Changing a function to take an input of cells rather than a matrix

So I have a Matlab function created by friend of mine and I need to change it such that it will work properly with an input of cells rather than just a single matrix.

the code is :

%%QDViewer v1.5 September 18 2009

function [] = QDViewer(data,threshold)
    close all %closes figures
    disp(' ');
    disp('Running QDViewer.m');
    disp('To terminate the program (a la ctrl-alt-del) hit CTRL+C.');
    tic;

      %Overload program to accept no input for threshold
      if nargin == 1
          disp('Threshold set to default (35%).');
          threshold = 35;
      else
          if threshold>100 || threshold<0
              error('The value for threshold is supposed to be a percentage of the highest peak, a value between 0 and 100. Please run the program again with an appropriate parameter.');
          else
          disp(['Threshold set to ' int2str(threshold) '%.']);       
          end
      end

      %This new version of the qd program knows how to take the metadata away
      %from the formatted matlab files.

      dotwavelength=data(1,2);
      pixelx = data(2,2);
      pixely = data(3,2);
  %     date = data(4,2);
  %     datestr = int2str(date);
  %     datestr = [datestr(1:4) '-' datestr(5:6) '-' datestr(7:8)];
      runname = ['blinky' int2str(data(5,2))];
      numcycle = data(6,2);
      time = data(7,2); %in seconds
      numwavelengths = data(8,2);
      wavelengths = data(9:(9+numwavelengths-1),2)';
       pointdata = ['[' int2str(dotwavelength) 'nm, ' '05222012' ': ' runname ' (' int2str(pixelx) ',' int2str(pixely) ')]'] %#ok<NOPRT> %This string will be used in plots later.

      %Now remove that data so that we have an array that looks like the old
      %version:
      data = data((9+numwavelengths):end,:);

      result = zeros(numcycle,3);
      maxintensity = max(data);
      maxintensity = maxintensity(2); %global max intensity
      threshold = maxintensity*threshold/100;
      time=(time/60)/numcycle; %time in minutes per cycle
      Z=zeros(numcycle,numwavelengths); %Variable used in the 3D plot

      % Initialize
      count = 0;
      intensity=zeros(1,numcycle);
      for i = 1:numcycle
          submatrix = data((i-1)*numwavelengths+1:i*numwavelengths,2);
          if (max(submatrix))>threshold
              count=count+1;
              result(count,2) = max(submatrix);  %max intensity for field
              intensity(i)    = max(submatrix);
              index=mod(find(submatrix==result(count,2)),numwavelengths);
              if index==0
                  index=numwavelengths;
              end
              result(count,3) = wavelengths(index(1));%wavelength
              result(count,1) = i*time + index(1)*time/numwavelengths;
              %time per cycle + time between wavelengths
          end

          for j=1:numwavelengths
              if submatrix(j)>threshold;
                  Z(i,j)=submatrix(j)/maxintensity*100;
              else
                  Z(i,j)=0;
              end
          end
      end

      %Normalized Intensity
      normintensity = zeros(size(intensity));%Loop normalizes telegraph to be 0 / 1
      for i=1:length(intensity)
          if intensity(i)~=0
              normintensity(i) = 1;
          else
              normintensity(i) = 0;
          end
      end

      %On / Off time probabilities
      %Initialize to zero
      jon = 0;
      kon = 0;
      joff = 0;
      koff = 0;
      oncount = zeros(1,length(normintensity));  %On-times, in cycles (not seconds)
      offcount = zeros(1,length(normintensity));
      for i=1:length(normintensity)
          if normintensity(i)==1
              jon = jon+1;
              koff = koff+1;
              offcount(koff)=joff;
              joff=0;
          elseif normintensity(i)==0
              joff = joff+1;
              kon = kon+1;
              oncount(kon)=jon;
              jon=0;
          end
      end
      oncount(oncount==0)=[]; %remove all zeros.
      offcount(offcount==0)=[];

      onfreq = zeros(1,length(oncount)); %Frequency of a blink being on this long
      offfreq= zeros(1,length(offcount));
      for i = 1:max(length(oncount),length(offcount))
         onfreq(i)= sum(oncount==i);
         offfreq(i)= sum(offcount==i);
      end
      onfreq = onfreq/sum(onfreq); %Normalize to percentages.
      offfreq = offfreq/sum(offfreq);

      %MLE Algorithm (fitting to Probs):
      S_on_MLE = 1/length(oncount)*sum(log(oncount));
      crit_on = (1+1/S_on_MLE) %#ok<NOPRT>
      S_off_MLE = 1/length(offcount)*sum(log(offcount));
      crit_off = (1+1/S_off_MLE) %#ok<NOPRT>

      t_min = 1;
      t_max = numcycle;

      tonlast = (find(onfreq~=0, 1, 'last'));
      if(isempty(tonlast)) 
          tonlast = numcycle;
      end
      t_on = linspace(1,tonlast); %set up our timedomain, starting from lowest resolution to highest occupied on-time.
      A_on = (crit_on-1)/(t_min^(1-crit_on)-t_max^(1-crit_on)); %normalizing constant
      P_on = A_on*t_on.^(-crit_on);

      tofflast = (find(offfreq~=0, 1, 'last'));
      if(isempty(tofflast)) 
          tofflast = numcycle;
      end
      t_off = linspace(1,tofflast);
      A_off = (crit_off-1)/(t_min^(1-crit_off)-t_max^(1-crit_off));
      P_off = A_off*t_off.^(-crit_off);

