How can I add figures from a looping script in the report generator toolbox?

I have a script that plots 4 different figures with each loop. When running this program from the editor, all my figures appear in the right order. However, when I am in report generator mode, the figures are placed in a .doc document randomly, even though the figures appear in the right order when the report generator runs using part of the string that generates these plots.

Here's what I have tried in the report generator tree:

 Report 'mynamereport'
  Title page(child from report 'mynamereport')

Chapter 1 (child from report 'mynamereport')
    %Evaluate Matlab expresion (child from Chapter 1) %this is where script runs loop and plots figures

 FigureLoop 1 Chapter All (sibling from Chapter 1)

    Figure Snapshot (child from FigureLoop 1)

 %Evaluate Matlab Expression (child from Report'mynamereport') ends the process

Is there a way to have these plots in the report in the order that they are generated? Many thanks!

Answer 1

[Naty S]

Hey This is a bit of a workaround. Ran a for loop from i=1:1:length(findobj('Type','figure')); Inside the for loop i used the figure(i) function to open the figure and than I snapped shot it.

[Julio Martinez]

Thank you Naty! I am back at it and I will try your solution. I had moved onto something else but this is still on my list of tasks to complete. I was told to use a MATLAB version prior to 2014b and works okay. However, I am interested in this workaround so that I can use R2015a which uses some functions that I believe are different in previous versions. I'll let you know how I did!

Answer 2

[Cindy Solomon]

Hi Julio,

Similar to how "drawnow" works when plotting figures, there is an equivalent function for the MATLAB Report Generator called "snapnow". Adding this line after the plot call in your loop should resolve the issue. For more information on this function, see this doc link.

Hope this helps!

[Julio Martinez]

Thanks Cindy!

I tried the 'snapnow' function after checking your post and the background on it. Maybe I am misplacing it because it has not solve this issue. Here's what this part of my script looks like -The whole thing is long- so I am posting just these sections:

The 'for' starts the loop for processing a set of text files where data from MRI images is stored.

    for nFile = 1:num
    Fullname = [Pathname Filename{nFile}];
    fID = fopen(Fullname,'r');

    firstline = fgetl(fID);  % has the sample name in it

    n=10;  % skip the rest of the header
    while n
        tline = fgetl(fID);
        if ~ischar(tline),   break,   end
       % disp(tline)
        n=n-1;
    end

    % now read in the numbers
    n = 1;
    %AA = 0
    while 1
        [AA, count] = fscanf(fID, '%d  %f %f %f %f %f %f %f %f %f %f %f %f',13);%there are 13 columns in each imported text file, file name plus 12 with numbers
        if (count == 0), break, end
        Series(n) = AA(1);
        M(1,n) = AA(2);
        A(1,n) = AA(3);
        S(1,n) = AA(4);
        M(2,n) = AA(5);
        A(2,n) = AA(6);
        S(2,n) = AA(7);
        M(3,n) = AA(8);
        A(3,n) = AA(9);
        S(3,n) = AA(10);
        M(4,n) = AA(11);
        A(4,n) = AA(12);
        S(4,n) = AA(13);
        n = n + 1;
    end

    fclose(fID);

    for n = 1:4
        M1T1 = M(n,1:5)';%Pick values 1-5 in column

        M1T2 = M(n,5:10)';%Pick values 5-10 in column, 5th value is recycled

%Uncertainties for T2 C(n) = 0; Csigma(n) = 0; T2(n) = 0; T2sigma(n) = 0; M0(n) = 0; M0sigma(n) = 0;

%Uncertainties for T1 SS(n) =0; SSsigma(n) = 0; T1(n) = 0; T1sigma(n) = 0; M0T1(n) = 0; M0T1sigma(n) = 0;

     if any(M1T2(n) == 0),  end; %continue,   % no data so move onto the next column

        [xData, yData] = prepareCurveData( T2taus, M1T2 );
        ffunT2 = fittype('M0*exp(-x/T2) + C', 'independent', 'x', 'dependent', 'y');
        options = fitoptions(ffunT2);
        options.Algorithm = 'Levenberg-Marquardt';
        options.Display = 'Off';
        options.Robust ='Bisquare';
        options.StartPoint = [1000 1 1000];
        %Fit model to data.
        [cfun, gof] = fit(xData,yData,ffunT2,options);

        %Create a figure for the plots for T1
        figure('Name', 'T2 fit 1,');

        % Plot fit with data for T2.
%         subplot( 2, 1, 1 ); not needed because we want larger figures
        h = plot(cfun, xData, yData, 'ob' );
        legend( h, 'Mean Signal Intensity vs. TE', 'Custom fit 1', 'Location', 'NorthEast' );
        % Label axes
        xlabel 'Echo Time, TE (ms)'
        ylabel 'Mean Signal Intensity,  (ms)'
        title 'T2'
        grid on
        snapnow
%       Plot residuals.
        figure('Name', 'T2 Residuals')
%         subplot( 2, 1, 2 ); not needed because we want larger figures

        h2 = plot( cfun, xData, yData, '*k','residuals');
        legend( h2, 'Custom fit 1 - residuals', 'Zero Line', 'Location', 'NorthEast' );
        % Label axes
        xlabel 'Echo Time,TE (ms)'
        ylabel 'Mean Signal Intensity (ms)'
        title 'T2'
        grid on
        snapnow

 if any(M1T1(n) == 0),  end; %continue,   % no data so move onto the next column

 %Curve fit for T1 (TR)

