pValueList= [pValueList; p2];
                            case 'Ttest2'
                              pValueList= [pValueList; p3];
                          end
                        catch ME
                          logMsg(strrep(getReport(ME),  sprintf('\n'), '<br/>'), 'w');
                        end
                      end
                    end
                    try
                      %n = -2*sum(log(pValueList));
                      %fval = chi2cdf(n,2*length(pValueList));
                      fval = mean(pValueList);
                      logMsg(sprintf('%s vs %s . Group: %s . Avg p-value: %.3g', xList{it}, xList{itt}, obj.fullGroupList{1}{git}, fval));
                      if(fval <= 0.05)
                        pList{git} = [pList{git}; fval];
                        grList{git}{end+1} = [it itt];
                      end
                    catch ME
                      logMsg(strrep(getReport(ME),  sprintf('\n'), '<br/>'), 'w');
                    end
                  end
                end
                if(obj.params.pipelineProject.computeIntraGroupComparisons)
                  for git = 1:length(fullData{it})
                    intraP = [];
                    for k = 1:length(fullData{it}{git})
                      for kk = (k+1):length(fullData{it}{git})
                        % Check that everything isn't nan
                        if(all(isnan(fullData{it}{git}{k})) || all(isnan(fullData{it}{git}{kk})))
                          intraP= [intraP; 1];
                          continue;
                        end
                        try
                          p = ranksum(fullData{it}{git}{k}, fullData{it}{git}{kk});
                          [~, p2] = kstest2(fullData{it}{git}{k}, fullData{it}{git}{kk});
                          %logMsg(sprintf('Intragroup: %s . Group: %s . Idx: %d vs %d Mann-Whitney U test P= %.3g - Kolmogorov-Smirnov test P= %.3g', xList{it}, objfullGroupList{git}, k, kk, p, p2));
                          switch obj.params.pipelineProject.significanceTest
                            case 'Mann-Whitney'
                              intraP= [intraP; p];
                            case 'Kolmogorov-Smirnov'
                              intraP= [intraP; p2];
                            case 'Ttest2'
                              intraP= [intraP; p3];
                          end
                        catch ME
                          logMsg(strrep(getReport(ME),  sprintf('\n'), '<br/>'), 'w');
                        end
                      end
                    end
                    if(mean(intraP) <= 0.05)
                      intraPlist{git} = [intraPlist{git}; mean(intraP)];
                      intraGrList{git}{end+1} = [it it];
                    end
                    logMsg(sprintf('Intragroup: %s . Group: %s . Avg p-value: %.3g', xList{it}, obj.fullGroupList{1}{git}, mean(intraP)));
                  end
                end
              end
          end
        otherwise
          switch obj.params.pipelineProject.showSignificance
            case 'none'
            case 'partial'
              for it = 1:size(subData, 2)
                for itt = (it+1):size(subData, 2)
                  for git = 1:size(subData, 3)
                    if(obj.params.pipelineProject.avoidCrossComparisons)
                      if(~any(cellfun(@(x)any(strcmp(x, strtrim(strsplit(xList{itt}, ',')))), strtrim(strsplit(xList{it}, ',')))))
                        continue;
                      else
                        nTests{git} = nTests{git} + 1;
                      end
                    else
                      nTests{git} = nTests{git} + 1;
                    end
                    try
                      p = ranksum(subData(:, it, git), subData(:, itt, git));
                      [h, p2] = kstest2(subData(:, it, git), subData(:, itt, git));
                      [h, p3] = ttest2(subData(:, it, git), subData(:, itt, git));
                      %subData(:, it, git)
                      %subData(:, itt, git)
                      %<b> </b>
                      MWtext = sprintf('P=%.3g', p);
                      KStext = sprintf('P=%.3g', p2);
                      TTtext = sprintf('P=%.3g', p3);
                      if(p <= 0.05)
                        MWtext = ['<b>' MWtext '</b>'];
                      end
                      if(p2 <= 0.05)
                        KStext = ['<b>' KStext '</b>'];
                      end
                      if(p3 <= 0.05)
                        TTtext = ['<b>' TTtext '</b>'];
