#JoinInteractExps3dev.R


#User prompt for std Value
cat("Enter a Standard Deviation value to filter data to be used by REMc ?\n")
inp <- readLines(file("stdin"), n = 1L)
cat(paste("Standard Deviation Value is", inp, "\n"))

#set std deviation multiplier default if no user entry
if(is.numeric(inp)){
  std= inp
  }else{std= 2}
   

#The input files should be entered in order from the greatest number of rows(Orfs) to least.
#Args <- commandArgs(TRUE)
#if(length(Args)==0){
 # std=0
  #}else{
   # std=Args[1]
#}

print(paste("SD=",std))
#Wstudy= getwd()
#dir.create("../JoinFiles")  #(paste0(Wstudy,"/JoinFiles"))
#outDir <- "../JoinFiles"  #paste0(Wstudy,"/JoinFiles")
#dir.create("../REMc")  #(paste0(Wstudy,"/JoinFiles"))
outDir <- "./"  #../REMc"  #paste0(Wstudy,"/JoinFiles")

print(outDir)
#Args= 'asdf'
#Args[1]= "../Exp1/ZScores/ZScores_Interaction.csv" #paste0(Wstudy,"/Exp1/ZScores/ZScores_Interaction.csv")
#Args[2]= "../Exp2/ZScores/ZScores_Interaction.csv" #paste0(Wstudy,"/Exp2/ZScores/ZScores_Interaction.csv")
#Args[3]= "../Exp3/ZScores/ZScores_Interaction.csv" #paste0(Wstudy,"/Exp1/ZScores/ZScores_Interaction.csv")
#Args[4]= "../Exp4/ZScores/ZScores_Interaction.csv" #paste0(Wstudy,"/Exp2/ZScores/ZScores_Interaction.csv")
#Args[3]= "../Exp3/ZScores/ZScores_Interaction.csv"  #paste0(Wstudy,"/Exp3/ZScores/ZScores_Interaction.csv")

inputFile= 'asdf'
if (file.exists('../Exp1/ZScores/ZScores_Interaction.csv')){
  inputFile <- '../Exp1/ZScores/ZScores_Interaction.csv'
}
if (file.exists('../Exp2/ZScores/ZScores_Interaction.csv')){
  inputFile[2] <- '../Exp2/ZScores/ZScores_Interaction.csv'
}
if (file.exists('../Exp3/ZScores/ZScores_Interaction.csv')){
  inputFile[3] <- '../Exp3/ZScores/ZScores_Interaction.csv'
}
if (file.exists('../Exp4/ZScores/ZScores_Interaction.csv')){
  inputFile[4] <- '../Exp4/ZScores/ZScores_Interaction.csv'
}

#Args= inputFile



#outDir <- ArgsJoin[1]  #Output Directory

print(length(inputFile)) #display the number of arguments on terminal

#open required library for the join function (libraries must already be install on R)
library(plyr)
library(dplyr)
library(sos)

#read in the files for your experiment and
#join the two files at a time as a function of how many inputFile, list the larger file first ? in this example X2 has the larger number of genes.
#if X1 has a larger number of genes, switch the order of X1 and X2
if(length(inputFile)==2) {
  X1 <- read.csv(file= inputFile[1],stringsAsFactors = FALSE)
  X2 <- read.csv(file= inputFile[2],stringsAsFactors = FALSE)
  X <- join(X1,X2,by="OrfRep")
  OBH=X[,order(colnames(X))]  #OrderByHeader
  headSel= select(OBH,  contains('OrfRep'), matches('Gene'), contains('Z_lm_K'), contains('Z_Shift_K'),contains('Z_lm_L'), contains('Z_Shift_L'))
  headSel= select(headSel, -'Gene.1') #remove 'Gene.1 column
  headSel2 = select(OBH,  contains('OrfRep'), matches('Gene')) #Frame for interleaving Z_lm with Shift colums
  headSel2 = select(headSel2, -'Gene.1') #remove 'Gene.1 column   #Frame for interleaving Z_lm with Shift colums
  
