valid_variables_script.R

rm(list=ls())

dir_name <- readline("Introduce the name of the directory please: ")

setwd(dir_name)

source("dependency_installer.R")
source("connection_parameters.R")
source("necessary_functions_connection.R")
#source("required_folder_checker.R")
#source("argument_hasher.R")


dep_list = c("jsonlite", "stringr","DSI","DSOpal","DSLite", "fields", "metafor", "ggplot2", "gridExtra", "data.table", "dsBaseClient", "openxlsx")
install_dependencies(dep_list)

codebook_file <- "20220315_Data Harmonisation.xlsb.xlsx"

codebook_demo <- read.xlsx(codebook_file , sheet = 2 )
codebook_com_and_rf <- read.xlsx(codebook_file , sheet = 3 ) 

codebook_home_med <- read.xlsx(codebook_file , sheet = 4 ) 
codebook_si_sympt <- read.xlsx(codebook_file , sheet = 5 ) 

codebook_treatments <- read.xlsx(codebook_file , sheet = 6 ) 
codebook_labo <- read.xlsx(codebook_file , sheet = 7 ) 

codebook_complications <- read.xlsx(codebook_file , sheet = 8 ) 
codebook_imaging_data <- read.xlsx(codebook_file , sheet = 9 ) 

codebook_lifestyle_diet <- read.xlsx(codebook_file , sheet = 10 ) 
codebook_dates <- read.xlsx(codebook_file , sheet = 11 )

codebook <- rbind(codebook_demo , codebook_com_and_rf)
codebook <- rbind(codebook , codebook_home_med)
codebook <- rbind(codebook , codebook_si_sympt)
codebook <- rbind(codebook , codebook_treatments)
codebook <- rbind(codebook , codebook_labo)
codebook <- rbind(codebook , codebook_complications)
codebook <- rbind(codebook , codebook_imaging_data)

codebook_lifestyle_diet <- codebook_lifestyle_diet[, !names(codebook_lifestyle_diet) %in% c("X2", "X4" , "X10")] 
codebook <- rbind(codebook , codebook_lifestyle_diet)
codebook <- rbind(codebook , codebook_dates)


codebook_col_names <- as.data.frame(codebook$Harmonised.variable.name)

names(codebook_col_names) <- c("col_names")

categoric_vars = c("DMRGENDR", "DMRBORN", "DMRRETH1", "DMROCCU", "DMRHREDU", "DSXOS", "DSXHO", "DSXIC", "TRXAV","TRXRIB","TRXLR","TRXRM","TRXIA","TRXIB","TRXCH","TRXAB","TRXCS","TRXHEP","TRXAF","TRXCP","TRXOT","TRXECM","TRXIV","TRXNIV","TRXNO","TRXOX","TRXRR","TRXTR","TRXVA","TRXPE","TRXPV","TRXIT","TRXNMB","TRXAC","TRXINA","TRXIS","TRXIM","TRXVC","TRXVD","TRXZN",                         "CSXCOT","CSXCTR","SMXASAH","SMXFEA","SMXCOA","SMXSTA","SMXSBA","SMXRNA","SMXMYA","SMXARA","SMXCPA","SMXAPA","SMXINA","SMXNAA","SMXDIA","SMXFAA","SMXHEA","SMXCNA","SMXACA","SMXSLA","SMXTLA","SMXSYA","SMXWHA","SMXLYA","SMXANA","SMXIWA","SMXSRA","SMXBLA","CMXPRG","CMXCVD","CMXCMP","CMXHT","CMXDI","CMXCKD","CMXCLD","CMXCPD","CMXASM","CMXCND","CMXRHE","CMXCCI","CMXCBD","CMXDE","CMXPU","CMXST","CMXLY","CMXAP","RFXSM","RFXFSM","RFXOB","RFXTB","RFXIMD","RFXHIV","RFXAIDS","RFXUI","RFXHC","RFXONC","RFXMN",                         "HMRACI","HMRARB","HMRAHO","HMRNS","HMROS","HMRCS","HMRIS","HMRAV","HMRAB","HMRCOV","IMDXCT","IMDXCTCR","IMDXCTTE","IMDXCTAB","IMDXXR","IMDXPN",                         "COXRD","COXAR","COXPM","COXMOD","COXPT","COXEC","COXSH","COXIO","COXPE","COXST","COXDIC","COXRIO","COXKF","COXHF","COXBC")