     [xData2, yData2] = prepareCurveData( T1taus, M1T1 );
        ffunT1 = fittype('M0T1*(1-SS*exp(-x/T1))', 'independent', 'x', 'dependent', 'y');
        options = fitoptions(ffunT1);
        options.Algorithm = 'Levenberg-Marquardt';
        options.Display = 'Off';
        options.Robust ='Bisquare';
        options.StartPoint = [1000 1 1000];
        %Fit model to data.
        [cfun2, gof] = fit(xData2,yData2,ffunT1,options);    

    % Plot fit with data for T1.

        figure('Name', 'T1 fit 1');
%       subplot( 2, 1, 1 ); not needed because we want larger figures
        h3 = plot(cfun2, xData2, yData2, 'sb' );
        legend( h3, 'Mean Signal Intensity vs. TR', 'Custom fit 1', 'Location', 'SouthEast' );
        % Label axes
        xlabel 'Repetition Time, TR (ms)'
        ylabel 'Mean Signal Intensity'
        title 'T1'
        grid on
        snapnow

% Plot residuals for T1. figure('Name', 'T1 Residuals') % subplot( 2, 1, 2 ); not needed because we want larger figures

        h4 = plot( cfun2, xData2, yData2, '*k','residuals');
        legend( h4, 'Custom fit 1 - residuals', 'Zero Line', 'Location',  'NorthEast' );
        % Label axes
        xlabel ('Repetition Time, TR (ms)')
        ylabel ('Mean Signal Intensity')
        title 'T1'
        grid on
        snapnow

        C(n) = cfun.C;
        M0(n) = cfun.M0;
        T2(n) = cfun.T2;
        uncert = confint(cfun, 0.683);
        T2sigma(n) = (uncert(2,3) - uncert(1,3))/2;
        Csigma(n) = (uncert(2,1) - uncert(1,1))/2;
        M0sigma(n) = (uncert(2,2) - uncert(1,2))/2;
        rsquared(n) = gof.rsquare;
        chisquared(n) = (gof.rmse)^2;

        T2Summary{n} = sprintf( '%.2f %.2f %.2f %f %.2f %.2f %f %.2f   ',   T2(n), T2sigma(n), M0(n), M0sigma(n), C(n), Csigma(n), rsquared(n),    chisquared(n));

        SS(n) = cfun2.SS;
        M0T1(n) = cfun2.M0T1;
        T1(n) = cfun2.T1;
        uncert = confint(cfun2, 0.683);
        T1sigma(n) = (uncert(2,3) - uncert(1,3))/2;
        SSsigma(n) = (uncert(2,2) - uncert(1,2))/2;
        M0T1sigma(n) = (uncert(2,1) - uncert(1,1))/2;
        rsquared(n) = gof.rsquare;
        chisquared(n) = (gof.rmse)^2;

        T1Summary{n} = sprintf( '%.2f %.2f %.2f %.2f %.4f %.4f %f %.2f   ', T1(n), T1sigma(n), M0T1(n), M0T1sigma(n), SS(n),   SSsigma(n),       rsquared(n), chisquared(n));

      end

    Series = sprintf( ' %d_%d   ', Series(1), Series(10));%display(Series)
    Patient = sprintf('%s ', pt);

    Study = sprintf('%.f ', Stdy);
    Temp = sprintf('%.f ', Te);
    Range = sprintf('%.f ', Rnge);

    Subject1 = sprintf('%s  ', sub1);
    State1 = sprintf('%s  ', state1);
    Tissue1 = sprintf('%s  ', tiss1);

    Subject2 = sprintf('%s  ', sub2);
    State2 = sprintf('%s  ', state2);
    Tissue2 = sprintf('%s  ', tiss2);

    Subject3 = sprintf('%s  ', sub3);
    State3 = sprintf('%s  ', state3);
    Tissue3 = sprintf('%s  ', tiss3);

    Subject4 = sprintf('%s  ', sub4);
    State4 = sprintf('%s  ', state4);
    Tissue4 = sprintf('%s  ', tiss4);

    fullLine = [Filename{nFile} Series Study Temp Range Temp Range Patient Subject1  State1 Tissue1 T1Summary{1} Subject2  State2 Tissue2 T1Summary{2} Subject3  State3 Tissue3 T1Summary{3} Subject4  State4 Tissue4 T1Summary{4} Subject1  State1 Tissue1 T2Summary{1} Subject2  State2 Tissue2  T2Summary{2} Subject3  State3 Tissue3 T2Summary{3} Subject4  State4 Tissue4 T2Summary{4} ];

    fprintf(fsaveID,'%s \r\n', fullLine); %send report to .txt file in the order that the loop completes analysis

 end % for nFile (this is the end for the loop)

I have tried placing the snapnow fx after h(n) = plot(xData...) In addition, I placed it after the last end having the same outcome. Any other ideas are most than welcomed!

Thanks a lot!

Answer 3

[Rob Edgar]

Someone at Mathworks figured this out for me. I was not having success with contourf plots in a loop working in report generator. The trick for me was getImpl. Without that, every iteration of fig2 was a duplicated of the first one I had added.

fig2 = Figure(f); fig2.SnapshotFormat = 'png';

fig2 = getImpl(fig2, rpt);