                      end
                      logMsg(sprintf('%s vs %s . Group: %s . Mann-Whitney U test %s - Kolmogorov-Smirnov test %s - Ttest2 %s', xList{it}, xList{itt}, obj.fullGroupList{1}{git}, MWtext, KStext, TTtext));
                      switch obj.params.pipelineProject.significanceTest
                        case 'Mann-Whitney'
                          if(p <= 0.05)
                            pList{git} = [pList{git}; p];
                            grList{git}{end+1} = [it itt];
                          end
                        case 'Kolmogorov-Smirnov'
                          if(p2 <= 0.05)
                            pList{git} = [pList{git}; p2];
                            grList{git}{end+1} = [it itt];
                          end
                        case 'Ttest2'
                          if(p3 <= 0.05)
                            pList{git} = [pList{git}; p3];
                            grList{git}{end+1} = [it itt];
                          end
                      end
                    catch ME
                      logMsg(strrep(getReport(ME),  sprintf('\n'), '<br/>'), 'w');
                    end
                  end
                end
              end
            case 'all'
              for it = 1:size(subData, 2)
                for itt = 1:size(subData, 2)
                  for git = 1:size(subData, 3)
                    if(it > itt)
                      p = ranksum(subData(:, it, git), subData(:, itt, git));
                      [h, p2] = kstest2(subData(:, it, git), subData(:, itt, git));
                      [h, p3] = ttest2(subData(:, it, git), subData(:, itt, git));
                      MWtext = sprintf('P=%.3g', p);
                      KStext = sprintf('P=%.3g', p2);
                      TTtext = sprintf('P=%.3g', p3);
                      if(p <= 0.05)
                        MWtext = ['<b>' MWtext '</b>'];
                      end
                      if(p2 <= 0.05)
                        KStext = ['<b>' KStext '</b>'];
                      end
                      if(p3 <= 0.05)
                        TTtext = ['<b>' TTtext '</b>'];
                      end
                      logMsg(sprintf('%s vs %s . Group: %s . Mann-Whitney U test %s - Kolmogorov-Smirnov test %s - Ttest2 %s', xList{it}, xList{itt}, obj.fullGroupList{1}{git}, MWtext, KStext, TTtext));
                      switch obj.params.pipelineProject.significanceTest
                        case 'Mann-Whitney'
                          pList{git} = [pList{git}; p];
                          grList{git}{end+1} = [it itt];
                        case 'Kolmogorov-Smirnov'
                          pList{git} = [pList{git}; p2];
                          grList{git}{end+1} = [it itt];
                        case 'Ttest2'
                          pList{git} = [pList{git}; p3];
                          grList{git}{end+1} = [it itt];
                      end
                      %grList{git}{end+1} = [it itt]; WHY WAS THIS HERE??
                    end
                  end
                end
              end
          end
      end
      % Holm-Bonferroni
      if(obj.params.pipelineProject.HolmBonferroniCorrection)
        switch obj.params.pipelineProject.showSignificance
          case {'partial', 'all'}
            for git = 1:length(nTests)
              Ncomparisons = nTests{git};
              if(isempty(Ncomparisons) || Ncomparisons == 0 || isempty(pList{git}))
                logMsg(sprintf('No significant data found for Holm-Bonferroni correction on group: %s', obj.fullGroupList{1}{git}));
                continue;
              end
              fullList = [pList{git}, cellfun(@(x)x(1), grList{git})', cellfun(@(x)x(2), grList{git})'];
              [fullList, idx] = sortrows(fullList, 1);
              validComparisons = 0;
              for it = 1:size(fullList, 1)
                if(fullList(it, 1) >= 0.05/(Ncomparisons-it+1))
                  break;
                else
                  validComparisons = validComparisons + 1;
                end
              end
              pList{git} = pList{git}(idx(1:validComparisons));
              grList{git} = grList{git}(idx(1:validComparisons));
              logMsg('Valid comparisons after Holm-Bonferroni correction:');
              for it = 1:length(pList{git})
                logMsg(sprintf('%s vs %s . Group: %s . P= %.3g', xList{grList{git}{it}(1)}, xList{grList{git}{it}(2)}, obj.fullGroupList{1}{git}, pList{git}(it)));
              end
            end
        end
      end
      
