setwd("./templates")

namefile <- "2.txt"

namestr <- paste("./../clustall9k/", namefile, sep = "")

b <- read.csv(namestr, header = F, row.names = NULL, stringsAsFactors = F)

for (i in 1:ncol(b)) { 

  for (j in 1:nrow(b)) {

    if (i > 1) {

      b[j,1] <- paste(b[j,1], b[j, i], sep = "")

    }

  }

}

b <- as.data.frame(b[[1]])

b[] <- lapply(b, as.character)

b[] <- lapply(b, tolower)

cbindPad <- function(...){

  args <- list(...)

  n <- sapply(args,nrow)

  mx <- max(n)

  pad <- function(x, mx){

    if (nrow(x) < mx){

      nms <- colnames(x)

      padTemp <- matrix(NA, mx - nrow(x), ncol(x))

      colnames(padTemp) <- nms

      if (ncol(x)==0) {

        return(padTemp)

      } else {

        return(rbind(x,padTemp))

      }

    }

    else{

      return(x)

    }

  }

  rs <- lapply(args,pad,mx)

  return(do.call(cbind,rs))

}

##Platelet Change To Rules

b[] <- lapply(b[], function(x) gsub("Platelet","Platelets",x, ignore.case = T))

b[] <- lapply(b[], function(x) gsub("Plateletss","Platelets",x, ignore.case = T))

b[] <- lapply(b[], function(x) gsub("Platelet count","Platelets",x, ignore.case = T))

b[] <- lapply(b[], function(x) gsub("Platelets:","Platelets",x, ignore.case = T))

b[] <- lapply(b[], function(x) gsub("\\(plts\\)","",x, ignore.case = T))

b[] <- lapply(b[], function(x) gsub("\\(plt\\)","",x, ignore.case = T))

b[] <- lapply(b[], function(x) gsub(">",">=",x, ignore.case = T))

b[] <- lapply(b[], function(x) gsub("greater than or equal to",">=",x, ignore.case = T))

b[] <- lapply(b[], function(x) gsub(">==",">=",x, ignore.case = T))

b[] <- lapply(b[], function(x) gsub(" cell/mm\\^3","/mcL",x, ignore.case = T))

b[] <- lapply(b[], function(x) gsub(" */ *mm\\^3","/mcL",x, ignore.case = T))

b[] <- lapply(b[], function(x) gsub("/microliters","mcL",x, ignore.case = T))

b[] <- lapply(b[], function(x) gsub(" *cells/mcL","/mcL",x, ignore.case = T))

b[] <- lapply(b[], function(x) gsub(" mm\\^3","/mcL",x, ignore.case = T))

b[] <- lapply(b[], function(x) gsub(" *x *10\\^9/L","000/mcL",x, ignore.case = T))

b[] <- lapply(b[], function(x) gsub("100000","100 000",x, ignore.case = T))

b[] <- lapply(b[], function(x) gsub("ul","mcL",x, ignore.case = T))

b[] <- lapply(b[], function(x) gsub("mcl","mcL",x, ignore.case = T))

l <- as.data.frame(matrix(0, ncol = 1, nrow = (nrow(b))))

for (i in 1:nrow(b)) {

  if (length( unlist(strsplit(unlist(strsplit (b[i,1], "[^[:alnum:]<>=./]")), "(?=[<>=./])", perl = TRUE))) > ncol(l)) {

    l <- as.data.frame(matrix(0, ncol = length( unlist(strsplit(unlist(strsplit (b[i,1], "[^[:alnum:]<>=./]")), "(?=[<>=./])", perl = TRUE))), nrow = (nrow(b))))

  }

}

n <- 1

for (i in 1:nrow(b)) {

  n <- as.numeric(as.numeric(ncol(l)) - as.numeric(length( unlist(strsplit(unlist(strsplit (b[i,1], "[^[:alnum:]<>=./]")), "(?=[<>=./])", perl = TRUE)))))

  l[i,] <- c( unlist(strsplit(unlist(strsplit (b[i,1], "[^[:alnum:]<>=./]")), "(?=[<>=./])", perl = TRUE)), rep("IGNORE", times = as.numeric(n)))

}

l <- cbind(l, rep(0, times = nrow(l)))

names(l)[ncol(l)] <- "wc"

for (i in 1:nrow(l)) {

  l[i,ncol(l)] <- length(grep("IGNORE", l[i,]))