#----------------------------------------------------------------------------

#Test if column names are valid
check_column_names <- function(codebook_param, colnames){
  
  str_res <- "The column names:"
  valid_colnames <- c()
  
  for(i in 1:(nrow(colnames))){
    colname <- colnames[i,1]
    number_of_column <- check_valid_name(colname ,  colnames) 
    if(number_of_column != 1){
      str_res<- paste(str_res, colname, sep=" ")
    }else{
      valid_colnames <- c(valid_colnames, colname)
    }
  }
  
  str_res<- paste(str_res,"are not registered in the harmonized data codebook \n", sep=" ")
  
  result <- list("not_colnames" = str_res , "colnames" = valid_colnames)
  
  return (result)
}

#Test if a single variable name is valid
check_valid_name <- function(col_name , col_names){
  
  valid <- 0
  
  if(col_name %in% codebook_col_names$col_names){
    
    valid <- length(grep(col_name, col_names))
    
  }
  
  return (valid)
  
}

remove_space <- function(x){
  searchString <- ' '
  replacementString <- ''
  res = sub(searchString,replacementString,x)
  return(res)
}

remove_spaces_from_ds <- function(ds){
  
  res<- lapply(ds,remove_space )
  
  return(as.data.frame(res))
  
}

is_number <- function(x){
  res <- FALSE
  
  
  if(length(x)!=0){
    x <- str_replace(x,",",".")
    
    aux <- as.numeric(x)
    
    
    if(!is.na(aux))
      res <- TRUE
  }
  
  
  return(res)
  
}

check_values_not_categoric <- function(values, colname , codebook_param){
  
  valid_vals <- values
  
  possible_vals <- possible_values(colname, codebook_param)
  
  for(i in  1:length(values)){
    
    res<- FALSE
    
    value <- values[[i]]
    
    if(is_number(value)){
      value <- str_replace(value,",",".")
      value <- as.numeric(value)
    }
    
    if(is.null(value)){
      res <- TRUE
    }
    
    else if( value == "NA" | value == "nan" | value == ".")
      res <- TRUE
    else{
      
      if(nrow(possible_vals) == 2 & (!grepl("DAT",colname, fixed=TRUE)) & colname !="CSXCTR"){
        
        if(colname=="LBXBEH" | colname=="LBXBEHn" | colname=="LBXBEM"){
          lower = possible_vals[1,1][[1]]
          higher = possible_vals[2,1][[1]]
          
        }else{
          bounds <- as.data.frame(strsplit(possible_vals[1,1], '-'))
          lowerAux <- str_replace(bounds [1,1],",",".")
          higherAux <- str_replace(bounds [2,1],",",".")
          lower <- as.numeric(remove_space(lowerAux))
          higher <- as.numeric(remove_space(higherAux))
        }
        
        
        if ((value >= lower & value <= higher))
          res <- TRUE
      }
      
      if(nrow(possible_vals) == 3 | colname=="CSXCTR"){
        if(value == 0 | value == 1)
          res <- TRUE
      }
      
      if(nrow(possible_vals) > 3){
        
        lower <- strtoi(remove_space(possible_vals[1,1]))
        higher_b <- nrow(possible_vals)-1
        higher <- strtoi(remove_space(possible_vals[higher_b,1]))
        
        if ((value >= lower & value <= higher))
          res <- TRUE
      }
    }
    
    if(res == FALSE)
      valid_vals[i] <- res
    
    
  }
  
  return(valid_vals)
}

possible_values <- function(colname , codebook_param){
  
  if(colname=="LBXBEH" | colname=="LBXBEHn" | colname=="LBXBEM")
    res <- t(as.data.frame(list(-20,20)))
  
  else{
    
    possible_value <- subset(codebook_param,codebook_param$Harmonised.variable.name==colname)[1,5]
    res <- strsplit(possible_value,split="/")
  }
  
  return(as.data.frame(res))
}

possible_values_categoric <- function(codebook_param){
  
  possible_value <- subset(harmonised_data,harmonised_data$Harmonised.variable.name==x)[1,4]
  res <- strsplit(x=possible_value,split="/")
  
  return(as.data.frame(res))
}

check_values_categoric <- function(values, colname){
  
  possible_vals <- possible_values_categoric(colname)
  
  res <- TRUE
  
  for(i in 1:length(values)){
    if(!(values[[i]] %in% as.matrix(remove_spaces_from_ds(possible_vals)))){
      res <- FALSE
    }
    
  }
  
  return(res)
  
}


get_values_from_quantiles <- function(x){
  
  data_summary <- ds.summary(x)
  
  low_quantile <- data_summary$HM$`quantiles & mean`[[1]]
  
  high_quantile <- data_summary$HM$`quantiles & mean`[[7]]
  
  return(list(low_quantile,high_quantile))
  