      % The actual figure creation
      obj.figureHandle = figure('Name', obj.figName, 'NumberTitle', 'off', 'Visible', obj.figVisible, 'Tag', 'netcalPlot');
      obj.figureHandle.Position = setFigurePosition(obj.guiHandle, 'width', obj.params.styleOptions.figureSize(1), 'height', obj.params.styleOptions.figureSize(2));
      obj.axisHandle = axes;
      hold on;
      box on;
      import iosr.statistics.*
      setappdata(gcf, 'subData', subData);

      switch obj.params.pipelineProject.plotType
        case 'univarscatter'
          % Could use reshape, but better be safe
          newData = zeros(size(subData, 1), size(subData, 2)*size(subData, 3));
          fullCmap = zeros(size(newData, 2), 3);
          newXPos = zeros(1, size(newData, 2));
          newXTickPos = zeros(1, size(subData, 2));
          newXPosSubdata = zeros(size(subData,2), size(subData, 3));
          curIt = 0;
          curPos = 0;
          for it1 = 1:size(subData, 2)
            newXTickPos(it1) = curPos + 1 + (size(subData, 3)-1)/2;
            for it2 = 1:size(subData, 3)
              curPos = curPos + 1;
              curIt = curIt + 1;
              newData(:, curIt) = subData(:, it1, it2);
              fullCmap(curIt, :) = cmap(it2, :);
              newXPos(curIt) = curPos;
              newXPosSubdata(it1, it2) = curPos;
            end
            % Gap of 1 between groups
            if(size(subData, 3) > 1)
              curPos = curPos+1;
            end
          end
          [xPositions, yPositions, Label, RangeCut, hList] = ...
            UnivarScatterV2(newData, 'XPositions', newXPos, ...
            'MarkerFaceColor', fullCmap, 'MarkerEdgeColor', fullCmap*0.5 ...
            );
            %,'RangeCut', [10 10]);
          %%% Quick hack - 2026 for paired measures
          %hold on;
          %plot(xPositions',yPositions','-', 'Color', [0 0.447 0.741])
          %plot(nanmean(newData,1), 'k', 'LineWidth', 2);
          %grList = cell(length(obj.fullGroupList{1}), 1);
          %intraGrList = cell(length(obj.fullGroupList{1}), 1);
          %%% The significance
          try
            if(~strcmpi(obj.params.pipelineProject.showSignificance, 'none'))
              switch obj.params.pipelineProject.significanceTest
                case {'Mann-Whitney', 'Kolmogorov-Smirnov', 'Ttest2'}
                  % The intragroup comparisons
                  if(obj.params.pipelineProject.computeIntraGroupComparisons && strcmpi(obj.params.pipelineProject.factor, 'mixed'))
                    for git = 1:size(subData, 3)
                      if(isempty(intraPlist{git}))
                        continue;
                      end
                      for kit = 1:length(pList{git})
                        tmpPos = intraGrList{git}{kit};
                        newPos = newXPosSubdata(tmpPos, git)+[-0.05 0.05];
                        sigstar(newPos, intraPlist{git}(kit));
                      end
                    end
                  end
                  % The normal comparisons
                  for git = 1:size(subData, 3)
                    if(isempty(pList{git}))
                      continue;
                    end
                    %newPos = newXPos([git, git+size(subData,3)]);
                    for kit = 1:length(pList{git})
                      tmpPos = grList{git}{kit};
                      newPos = newXPosSubdata(tmpPos, git);
                      sigstar(newPos, pList{git}(kit));
                    end
                  end
              end
            end
          catch ME
            logMsg(strrep(getReport(ME),  sprintf('\n'), '<br/>'), 'e');
          end
          
          %%% Figure properties
          if(length(legendList) == 1)
            l = legend(hList(1:size(subData, 3)), legendList{1});
          else
            l = legend(hList(1:size(subData, 3)), legendList);
          end
          l.Location = 'best';
          l.Box = 'off';
          set(obj.axisHandle, 'XTick', newXTickPos);
          set(obj.axisHandle, 'XTickLabel', xList);
          set(hList, 'SizeData', 16);
          title(obj.axisHandle, obj.figName, 'interpreter', 'none');

          if(isempty(obj.params.styleOptions.xLabel))
          else
            xlabel(obj.params.styleOptions.xLabel);
          end
          if(isempty(obj.params.styleOptions.yLabel))
            ylabel(obj.statisticsName);
          else
            ylabel(obj.params.styleOptions.yLabel);
          end

          set(obj.figureHandle,'Color','w');

          set(obj.axisHandle, 'XTickLabelRotation', obj.params.styleOptions.XTickLabelRotation);
          set(obj.axisHandle, 'YTickLabelRotation', obj.params.styleOptions.YTickLabelRotation);

          title(obj.axisHandle, obj.figName, 'interpreter', 'none');
          set(obj.axisHandle,'Color','w');
          set(obj.figureHandle,'Color','w');
            
          %%% The menu
          ui = uimenu(obj.figureHandle, 'Label', 'Export');
          uimenu(ui, 'Label', 'Figure',  'Callback', {@exportFigCallback, {'*.pdf';'*.eps'; '*.tiff'; '*.png'}, strrep([obj.figFolder, obj.figName], ' - ', '_'), obj.params.saveOptions.saveFigureResolution});
          uimenu(ui, 'Label', 'Data (statistics)',  'Callback', @(h,e)obj.exportDataAggregates(bpData, obj.exportFolder));
          uimenu(ui, 'Label', 'Data (full)',  'Callback', @(h,e)obj.exportDataFull(bpData, obj.exportFolder));
          