}

row1 <- unlist(strsplit(unlist(strsplit(b[which.max(l$wc),1], "[^[:alnum:]<>=./]")), "(?=[<>=./])", perl = TRUE))

wc_df <- as.data.frame(l$wc) 

for (i in 1:nrow(wc_df)) {

  wc_df[i,1] <- ncol(l) - wc_df[i,1] - 1

  wc_df[i,2] <- 0

}

l <- l[1:(ncol(l)-1)]

base_str <- row1

text_df <- l

text_df_2 <- text_df

loc_df <- as.data.frame(matrix(0, ncol = 2*length(base_str), nrow = nrow(text_df)), stringsAsFactors = F)

rem_df <- as.data.frame(matrix('f', ncol = 1, nrow = nrow(text_df)), stringsAsFactors = F)

fillins_df <- as.data.frame(matrix(NA, ncol = length(base_str), nrow = nrow(text_df)), stringsAsFactors = F)

temp_df <- as.data.frame(matrix(NA, ncol = length(base_str) + 1, nrow = 1), stringsAsFactors = F)

i <- 1

j <- 1

diff <- 0

count <- 0

threshold <- 0.1

temp_loc <- 1

fillins_loc <- 1

loc <- 1

blank_check <- F

match_check <- F

while (j <= length(base_str)) {

  n <- 1

  k <- 1

  while (n <= nrow(text_df)) {

    #print("n:")

    #print(n)

    while (k <= (ncol(text_df)-diff)) {

      #print("k:")

      #print(k)

      if ( ((paste(text_df[n, k:(k+diff)], collapse = ' ')) == (paste(base_str[i:j], collapse = ' '))) ) {

        if (loc == 1) {

          match_check <- T

          count <- count + 1

          loc_df[n,loc] <- k

          loc_df[n,loc+1] <- k + diff

          k <- ncol(text_df)-diff+1

        }

        else if (diff == 0) {

          if (k > loc_df[n, loc-1]) {

            match_check <- T

            count <- count + 1

            loc_df[n,loc] <- k

            loc_df[n,loc+1] <- k + diff

            k <- ncol(text_df)-diff+1

          }

        } else {

            if (k == (loc_df[n,loc-2])) {

              match_check <- T

              count <- count + 1

              loc_df[n,loc] <- k

              loc_df[n,loc+1] <- k + diff

              k <- ncol(text_df)-diff+1

            }   

          }

        }

      k <- k + 1

    }

    if (match_check == F) {

      rem_df[n,1] <- 't'

    }else {

      print("YOU HIT A MATCH")

    }

    n <- n + 1

    k <- 1

    match_check <- F

  }

  if (count >= (nrow(text_df)*threshold)) {

    if (count < nrow(text_df)) {

      threshold <- ((threshold)*nrow(text_df_2))/(sum(rem_df$V1 == 'f'))

    }

    text_df <- text_df[which(rem_df$V1 == 'f'),]

    loc_df <- loc_df[which(rem_df$V1 == 'f'),]

    fillins_df <- fillins_df[which(rem_df$V1 == 'f'),]

    wc_df <- wc_df[which(rem_df$V1 == 'f'),]

    rem_df <- subset(rem_df, V1 == 'f')

    if (loc == 1) {

      for (n in 1:nrow(loc_df)) {

        if (loc_df[n, loc] != 1) {

          temp_df[1,temp_loc]  <- "BLANK"

          temp_loc <- temp_loc + 1

          blank_check <- T

          fillins_loc <- fillins_loc + 1

          break

        }

      }

    }else {

        for (n in 1:nrow(loc_df)) {

          if ((loc_df[n,loc]) - (loc_df[n,loc-1]) > 1) {

            temp_df[1,temp_loc]  <- "BLANK"

            temp_loc <- temp_loc + 1

            blank_check <- T

            fillins_loc <- fillins_loc + 1

            break

          }

        }

    }

    if (diff > 0) {

      temp_loc <- temp_loc - 1

      for (x in 1:nrow(loc_df)) {

        loc_df[x, loc - 2] <- loc_df[x, loc]

        loc_df[x, loc - 1] <- loc_df[x, loc + 1]

        loc_df[x, loc] <- 0 

        loc_df[x, loc + 1] <- 0 

      }

      loc <- loc - 2

    }

    if (blank_check == T) {

      print ("BLANK_CHECK")

      print (loc)

      print("161")

      for (x in 1:nrow(fillins_df)) {

        print("163")

        if (loc == 1) {

          print("165")

          if (loc_df[x, loc] != 1) {

            print("167")

            fillins_df[x, fillins_loc] <- paste(text_df[x, 1:(loc_df[x,loc]-1)], collapse = ' ')

          }

        }

        else {

          print("172")

          if ((loc_df[x,loc]) - (loc_df[x,loc-1]) > 1) {

            print("174")

            fillins_df[x, fillins_loc] <- paste(text_df[x, (loc_df[x,loc-1] + 1):(loc_df[x,loc]-1)], collapse = ' ')

          }

        }

      }

    }

    print("180")

    temp_df[temp_loc] <- paste(base_str[i:j], collapse = ' ')

    temp_loc <- temp_loc + 1

    loc <- loc + 2

    j <- j + 1

    #threshold <- ((threshold)*nrow(text_df_2))/(sum(rem_df$V1 == 'f'))

    #text_df <- text_df[which(rem_df$V1 == 'f'),]

    #loc_df <- loc_df[which(rem_df$V1 == 'f'),]

    #fillins_df <- fillins_df[which(rem_df$V1 == 'f'),]

    #wc_df <- wc_df[which(rem_df$V1 == 'f'),]