  
}

notify_error <- function(invalid_name_list,invalid_value_list,wrong_categoric, wrong_categoric_values, missing_numeric, codebook_param){
  
  res <- ""
  
  if(length(invalid_name_list) !=0){
    res <- "There are invalid values in the numeric fields:"
    
    
    for(i in 1:length(invalid_name_list)){
      res <- paste(res, invalid_name_list[i], sep=" ")
    }
    
    res <- paste(res, "\n", sep="")
    
    for(i in 1:length(invalid_name_list)){
      res <- paste(res, error_message(invalid_name_list[i], invalid_value_list[[i]]), sep=" ")
    }
  }
  
  
  if(length(wrong_categoric)!=0){
    
    res <- paste(res, "\n############################################################################ \n", sep="")
    
    res <- paste(res,"\nThe following categoric values are invalid:", sep=" ")
    
    for(i in 1:length(wrong_categoric)){
      res <- paste(res, wrong_categoric[i], sep=" ")
    }
    
    res <- paste(res, "\n", sep="")
    
    for(i in 1:length(wrong_categoric)){
      res <- paste(res, error_message_categoric(wrong_categoric[i], wrong_categoric_values[[i]], codebook_param), sep=" ")
    }
  }
  
  if(length(missing_numeric)!=0){
    
    res <- paste(res, "\n############################################################################ \n", sep="")
    
    res<- paste(res, "\nThe following fields are missing a numeric field:")
    
    for(i in 1:length(wrong_categoric)){
      res <- paste(res, missing_numeric[i], sep=" ")
    }
  }
  
  res <- paste(res, "\n", sep="")
  
  
  return(res)
  
}

error_message_categoric <- function(colname, invalid_values, codebook_param){
  
  res<- "\nValues in the field"
  res<- paste(res, colname, sep=" ")
  res<- paste(res, "should be", sep=" ")
  
  range <- possible_values_categoric(colname , codebook_param)
  
  for(i in 1:nrow(range)){
    res <- paste(res, remove_space(range[i,1]), sep=" ")
  }
  
  
  res<- paste(res, "\nBut values were:", sep=" ")
  
  for(j in 1:length(invalid_values)){
    res<- paste(res, invalid_values[[j]], sep=" ")
  }
  
  res<- paste(res, "\n\n", sep="")
  
  
  return(res)
  
}

error_message <- function(colname, invalid_values){
  
  res<- "\nValues in the field"
  res<- paste(res, colname, sep=" ")
  res<- paste(res, "should be", sep=" ")
  
  if(grepl("numeric", colname,fixed=TRUE))
    new_colname <- strsplit(x=colname,split="_")[[1]][1]
  else
    new_colname <- colname
  
  range <- subset(harmonised_data, harmonised_data$Harmonised.variable.name == new_colname)
  range <- range[5]
  range <- as.data.frame(strsplit(range[1,1], '/'))
  
  #Range of values or null
  if(nrow(range) == 2 & !grepl("DAT",colname, fixed=TRUE)){
    bounds <- as.data.frame(strsplit(range[1,1], '-'))
    lower <- remove_space(bounds [1,1])
    higher <- remove_space(bounds [2,1])
    res<- paste(res, "numbers between", sep=" ")
    res<- paste(res, lower, sep=" ")
    res<- paste(res, "and", sep=" ")
    res<- paste(res, higher, sep=" ")
    res<- paste(res, "(both included)", sep=" ")
  }
  
  if(nrow(range) == 3){
    res<- paste(res, "0 or 1", sep=" ")
  }
  
  if(nrow(range) > 3){
    lower <- strtoi(remove_space(range[1,1]))
    higher_b <- nrow(range)-1
    higher <- strtoi(remove_space(range[higher_b,1]))
    res<- paste(res, "numbers between", sep=" ")
    res<- paste(res, lower, sep=" ")
    res<- paste(res, "and", sep=" ")
    res<- paste(res, higher, sep=" ")
    res<- paste(res, "(both included)", sep=" ")
  }
  
  if(grepl("DAT",colname, fixed=TRUE)){
    res<- paste(res, "dates with the following format: dd/mm/yyyy", sep=" ")
  }
  
  
  res<- paste(res, "\nBut values were:", sep=" ")
  
  for(j in 1:length(invalid_values)){
    res<- paste(res, invalid_values[[j]], sep=" ")
  }
  
  res<- paste(res, "\n", sep="")
  
  if(!is_number(invalid_values[[1]]))
    res<- paste(res, "(It's missing a \"numeric\" field)", sep="")
  
  res<- paste(res, "\n", sep="")
  
  return(res)
  