}else if(length(inputFile)==3){
  X1 <- read.csv(file= inputFile[1],stringsAsFactors = FALSE) #exp1File,stringsAsFactors = FALSE)
  X2 <- read.csv(file= inputFile[2],stringsAsFactors = FALSE) #exp2File,stringsAsFactors = FALSE)
  X3 <- read.csv(file= inputFile[3],stringsAsFactors = FALSE) #exp3File,stringsAsFactors = FALSE)
  X <- join(X1,X2,by="OrfRep")
  X <- join(X,X3,by="OrfRep")
  OBH=X[,order(colnames(X))]  #OrderByHeader
  headSel= select(OBH,  contains('OrfRep'), matches('Gene'), contains('Z_lm_K'), contains('Z_Shift_K'),contains('Z_lm_L'), contains('Z_Shift_L'))
  headSel= select(headSel, -'Gene.1',-'Gene.2') 
  headSel2 = select(OBH,  contains('OrfRep'), matches('Gene')) 
  headSel2 = select(headSel2, -'Gene.1',-'Gene.2') 
  
}else if(length(inputFile)==4){
  X1 <- read.csv(file= inputFile[1],stringsAsFactors = FALSE) #exp1File,stringsAsFactors = FALSE)
  X2 <- read.csv(file= inputFile[2],stringsAsFactors = FALSE) #exp2File,stringsAsFactors = FALSE)
  X3 <- read.csv(file= inputFile[3],stringsAsFactors = FALSE) #exp3File,stringsAsFactors = FALSE)
  X4 <- read.csv(file= inputFile[4],stringsAsFactors = FALSE) #exp4File,stringsAsFactors = FALSE)
  X <- join(X1,X2,by="OrfRep")
  X <- join(X,X3,by="OrfRep")
  X <- join(X,X4,by="OrfRep")
  OBH=X[,order(colnames(X))]  #OrderByHeader
  headSel= select(OBH,  contains('OrfRep'), matches('Gene'), contains('Z_lm_K'), contains('Z_Shift_K'),contains('Z_lm_L'), contains('Z_Shift_L'))
  headSel= select(headSel, -'Gene.1',-'Gene.2',-'Gene.3') 
  headSel2 = select(OBH,  contains('OrfRep'), matches('Gene')) 
  headSel2 = select(headSel2, -'Gene.1',-'Gene.2',-'Gene.3') 
  
}

#headSel$contains('Z_Shift') %>% replace_na(0.001)
headers<-colnames(headSel)
i=0
for(i in 1:length(headers)){
  
  if(grepl("Shift",headers[i])) {
    headSel[headers[i]][is.na(headSel[headers[i]])] =  0.001
  }
  if(grepl("Z_lm_",headers[i])) {
    headSel[headers[i]][is.na(headSel[headers[i]])] =  0.0001
  }
  
}


#2SD option code to exclude Z_lm values less than 2 standard Deviations


#+++++++++++++

REMcRdy= select(headSel,  contains('OrfRep'), matches('Gene'), contains('Z_lm_'))
shiftOnly= select(headSel,  contains('OrfRep'), matches('Gene'), contains('Z_Shift'))

#Code to replace the numeric (.1 .2 .3) headers with experiment names from StudyInfo.txt
Labels <- read.csv(file= "../Code/StudyInfo.csv",stringsAsFactors = FALSE,sep= ",")


#R can't do math reliably within a nested loop, Therefore requires a complex text based work around+++++++++++++++
#Do it with text search and replace or modify. 
#R is just a badly designed or rather badly evolved mutant language!!! Full of problems and inconsistencies!!!
#R causes huge waste of time. 
#R is not worth fixing. It needs to be discontinued both in use and development before it does more harm.
#Using Text search grepl to relabel headers+++++++++++++++++++++++++++++++++++++++++
REMcRdyHdr= colnames(REMcRdy)
REMcRdyLabels= 'asdf'
shftHdr= colnames(shiftOnly)
shiftLabels='asdf'
shiftLabels[1:2]<-shftHdr[1:2]
REMcRdyLabels[1:2]<-REMcRdyHdr[1:2]
  
