#x0=1:J
#xx0=x0=c(-1.69,-0.16,0.58,0.626,0.634)#midpoints
#xx0=x0=c(-0.9, -0.72, -0.38, 0.48, 1.52)#midrank
#xx0=x0=c(0.96,  1.13, -0.36, -0.57, -1.16)#ratio score

pvalues<-function(data=data,x0=x0){

J=ncol(data)
N=10000 
allt=NULL


###################################################

###Permutation of a 2*J contigency table###
permute=function(data,score=c(1:ncol(data))){

J=ncol(data)
y=c(rep(1,sum(data[1,])),rep(0,sum(data[2,])))
x=c(rep(score,data[1,]),rep(score,data[2,])) 

y.per=sample(y)

data.per= table(y.per,x)
rbind(data.per[2,],data.per[1,])->data.per
data.per

}

################################





S0=sum(data[1,]*x0)
T10=CATK2(data,x=x0)
T20=mychisq(data)
P10=1-pchisq(T10,df=1)
P20=1-pchisq(T20,df=J-1)
MIN0=min(P10,P20)



 for (i in 1:N){
	permute(data,score=x0)->data.per
     
        T1=CATK2(data.per,x=x0)
        T2=mychisq(data.per)
        P1=1-pchisq(T1,df=1)
        P2=1-pchisq(T2,df=J-1)
        MIN=min(P1,P2)
       tmp=c(catt=T1,Pearson=T2, min2=MIN)
      allt=rbind(allt,tmp)
       }
 

pvalue=c(catt=mean(allt[,1]>T10), pearson=
mean(allt[,2]>=T20), min2=mean(allt[,3]<=MIN0))

pvalue
 
}