    #rem_df <- subset(rem_df, V1 == 'f')

  }else {

    #print("YOU HIT ELSE")

    if (diff > 0) {

      i <- j

      j <- i

    }else {

      i <- i + 1

      j <- i

    }

    loc_df[loc] <- 0

    loc_df[loc+1] <- 0

    rem_df$V1 <- 'f'

  }

  print("this is i: ")

  print (i)

  count <- 0

  diff <- j - i

  print("this is diff:")

  print(diff)

}

##This part could be cleaner

loc_df <- loc_df[, colSums(loc_df != 0) > 0]

blank_check_end <- F

for (n in 1:nrow(loc_df)) {

  if (loc_df[n, ncol(loc_df)] < wc_df[n, 1]) {

    blank_check_end <- T

  }

}

if (blank_check_end == T) {

  fillins_loc = fillins_loc + 1

  temp_df[temp_loc] <- "BLANK"

  for (x in 1:nrow(fillins_df)) {

    fillins_df[x, fillins_loc] <- paste(text_df[x, (loc_df[x,ncol(loc_df)] + 1):ncol(text_df)], collapse = ' ')

  }

}

for (x in 1:nrow(fillins_df)) {

  fillins_df[x, fillins_loc] <- gsub("IGNORE", "", fillins_df[x, fillins_loc])

}

##Part where we do the counting

fillins_df <- fillins_df[, colSums(is.na(fillins_df)) != nrow(fillins_df)]

for (i in 1:ncol(fillins_df)) {

  names(fillins_df)[i] <- paste("Blank", as.character(i), collapse = ' ')

}

fillins_df <- sapply(fillins_df, as.character)

fillins_df[is.na(fillins_df)] <- ""

i <- 1

while (i <=  ncol(fillins_df)) {

  fillins_df <- as.data.frame(cbind (fillins_df[,c(1:i)], rep(1, times = nrow(fillins_df)), fillins_df[,-c(1:i)]))

  i <- i+ 2

}

for (i in 1:ncol(fillins_df)) {

  if (i %% 2 == 0) {

    colnames(fillins_df)[i] <- "Frequency"

    fillins_df[,i] <- sapply(fillins_df[,i], as.character)

    fillins_df[,i] <- sapply(fillins_df[,i], as.numeric)

  }

}

for (i in 1:ncol(fillins_df)) {

  if (i %% 2 == 1) {

    colnames(fillins_df)[i] <- paste("Blank", (i+1)/2, collapse ="")

    fillins_df[,i] <- sapply(fillins_df[,i], as.character)

  }

}

for (j in 1:ncol(fillins_df)) {

  for (i in 1:nrow(fillins_df)) {

    if (j %% 2 == 1) {

      for (k in 1:i) {

        if (k < i) {

          if (trimws(fillins_df[i, j], which = "both") == trimws(fillins_df[k,j], which = "both")) {

            fillins_df[k, j+1] <- fillins_df[k, j+1] + 1

            fillins_df[i, j] <- paste("IGNORE", i, collapse ="")

            fillins_df[i, j + 1] <- 0 

          }

        }

      }

    }

  }

}

for (j in 1:ncol(fillins_df)) {

  for (i in 1:nrow(fillins_df)) {

    if (j %% 2 == 0) {

      if (fillins_df[i, j] == 0) {

        fillins_df[i, j] <- ""  

      }

    } else {

      fillins_df[i, j] <- gsub("IGNORE \\d+", "", fillins_df[i, j])

    }

  }

}

test2 <- as.data.frame(matrix(0, ncol = 0, nrow = (nrow(fillins_df))))

for (j in 1:ncol(fillins_df)) {

  if (j %% 2 == 1) {

    test <- fillins_df[, c(j, j+1)]

    test[,2] <- as.numeric(test[,2])

    test <- test[order(-test$Frequency),]

    test2 <- cbind(test2, test)

  }

}

fillins_df <- test2[rowSums(is.na(test2))!=(ncol(test2)/2),]

#for (j in 1:ncol(fillins_df)) {

#  for (i in 1:nrow(fillins_df)) {

#    if (j %% 2 == 1) {

#      while (fillins_df[i, j + 1] == "") {

#        for (l in i:nrow(fillins_df)) {

#          if (fillins_df[l, j + 1] != "") {

#            fillins_df[i, j] <- fillins_df[l, j] 

#            fillins_df[i, j + 1] <- fillins_df[l, j + 1] 

#          }

#        }

#        break

#      }

#      fillins_df[l, j] <- ""

#      fillins_df[l, j + 1] <- ""

#    } 

#  }

#}

#temp_df <- cbindPad(temp_df, b)

#fillins_df <- cbindPad(fillins_df, b)

#names(temp_df)[ncol(temp_df)] <- "Cluster Text"

#names(fillins_df)[ncol(fillins_df)] <- "Cluster Text"

write.csv(temp_df, file = paste0(namefile, "TEMPLATE.csv"), row.names = F)

write.csv(fillins_df, file = paste0(namefile, "FILLINS.csv"), col.names = F)