        case 'boxplot'
          try
            obj.plotHandles = boxPlot(xList, subData, ...
                            'symbolColor','k',...
                            'medianColor','k',...
                            'symbolMarker','+',...
                            'showLegend',false, ...
                            'showOutliers', true, ...
                            'groupLabels', legendList, ...
                            'showLegend', true, ...
                            'notch', obj.params.styleOptions.notch, ...
                            'boxcolor', cmap);
          catch
            obj.plotHandles = [];
          end
          if(obj.params.pipelineProject.showMeanError)
            try
              obj.plotHandles.showMean = true;
              obj.plotHandles.meanColor = [1 0 0];
              boxes = obj.plotHandles.handles.box;
              boxesPositions = arrayfun(@(x)mean(x.Vertices(:,1)), boxes(:));
              xcoords = boxesPositions;
              avgy = nanmean(obj.plotHandles.y,1);
              erry = nanstd(obj.plotHandles.y,1)./sqrt(sum(~isnan(obj.plotHandles.y)));
              hold on;
              h = errorbar(xcoords, avgy, erry,'o');
              h.Color = [1 0 0];
            catch ME
              logMsg(strrep(getReport(ME),  sprintf('\n'), '<br/>'), 'e');
            end
          end
          try
            % Now let's fix the patches
            boxes = obj.plotHandles.handles.box;
            if(all(arrayfun(@(x)length(unique(x.Vertices(:, 2))) == 1, boxes,  'UniformOutput', true) | arrayfun(@(x)all(isnan(x.Vertices(:, 2))), boxes,  'UniformOutput', true)))
              singleStatistic = true;
            else
              singleStatistic = false;
            end

            % Turn the patches into simple bars if only 1 data point
            if(singleStatistic)
              for it = 1:numel(boxes)
                boxes(it).Vertices(1,2) = 0;
                boxes(it).Vertices(end,2) = 0;
              end
            end
            if(~strcmpi(obj.params.pipelineProject.showSignificance, 'none'))
              switch obj.params.pipelineProject.significanceTest
                case {'Mann-Whitney', 'Kolmogorov-Smirnov', 'Ttest2'}
                  % The intragroup comparisons
                  if(obj.params.pipelineProject.computeIntraGroupComparisons && strcmpi(obj.params.pipelineProject.factor, 'mixed'))
                    for git = 1:size(subData, 3)
                      if(isempty(intraPlist{git}))
                        continue;
                      end
                      boxesPositions = arrayfun(@(x)mean(x.Vertices(:,1)), boxes(1, :, git));
                      newPos = cellfun(@(x)boxesPositions(x)+[-0.01 0.01], intraGrList{git}, 'UniformOutput', false);
                      sigstar(newPos, intraPlist{git});
                    end
                  end
                  % The normal comparisons
                  for git = 1:size(subData, 3)
                    if(isempty(pList{git}))
                      continue;
                    end
                    boxesPositions = arrayfun(@(x)mean(x.Vertices(:,1)), boxes(1, :, git));
                    newPos = cellfun(@(x)boxesPositions(x), grList{git}, 'UniformOutput', false);
                    if(numel(newPos) == 2 && numel(pList{git}) == 1)
                      newPos = newPos{1};
                    end
                    sigstar(newPos, pList{git});
                  end
              end
            end
            obj.plotHandles.handles.box =  boxes;
            setappdata(obj.figureHandle, 'boxData', obj.plotHandles);
            title(obj.axisHandle, obj.figName, 'interpreter', 'none');

            if(isempty(obj.params.styleOptions.xLabel))
            else
              xlabel(obj.params.styleOptions.xLabel);
            end
            if(isempty(obj.params.styleOptions.yLabel))
              ylabel(obj.statisticsName);
            else
              ylabel(obj.params.styleOptions.yLabel);
            end

            set(obj.figureHandle,'Color','w');

            set(obj.axisHandle, 'XTickLabelRotation', obj.params.styleOptions.XTickLabelRotation);
            set(obj.axisHandle, 'YTickLabelRotation', obj.params.styleOptions.YTickLabelRotation);

            title(obj.axisHandle, obj.figName, 'interpreter', 'none');
            set(obj.axisHandle,'Color','w');
            set(obj.figureHandle,'Color','w');
          catch ME
            logMsg(strrep(getReport(ME),  sprintf('\n'), '<br/>'), 'e');
          end

          ui = uimenu(obj.figureHandle, 'Label', 'Export');
          uimenu(ui, 'Label', 'Figure',  'Callback', {@exportFigCallback, {'*.pdf';'*.eps'; '*.tiff'; '*.png'}, strrep([obj.figFolder, obj.figName], ' - ', '_'), obj.params.saveOptions.saveFigureResolution});
          uimenu(ui, 'Label', 'To workspace',  'Callback', @exportToWorkspace);
          uimenu(ui, 'Label', 'Data (statistics)',  'Callback', @(h,e)obj.exportDataAggregates(bpData, obj.exportFolder));
          uimenu(ui, 'Label', 'Data (full)',  'Callback', @(h,e)obj.exportDataFull(bpData, obj.exportFolder));
        case 'distribution'
        % Since the groups are independent. One subplot per group
          nSquares = ceil(sqrt(size(subData, 3)));