%%Marcel's Stuff    

      offtimes((1:length(offfreq)),1) = (1:length(offfreq))*time;
      offtimes((1:length(offfreq)),2) = offfreq;
      ontimes((1:length(onfreq)),1) = (1:length(onfreq))*time;
      ontimes((1:length(onfreq)),2) = onfreq;

      [row col] = find(offtimes == 0);
      offtimes(row,:) = [];
      [row col] = find(ontimes == 0);
      ontimes(row,:) = [];

      xlswrite('offtime.xls',offtimes);
      xlswrite('ontime.xls',ontimes);

      F = @(a,x)(a(2)*x.^(-a(1)));
      param = [6 1];
      [fit_off] = nlinfit(offtimes(:,1),offtimes(:,2),F,param);
      F = @(a,x)(a(2)*x.^(-a(1)));
      param = [6 1];
      [fit_on] = nlinfit(ontimes(:,1),ontimes(:,2),F,param);

      display(fit_off);
      display(fit_on);

      %%Plot results
      result=result(1:count,1:3);
      result(:,2)=result(:,2)/maxintensity*100;
      intensity = intensity/maxintensity*100;
      threshold = threshold/maxintensity*100;
      regression=[ones(count,1) result(:,1)]\result(:,3); %#ok<NOPRT>
      trendline(1:numcycle) = regression(1)+regression(2)*time*(1:numcycle);

%     %Blinking telegraph
    figure(1)
    hold on;
    title(['Intensity vs. Time ' pointdata], 'FontSize',16);
    plot(time*(1:numcycle),intensity,'-black');
    plot(time*(1:numcycle),threshold,'-r'); %show threshold line
    xlabel('Time (min)','FontSize',14);
    ylabel('Intensity (normalized a.u.)','FontSize',14);
    axis([0 time*numcycle 0 105]);
    hold off;
end

What I really need to know is how to properly edit the recursive stuff so that my the stuff I'm displaying (fit_off, fit_on, crit_on and crit_off) end up as nice outputs.

Sorry if this isn't so clear. Its hard to explain particularly because I didn't write the code myself and I'm not a pro at Matlab yet.

Thanks,

Marcel

[Matt Fig]

Please give an example of inputs as the function is now and the inputs you would like to pass.

[Marcel]

I want to be able to input an array of cells in 1xN form and within each cell is a Nx2 matrix of times and intensities from an image.

right now I have to manually input all each different matrix I want to input a cell array

Answer 1

[Babak]

All you need to do is to change all the occurances of data(.,.) in your function to data{.,.}. This way wherever you are accessing the elemnts of the matrix with (.,.) you access the coresponding element of the cell with {.,.} instead.

As for where you're saying:

wavelengths = data(9:(9+numwavelengths-1),2)';

you need to keep it same as it is, but again wavelengths will still be a cell and later you need to change all the occurances of wavelengths(.,.) to wavelengths{.,.}

YOu got to do same for the line

submatrix = data((i-1)*numwavelengths+1:i*numwavelengths,2);

[Marcel]

In this case I would need a for loop around everything to access each cell. Is that correct?

[Marcel]

So I selected all the 'data()' and changed them to data{1,k}() where k is the index for my for loop around the algorithm that is: k = 1:length(data).

For some reason with this, the first iteration works just fine but the rest dont.

any ideas?

thanks

[Babak]

You don't need for loop! data{.,.} accesses a component of data object which is a cell similar to data(.,.) which accesses an element of a matrix. by replacing I mean do the following for example, replace:

dotwavelength=data(1,2); pixelx = data(2,2); pixely = data(3,2);

with

dotwavelength=data{1,2}; pixelx = data{2,2}; pixely = data{3,2};

After you do all these replacements, data is considered a cell and you can no longer use this function for input of data as a matrix! the data that is the input to the function has to be a cell object.

Answer 2

[per isakson]

Do you really need to change your friends function?

    function My_QDViewer( my_cell_array, threshold)
        data = ...
        QDViewer( data, threshold );
    end

.

or better something like

    function My_QDViewer( my_input, threshold)
        if isa( my_input, 'double' )
            data = my_input;
        elseif isa( my_input, 'cell' ) 
            % convert to double array
            data = ...
        else
            error( 'My_QDViewer:...', 'Illegal input: ... ),   
        end
        QDViewer( data, threshold );
    end

[Marcel]

So I tried this as well and it wont accept my cell input or my double input anymore. I have what you have but with a for loop in a similar fashion to above. I can't seem to figure out why it won't accept my input. Maybe its just evening brain.

[per isakson]

From my point of view:

1. make sure that your friends QDViewer runs with some test data
2. figure out the exact requirement on the input data of QDViewer
3. write your own program, My_QDViewer, which is a wrapper around QDViewer. The purpose of My_QDViewer is to convert your cell array input data to the double array form required by QDViewer.
4. do not change QDViewer
5. My_QDViewer may call QDViewer several times in a loop