for(i in 3:(length(shftHdr))){
  if(i==3){
    shiftLabels[3]<-paste0(Labels[1,2],".",shftHdr[3])
    REMcRdyLabels[3]<-paste0(Labels[1,2],".",REMcRdyHdr[3]) }
  if(i==5){
    shiftLabels[5]<-paste0(Labels[1,2],".",shftHdr[5])
    REMcRdyLabels[5]<-paste0(Labels[1,2],".",REMcRdyHdr[5])
  }
  if(i==7){
    shiftLabels[7]<-paste0(Labels[1,2],".",shftHdr[7])
    REMcRdyLabels[7]<-paste0(Labels[1,2],".",REMcRdyHdr[7])
  }
  if(grepl(".1",shftHdr[i],fixed=true)){
    shiftLabels[i]<-paste0(Labels[2,2],".",shftHdr[i])
    REMcRdyLabels[i]<-paste0(Labels[2,2],".",REMcRdyHdr[i])}
  if (grepl(".2",shftHdr[i],fixed=true)){
    shiftLabels[i]<-paste0(Labels[3,2],".",shftHdr[i])
    REMcRdyLabels[i]<-paste0(Labels[3,2],".",REMcRdyHdr[i])}
  if(grepl(".3",shftHdr[i],fixed=true)){
    shiftLabels[i]<-paste0(Labels[4,2],".",shftHdr[i])
    REMcRdyLabels[i]<-paste0(Labels[4,2],".",REMcRdyHdr[i])}
}

for(i in 3:(length(REMcRdyLabels))){
  j=as.integer(i)
  REMcRdyLabels[j]<- gsub("[.]", "_", REMcRdyLabels[j])
  shiftLabels[j]<- gsub("[.]", "_", shiftLabels[j])
}

 colnames(shiftOnly)<- shiftLabels  
 colnames(REMcRdy)<- REMcRdyLabels

#+++++++++++++++++++++++
combI= headSel2 #Starting Template orf, Genename columns
#headersRemc<-colnames(REMcRdy)
#Reoder columns to produce an interleaved set of Z_lm and Shift data for all the cpps.
for(i in 3:length(colnames(REMcRdy))){
  combI=cbind.data.frame(combI, shiftOnly[i])
  combI=cbind.data.frame(combI, REMcRdy[i])
}

#Effort to use R vectorization select data greater than 2 Standard deviations.
#R can't handle variables in the...$(varible). The column names must be explicit. Just another R issue of thoughtlessness.
#REMcGT2<- REMcRdy[abs(REMcRdy$(REMcRdyLabels[3]))>=2 |abs(REMcRdy$REMcRdyLabels[4])>=2 |abs(REMcRdy$REMcRdyLabels[5])>=2 |abs(REMcRdy$REMcRdyLabels[6])>=2]
#another failed R work-around
#REMcGT2<- REMcRdy[abs(REMcRdy[,3])>=2 | abs(REMcRdy[,4])>=2 | abs(REMcRdy[,5])>=2 | abs(REMcRdy[,6])>=2]
#A simple task made difficult by R
#Well R doesn't handle loops well but I may have to try some extreme measures.
#a Fundamental reason RisFuckedUp StackOverflow ...$ and everything else (for, if, +,^,...) that is treated as a function can not accept argument which must be evaluated.
#Therefore just another reason R is shit! This alone should qualify the language to the  garbage bin.
#Working in R is a constant battle to workaround its shortsighted deficits.
#REMcGT2<- REMcRdy[abs(REMcRdy$(REMcRdyLabels[3]))>=2 |abs(REMcRdy$REMcRdyLabels[4])>=2 |abs(REMcRdy$REMcRdyLabels[5])>=2 |abs(REMcRdy$REMcRdyLabels[6])>=2,]
#Vectorization with variable adaptation of Sean's string explicit code will not work because R cannot accept variables. Piss poor language.
#So what the hell to do???
#Basically we just need to 'OR' all the column vectors that are >=2 'True' and let that be a binary vector to apply across the data.frame REMcRdy
#R doesn't allow a vector of vectors! 
#I guess the only way to do this simple task in R is to create a set of explicit variables, 
#then fill each variable with the binary results, next 'OR those results and use that vector of bolean results to select values from data.frame.
#This is totally stupid crazy. But this is R.
#R violates all the fundamental rules of automation. R Requires specific explicit parameters. 
#Since the largest REMc is four experiments make eight discrete vectors(K and L for each of the experiments).
#The totally stupid work-around for R