        % Establish cmap entries - NOT this way
          switch obj.params.pipelineProject.factor
            case 'experiment'
              %subData: array sample - label - group
              numColorPlots = size(subData, 2);
            case 'event'
              %subData: array sample - label - group
              numColorPlots = size(subData, 2);
            case 'mixed'
              %fullData: cell label - group - experiment
              numColorPlots = length(fullData);
          end
          for git = 1:size(subData, 3)
            subplotHandles = [];
            subplot(nSquares, nSquares, git);
            hold on;
            if(isempty(obj.params.styleOptions.xLabel))
              xlabel(obj.params.statistic);
            else
              xlabel(obj.params.styleOptions.xLabel);
            end
            if(isempty(obj.params.styleOptions.yLabel))
              ylabel('PDF');
            else
              ylabel(obj.params.styleOptions.yLabel);
            end
            if(isfield(obj.params.styleOptions, 'colormap') && ~isempty(obj.params.styleOptions.colormap))
              cmap = eval(sprintf('%s (%d)', obj.params.styleOptions.colormap, numColorPlots));
            else
              cmap = lines(numColorPlots);
            end
            if(obj.params.styleOptions.invertColormap)
              if(size(cmap, 1) == 1)
                % If the colormap has a single entry we need to add another entry
                if(isfield(obj.params.styleOptions, 'colormap') && ~isempty(obj.params.styleOptions.colormap))
                  cmap = eval(sprintf('%s (%d)', obj.params.styleOptions.colormap, numColorPlots+1));
                else
                  cmap = lines(numColorPlots+1);
                end
                cmap = cmap(end:-1:1, :);
              else
                % Default behavior
                cmap = cmap(end:-1:1, :);
              end
            end
            if(numColorPlots > 1)
              alpha = 0.5;
            else
              alpha = 1;
            end
            validGroups = [];
            %for git = 1:length(obj.groupList)
            % Get the data
            for it = 1:numColorPlots
               switch obj.params.pipelineProject.factor
                case {'experiment', 'event'}
                  numReplicas = 1;
                case 'mixed'
                  %fullData: cell label - group - experiment
                  numReplicas = length(fullData{it}{git})+1;
               end
              for cit = 1:numReplicas
                switch obj.params.pipelineProject.factor
                  case {'experiment', 'event'}
                    curData = squeeze(subData(:, it, git)); %cit is always 1
                  case 'mixed'
                    %fullData: cell label - group - experiment
                    if(cit == 1)
                      curData = squeeze(subData(:, it, git)); % First one is the average
                    else
                      curData = fullData{it}{git}{cit-1};
                    end
                end
                if(isempty(curData))
                  continue;
                end
                % Compute the distribution
                [f, xi] = obj.computeDistribution(curData);
                % The actual plot
                switch obj.params.distributionEstimation.mode
                  case {'unbounded', 'positive', 'histogram line'}
                    h = plot(xi, f, 'Color', cmap(it, :));
                  case 'histogram'
                    h = bar(xi, f, 'FaceColor', cmap(it, :), 'EdgeColor', cmap(it, :)*0.5, 'FaceAlpha', alpha, 'EdgeAlpha', alpha);
                  case 'raw'
                    h = bar(xi, f, 'FaceColor', cmap(it, :), 'EdgeColor', cmap(it, :)*0.5, 'FaceAlpha', alpha, 'EdgeAlpha', alpha);
                    if(isempty(obj.params.styleOptions.yLabel))
                      ylabel('count');
                    else
                      ylabel(obj.params.styleOptions.yLabel);
                    end
                 case 'raw line'
                    h = plot(xi, f, 'Color', cmap(it, :));
                    if(isempty(obj.params.styleOptions.yLabel))
                      ylabel('count');
                    else
                      ylabel(obj.params.styleOptions.yLabel);
                    end
                end
                if(cit == 1)
                  % Stronger average
                  if(strcmpi(obj.params.pipelineProject.factor, 'mixed'))
                    h.LineWidth = 2;
                  end
                  h.DisplayName = xList{it};
                  obj.plotHandles = [obj.plotHandles; h];
                  subplotHandles = [subplotHandles; h];
                end
              end
            end
            l = legend(gca, subplotHandles);
            l.Location = 'best';
            l.Box = 'off';
            title([obj.figName ' - ' legendList{git}], 'interpreter', 'none');
            set(gca, 'XTickLabelRotation', obj.params.styleOptions.XTickLabelRotation);
            set(gca, 'YTickLabelRotation', obj.params.styleOptions.YTickLabelRotation);
            box on;
            set(gca,'Color','w');
          end
          set(obj.figureHandle,'Color','w');
          ui = uimenu(obj.figureHandle, 'Label', 'Export');
          uimenu(ui, 'Label', 'Figure',  'Callback', {@exportFigCallback, {'*.pdf';'*.eps'; '*.tiff'; '*.png'}, strrep([obj.figFolder, obj.figName], ' - ', '_'), obj.params.saveOptions.saveFigureResolution});
      end