}

check_valid_values <- function(valid_colnames , codebook_param){
  
  invalid_name_list <- c()
  cannot_analyse_list <- c()
  invalid_values_list <- list()
  wrong_categoric_values <- list()
  wrong_categoric <- c()
  missing_numeric <- c()
  j<- 1
  k <- 1
  
  for(i in  1:(nrow(valid_colnames))){
    
    data_available <- FALSE
    data_table <- "empty"
    colname <- valid_colnames[i, 1]
    
    if(grepl("DMRBORN",colname, fixed=TRUE) | (grepl("DAT",colname, fixed=TRUE)) | (grepl("ISO",colname, fixed=TRUE)) | (grepl("BEF", colname, fixed=TRUE))){
      next 
    }
    
    column <- "data$"
    column <- paste(column, colname, sep="")
    
    tryCatch(
      error = function(cnd) {
        print(paste("Unable to analyse data" , cnd , sep = " "))
        res <- FALSE
      },
      {
      data_table <- as.data.frame(ds.table(column))
      data_available <- TRUE
      }
    )
    
    if(data_available){
      
      cannot_analyse_list <- c(cannot_analyse_list, colname)
      
    }else{
      
      
      if (data_table[[1]] == "All studies failed for reasons identified below")
        values <- get_values_from_quantiles(column)
      else
        values <- row.names(data_table)
      
      numeric_col<- paste(colname,"_numeric", sep="")
      
      if( colname %in% categoric_vars ){
     
        has_numeric <-  numeric_col %in% valid_colnames$`valid_data_colnames(data_colnames)`
        
        if(!has_numeric)
          missing_numeric <- c(missing_numeric, colname)
        
        
        if (data_table[[1]] == "All studies failed for reasons identified below"){
          
          cannot_analyse_list <- c(cannot_analyse_list,colname)
          
        }else  if(!check_values_categoric(values,colname)){
          
          print("Wrong categoric value:")
          print(colname)
          
          wrong_categoric <- c(wrong_categoric, colname)
          wrong_categoric_values[[k]] <- values
          k <- k+1
        }
        
      }else{
        
        if(grepl("numeric", colname,fixed=TRUE))
          new_colname <- strsplit(x=colname,split="_")[[1]][1]
        else
          new_colname <- colname
        
        valid <- check_values_not_categoric(values, new_colname , codebook_param)
        
        if (FALSE %in% valid){
          invalid_name_list <- c(invalid_name_list,colname)
          invalid_values_list[[j]] <- values
          j <- j+1
        }
        
        
        #print(colname)
        #print(values)
        
      }#else
      # print("This variable has a numeric version")
    }
    
  }
  
  missing_numeric
  if(length(invalid_name_list)>0 | length(wrong_categoric)>0 | length(missing_numeric)>0){
    res <- notify_error(invalid_name_list, invalid_values_list, wrong_categoric, wrong_categoric_values, missing_numeric, codebook_param)
  }else{
    res <- "All values are valid \n"
  }
  
  if(length(cannot_analyse_list)>0){
    res <- paste(res, "\n############################################################################ \n", sep="")
    
    res <- paste(res, notify_unable_analyse(cannot_analyse_list), sep="\n" )
  }
  
  
  
  
  return(res)
  
}

notify_unable_analyse <- function(x){
  
  res <- "\nCould not obtain data from the fields:"
  
  for(i in 1:length(x)){
    res <- paste(res, x[i], sep=" ")
  }
  
  return (res)
  
  
}

auxConnections <- connect()
connections <- auxConnections[[1]]
inp <- auxConnections[[2]]

#Conexión a la base de datos

ds.dim("data", datasources = connections)
ds.colnames("data")

#----------------------------------------------------------------------------

#Check valid column names
datastructure_name <- "data"
data_colnames <- ds.colnames(x=datastructure_name, datasources= connections)

data_colnames <- as.data.frame(data_colnames)

check_valid_columns <- check_column_names(codebook ,data_colnames)
valid_columns <- as.data.frame(check_valid_columns$colnames)


result <- ""
result<-check_valid_values(valid_columns, codebook)
print(check_valid_columns)
datashield.logout(connections) 
cat(result)

file_name<- paste(hospital_name,"_invalid_values.txt", sep="")

dir.create("../invalid_values", showWarnings = FALSE)
setwd("../invalid_values")


cat(check_valid_columns,file=file_name,sep="\n")
cat(result,file=file_name,append=TRUE)

datashield.logout(connections)