clear all;
javaaddpath 'd:\SSS\Bin\sss.jar'
javaaddpath 'd:\SSS\Bin\AbsoluteLayout.jar' 
addpath 'd:\SSS\Bin'
addpath 'd:\utilities'
addpath d:\bfrm\bfrm.utilities.matlab\;
addpath d:\projects\20080401-candidaExpression;
javaaddpath 'd:\bfrm\Bin\dataframe.jar'
import dataframe.*;

cd d:\projects\20080122-copyNum_expression\sss;

dPth='l:\lucas\dataRepository\tissueSamples\breast\miller\rmaA.txt.dfm';
[x sids]=getFacScores('l:\lucas\bfrm\breastGray\myc',dPth);

clin=importdata('l:\lucas\dataRepository\tissueSamples\breast\miller\clinData.txt');
% clin=importdata('l:\lucas\dataRepository\tissueSamples\breast\gray\clinData.txt');
clin.textdata((end-size(clin.data,1)+1):end,(end-size(clin.data,2)+1):end)=num2cell(clin.data);
clin=clin.textdata;
keyHeader=clin(1,:);
clin=clin(2:end,:);

[a b c]=intersect(clin(:,1),sids);
x=x(c,:); sids=sids(c);
clin=clin(b,:);

Des=num2str((1:size(x,2))');
% y=strcmpi(clin(2:end,2),'basal')*1;
y=cell2mat(clin(:,strcmp(keyHeader,'deathTime')==1));
r=cell2mat(clin(:,strcmp(keyHeader,'deathBCInd')==1));
% y=cell2mat(clin(:,strcmp(keyHeader,'disease.time')==1));
% r=cell2mat(clin(:,strcmp(keyHeader,'disease.binary')==1));
kp=find(y>0 & r>=0);
y=y(kp);
r=r(kp);
x=x(kp,:);

n=size(x,1);
itrain=1:n; ivalid=(1:n); ivalid(itrain)=[]; 
% itrain=randsample(length(y),floor(length(y)*2/3)); ivalid=(1:n); ivalid(itrain)=[];
w=ones(n,1); w(ivalid)=0; 

% save 'data.txt' x -ASCII -TABS;
% save 'responses.txt' y -ASCII -TABS;
% save 'r.txt' r -ASCII -TABS;
% save 'weights.txt' w -ASCII -TABS;
% 
% size(x)
% % a=sss.Model; a.start;
% !modelsearch sss.survival.txt


N=size(x,2);
models = TextArray('modelsummary.txt','default','NA',false); models = models.data; 
nm=size(models,1); pm=models(:,2);pm=exp(pm-max(pm));pm=pm/sum(pm);pmax=max(models(:,1)); 
k=min(20,size(models,1));

survivalsummary;
% binaryfitpredict;
tps=zeros(size(y));
tps(strcmpi(clin(2:end,2),'Norm')==1)=1;
tps(strcmpi(clin(2:end,2),'LumA')==1)=2;
tps(strcmpi(clin(2:end,2),'lumb')==1)=3;
tps(strcmpi(clin(2:end,2),'erbb2')==1)=4;
tps(strcmpi(clin(2:end,2),'basal')==1)=5;
tpnm=[cellstr('Normal Type');
    cellstr('Luminal A');
    cellstr('Luminal B');
    cellstr('ERBB2');
    cellstr('Basal')
    ];

%%%%%%%%%%%%redraw with colors and such
% for(i=1:size(x,2))
figure(1)
clf;
[jnk ord]=sortrows(tps);
pttps=tps(ord);
af=avepFit(ord);
% af=x(ord,i);
clr='krbgm';
loc=[];
for(i=1:5)
    hold on;
    plot(find(pttps==i),af(find(pttps==i)),'o','color',clr(i),'markersize',10,'linewidth',3);
    hold off;
    loc=[loc; mean(find(pttps==i))];
end;
set(gca,'fontsize',16,'linewidth',2)
xlabel('Samples','fontsize',16)
ylabel('Probability - Luminal B','fontsize',16);
% legend(tpnm)
% pause;
% end;