      if(obj.params.saveOptions.saveFigure)
        %[obj.figFolder, obj.figName, '.', obj.params.saveOptions.saveFigureType]
        export_fig([obj.figFolder, obj.figName, '.', obj.params.saveOptions.saveFigureType], ...
                    sprintf('-r%d', obj.params.saveOptions.saveFigureResolution), ...
                    sprintf('-q%d', obj.params.saveOptions.saveFigureQuality), '-nocrop', obj.figureHandle);
      end
      if(obj.params.pipelineProject.automaticallyExportData)
        try
          obj.exportDataFull(bpData, obj.exportFolder, true)
        catch ME
          logMsg(strrep(getReport(ME),  sprintf('\n'), '<br/>'), 'w');
          logMsg('Could not export the file', 'e');
        end
      end
      %--------------------------------------------------------------------
      function exportToWorkspace(~, ~)
        assignin('base', 'boxPlotData', obj.plotHandles);
        if(~isempty(obj.guiHandle))
          logMsg('boxPlotData data succesfully exported to the workspace. Modify its components to modify the figure', obj.guiHandle, 'w');
        else
          logMsg('boxPlotData data succesfully exported to the workspace. Modify its components to modify the figure', 'w');
        end
      end
    end
    %----------------------------------------------------------------------
    function [f, xi] = computeDistribution(obj, data)
      if(~isempty(obj.params.distributionEstimation.additionalOptions))
        additionalOptions = obj.params.distributionEstimation.additionalOptions;
      else
        additionalOptions = [];
      end
      obj.params.distributionEstimation
      switch obj.params.distributionEstimation.mode
        case 'unbounded'
          if(isempty(additionalOptions))
            [f, xi] = ksdensity(data);
          else
            try
              evalLine = sprintf('ksdensity(data, %s)', additionalOptions);
              [f, xi] = eval(evalLine);
            catch
              logMsg(sprintf('Invalid additionalOptions. Eval line: %s', evalLine), 'e');
              return;
            end
          end
           %h = plot(xi, f, 'Color', cmap(git, :));
        case 'positive'
          if(isempty(additionalOptions))
            [f, xi] = ksdensity(data, 'support', 'positive');
          else
            try
              evalLine = sprintf('ksdensity(data, ''support'', ''positive'', %s)', additionalOptions);
              [f, xi] = eval(evalLine);
            catch
              logMsg(sprintf('Invalid additionalOptions. Eval line: %s', evalLine), 'e');
              return;
            end
          end
          %h = plot(xi, f, 'Color', cmap(git, :));
        case {'histogram', 'histogram line'}
          if(isempty(obj.params.distributionEstimation.distributionBins) || (~ischar(obj.params.distributionEstimation.distributionBins) && obj.params.distributionEstimation.distributionBins == 0))
            try
              bins = sshist(data);
            catch
              bins = 10;
            end
          elseif(ischar(obj.params.distributionEstimation.distributionBins))
            bins = eval(obj.params.distributionEstimation.distributionBins);
            if(bins == 0)
              try
                bins = sshist(data);
              catch
                bins = 10;
              end
              if(isempty(bins))
                bins = 10;
              end
            end
          else
            bins = obj.params.distributionEstimation.distributionBins;
          end
          if(isempty(additionalOptions))
            [f, xi] = hist(data, bins);
          else
            try
              evalLine = sprintf('hist(data, bins, %s)', additionalOptions);
              [f, xi] = eval(evalLine);
            catch
              logMsg(sprintf('Invalid additionalOptions. Eval line: %s', evalLine), 'e');
              return;
            end
          end
          % Now normalize the histogram
          area = trapz(xi, f);
          f = f/area;
          %h = bar(xi, f/area, 'FaceColor', cmap(git, :), 'EdgeColor', cmap(git, :)*0.5, 'FaceAlpha', alpha, 'EdgeAlpha', alpha);
        case {'raw', 'raw line'}
          if(isempty(obj.params.distributionEstimation.distributionBins) || (~ischar(obj.params.distributionEstimation.distributionBins) && obj.params.distributionEstimation.distributionBins == 0))
            try
              bins = sshist(data);
            catch
              bins = 10;
            end
          elseif(ischar(obj.params.distributionEstimation.distributionBins))
            bins = eval(obj.params.distributionEstimation.distributionBins);
            if(bins == 0)
              try
                bins = sshist(data);
              catch
                bins = 10;
              end
              if(isempty(bins))
                bins = 10;
              end
            end
          else
            bins = obj.params.distributionEstimation.distributionBins;
          end
          if(isempty(additionalOptions))
            [f, xi] = hist(data, bins);
          else
            try
              evalLine = sprintf('hist(data, bins, %s)', additionalOptions);
              [f, xi] = eval(evalLine);
            catch
              logMsg(sprintf('Invalid additionalOptions. Eval line: %s', evalLine), 'e');
              return;
            end
          end
        otherwise
          xi = [];
          f = [];
      end
    end  
    %----------------------------------------------------------------------
    function updateFigure(obj, params)
      