Vec1= NA
Vec2= NA
Vec3= NA
Vec4= NA
Vec5= NA
Vec6= NA
Vec7= NA
Vec8= NA

if(length(REMcRdy)== 6){
  Vec1= abs(REMcRdy[,3])>=std
  Vec2= abs(REMcRdy[,4])>=std
  Vec3= abs(REMcRdy[,5])>=std
  Vec4= abs(REMcRdy[,6])>=std
  bolVec= Vec1 | Vec2
  REMcRdyGT2= REMcRdy[bolVec,1:2]
  REMcRdyGT2[ ,3:6]= REMcRdy[bolVec,3:6]
  shiftOnlyGT2= shiftOnly[bolVec,1:2]
  shiftOnlyGT2[ ,3:6]= shiftOnly[bolVec,3:6]
}
if(length(REMcRdy)== 8){
  Vec1= abs(REMcRdy[,3])>=std
  Vec2= abs(REMcRdy[,4])>=std
  Vec3= abs(REMcRdy[,5])>=std
  Vec4= abs(REMcRdy[,6])>=std
  Vec5= abs(REMcRdy[,7])>=std
  Vec6= abs(REMcRdy[,8])>=std
  bolVec= Vec1 | Vec2 |Vec3
  REMcRdyGT2= REMcRdy[bolVec,1:2]
  REMcRdyGT2[ ,3:8]= REMcRdy[bolVec,3:8]
  shiftOnlyGT2= shiftOnly[bolVec,1:2]
  shiftOnlyGT2[ ,3:8]= shiftOnly[bolVec,3:8]
}
if(length(REMcRdy)== 10){
  Vec1= abs(REMcRdy[,3])>=std
  Vec2= abs(REMcRdy[,4])>=std
  Vec3= abs(REMcRdy[,5])>=std
  Vec4= abs(REMcRdy[,6])>=std
  Vec5= abs(REMcRdy[,7])>=std
  Vec6= abs(REMcRdy[,8])>=std
  Vec7= abs(REMcRdy[,9])>=std
  Vec8= abs(REMcRdy[,10])>=std
  bolVec= Vec1 | Vec2 |Vec3 |Vec4|Vec5|Vec6|Vec7|Vec8 
  REMcRdyGT2= REMcRdy[bolVec,1:2]
  REMcRdyGT2[ ,3:10]= REMcRdy[bolVec,3:10]
  shiftOnlyGT2= shiftOnly[bolVec,1:2]
  shiftOnlyGT2[ ,3:10]= shiftOnly[bolVec,3:10]
}


if(std!=0){
  REMcRdy= REMcRdyGT2  #[,2:length(REMcRdyGT2)]
  shiftOnly= shiftOnlyGT2  #[,2:length(shiftOnlyGT2)]
}

if(std==0){
  REMcRdy= REMcRdy  #[,2:length(REMcRdy)]
  shiftOnly= shiftOnly  #[,2:length(shiftOnly)]
}
#Yet again R creates a problem, placing hidden "" around the header names. The following 
# is intended to remove those quote so that the "" donot blow up the Java REMc.
#Use ,quote=F in the write.csv statement to fix R output file.

print(paste("SD=",std))
print(getwd())
#write.csv(combI,file = file.path(outDir,"CombinedKLzscores.csv"),row.names = FALSE)
write.csv(REMcRdy,file = file.path(outDir,"REMcRdy_lm_only.csv"),row.names = FALSE, quote=F)
write.csv(shiftOnly,file = file.path(outDir,"Shift_only.csv"),row.names = FALSE, quote=F)

#LabelStd <- read.table(file= "./Parameters.csv",stringsAsFactors = FALSE,sep= ",")
pwd=getwd()
print(getwd)

LabelStd<- read.csv(file= "../Code/StudyInfo.csv",stringsAsFactors = FALSE)
print(std)
LabelStd[,4]= as.numeric(std)
write.csv(LabelStd,file="../Code/Parameters.csv",row.names = FALSE)
write.csv(LabelStd,file="../Code/StudyInfo.csv",row.names = FALSE)
cat(paste("Standard Deviation Value was set as", std, "\n"))