    end
    
    %----------------------------------------------------------------------
    function init(obj, projexp, optionsClass, msg, varargin)
      %--------------------------------------------------------------------
      [obj.params, var, newClass] = processFunctionStartup(optionsClass, varargin{:});
      % Define additional optional argument pairs
      obj.params.pbar = [];
      obj.params.gui = [];
      obj.params.loadFields = {};
      obj.params.multiStatistic = 'none'; % none / init / present
      obj.params.plotDataFull = [];
      % Parse them
      obj.params = parse_pv_pairs(obj.params, var);
      obj.params = barStartup(obj.params, msg);
      %obj.params = obj.params;
      obj.guiHandle = obj.params.gui;
      %--------------------------------------------------------------------
      
      % Fix in case for some reason the group is a cell
      if(iscell(obj.params.group))
        obj.mainGroup = obj.params.group{1};
      else
        obj.mainGroup = obj.params.group;
      end
      obj.fullGroupList = [];
      % For the multi statistic
      obj.multiStatistic = obj.params.multiStatistic;
      obj.plotDataFull = obj.params.plotDataFull;
      
      % Consistency checks
      if(obj.params.saveOptions.onlySaveFigure)
        obj.params.saveOptions.saveFigure = true;
      end
      if(ischar(obj.params.styleOptions.figureSize))
        obj.params.styleOptions.figureSize = eval(obj.params.styleOptions.figureSize);
      end
      
      % Check if its a project or an experiment
      if(isfield(projexp, 'saveFile'))
        [~, ~, fpc] = fileparts(projexp.saveFile);
        if(strcmpi(fpc, '.exp'))
          obj.mode = 'experiment';
          experiment = projexp;
          baseFolder = experiment.folder;
        else
          obj.mode = 'project';
          project = projexp;
          baseFolder = project.folder;
        end
      else
        obj.mode = 'project';
        project = projexp;
        baseFolder = project.folder;
      end
      
      % Create necessary folders
      if(obj.params.saveOptions.saveFigure)
        if(strcmpi(obj.mode, 'experiment'))
          switch obj.params.saveOptions.saveBaseFolder
            case 'experiment'
              baseFolder = experiment.folder;
            case 'project'
              baseFolder = [experiment.folder '..' filesep];
          end
        else
          baseFolder = project.folder;
        end
        if(~exist(baseFolder, 'dir'))
          mkdir(baseFolder);
        end
        obj.figFolder = [baseFolder 'figures' filesep];
        if(~exist(obj.figFolder, 'dir'))
          mkdir(obj.figFolder);
        end
        obj.exportFolder = [baseFolder 'exports' filesep];
        if(~exist(obj.exportFolder, 'dir'))
          mkdir(obj.exportFolder);
        end
      end
      % This block should be the same for any statistics plot
      stat = obj.params.statistic;
      obj.params
      if(strcmpi(stat, 'ask') || strcmpi(stat, 'all'))
        if(isfield(projexp, 'checkedExperiments'))
          exp = [projexp.folderFiles projexp.experiments{find(projexp.checkedExperiments, 1, 'first')} '.exp'];
          tmpClass = optionsClass.setExperimentDefaults(exp);
        else
          tmpClass = optionsClass.setExperimentDefaults(projexp);
        end
        %tmpClass = defClass.setExperimentDefaults([]);
        statList = tmpClass.statistic(1:end-2); % Removing last one since it's going to be empty and 'all'
        if(strcmpi(stat, 'ask'))
          [selection, ok] = listdlg('PromptString', 'Select statistics to plot', 'ListString', statList, 'SelectionMode', 'multiple');
          if(~ok)
            return;
          end
        elseif(strcmpi(stat, 'all'))
          selection = 1:length(statList);
        end
        externalPlotDataFull = {};
        for it = 1:length(selection)
          obj.fullGroupList = [];
          logMsg(sprintf('%s for: %s', msg, statList{selection(it)}));
          obj.params.statistic = statList{selection(it)};
          if(it == 1)
           obj.multiStatistic = 'init';
          else
            obj.multiStatistic = 'present';
            obj.plotDataFull = externalPlotDataFull{it};
          end
          if(obj.getData(newClass, projexp, statList(selection)))
            if(it == 1)
              externalPlotDataFull = obj.plotDataFull;
            end
            obj.createFigure();
          end
          obj.cleanup();
          autoArrangeFigures();
        end
      else
        if(obj.getData(newClass, projexp, stat))
          obj.createFigure();
        end
        obj.cleanup();
      end
      %--------------------------------------------------------------------
      barCleanup(obj.params);
      %--------------------------------------------------------------------
    end
    
    %----------------------------------------------------------------------
    function cleanup(obj)
      % Execute additional figure commands
      if(~isempty(obj.params.additionalFigureOptions) && ischar(obj.params.additionalFigureOptions))
        obj.params.additionalFigureOptions = java.io.File(obj.params.additionalFigureOptions);
      end
      if(~isempty(obj.params.additionalFigureOptions) && obj.params.additionalFigureOptions.isFile)
        %obj.params.additionalFigureOptions.getAbsoluteFile
        run(char(obj.params.additionalFigureOptions.getAbsoluteFile));
      end
      if(obj.params.saveOptions.onlySaveFigure)
       close(obj.figureHandle);
      end
    end
    
    %--------------------------------------------------------------------
    function exportDataAggregates(obj, bpData, baseFolder)
      [fileName, pathName] = uiputfile('.csv', 'Save data', [baseFolder obj.params.statistic obj.params.saveOptions.saveFigureTag '_aggregates.csv']);
      if(fileName == 0)
        return;
      end
      fID = fopen([pathName fileName], 'w');
      names = {'label', 'group', 'median', 'mean', 'std', 'N', 'Q1', 'Q3', 'IQR', 'min', 'max'};

      data = obj.plotHandles.statistics;
      for it = 1:length(names)
        lineStr = sprintf('"%s"', names{it});
        for cit = 1:size(bpData, 2)
          for git = 1:size(bpData, 3)
            if(it == 1)
              lineStr = sprintf('%s,"%s"', lineStr, strrep(obj.groupLabels{git}, ',', ' -'));
            elseif(it == 2)
              lineStr = sprintf('%s,"%s"', lineStr, obj.fullGroupList{1}{cit});
            else
              %data.(names{it})
              if(ndims(data.(names{it})) == 3)
                lineStr = sprintf('%s,%.4f', lineStr, data.(names{it})(1, git, cit));
              else
                lineStr = sprintf('%s,%.4f', lineStr, data.(names{it})(cit, git));
              end
            end
          end
        end
        lineStr = sprintf('%s\n', lineStr);
        fprintf(fID, lineStr);
      end
      fclose(fID);
    end

    %--------------------------------------------------------------------
    function exportDataFull(obj, bpData, baseFolder, varargin)
      if(nargin > 3)
        automatic = varargin{1};
      else
        automatic = false;
      end
      if(~automatic)
        [fileName, pathName] = uiputfile('.csv', 'Save data', [baseFolder obj.params.statistic obj.params.saveOptions.saveFigureTag '_full.csv']);
        if(fileName == 0)
          return;
        end
      else
        pathName = baseFolder;
        fileName = [obj.params.statistic obj.params.saveOptions.saveFigureTag '_autoExport.csv'];
      end
      fID = fopen([pathName fileName], 'w');
      % +2 for the headers
      for it = 1:(size(bpData, 1)+2) 
        mainIdx = it-2;
        lineStr = '';
        for cit = 1:size(bpData, 2)
          for git = 1:size(bpData, 3)
            if(it == 1)
              if(git == 1)
                lineStr = sprintf('%s,"%s"', lineStr, strrep(obj.groupLabels{git}, ',', ' -'));
              else
                lineStr = sprintf('%s,"%s"', lineStr, strrep(obj.groupLabels{git}, ',', ' -'));
              end
            elseif(it == 2)
              lineStr = sprintf('%s,"%s"', lineStr, obj.fullGroupList{1}{cit});
            else
              if(isnan(bpData(mainIdx, cit, git)))
                lineStr = sprintf('%s,%s', lineStr, '');
              else
                lineStr = sprintf('%s,%.4f', lineStr, bpData(mainIdx, cit, git));
              end
            end
          end
        end

        % Stop when everything is NaN - DONT
%         if(mainIdx >= 1 && all(all(isnan(bpData(mainIdx, :, :)))))
%            lineStr = sprintf('%s\r\n', lineStr(2:end));
%            break;
%         end
        % 2:end to avoid the first comma NOT ANYMORE - WHAT?
        lineStr = sprintf('%s\r\n', lineStr(2:end));
        fprintf(fID, lineStr);
      end
      fclose(fID);
    end
  end
end