R/formattingFunctions.R

In ncats/RaMP-DB: RaMP (Relational Database of Metabolomic Pathways)

#' Produce RaMP input from metabolite metadata data.frame object or .csv/.xlsx files
#'
#' @description
#' Converts data.frame,  .csv, or .xlsx formatted metabolite metadata into RaMP data input format. The input should
#' have ID sources (e.g. hmdb, kegg, entrez) as column names and the corresponding rows filled with IDs from that
#' source.
#'
#' @details
#' RaMP input format is a character() of entries that look like prefix:ID where the prefix is a code for the ID 
#' system the ID belongs to (e.g. hmdb, kegg, entrez). This function expects input with prefixes as column names, 
#' with IDs belonging to that prefix as rows in the corresponding column. Data can be supplied as a csv file, excel
#' file, or data.frame object. NA values, 'NA' values, columns named with an unsupported prefix, and columns named 
#' with a supported prefix but no data will be ignored by the function. Two IDs in the same string separated by a 
#' semicolon are separated into two entries.
#'
#' A complete and current list of currently supported prefixes for metabolites can be found by running the 
#' getPrefixesFromAnalytes() function, as shown in examples. These will include the following genes/proteins: ensembl,
#' gene_symbol, uniprot, entrez, hmdb, wikidata, EN, ncbiprotein, brenda, chebi. They will also include the following 
#' metabolites: hmdb, chebi, chemspider, kegg, pubchem, CAS, wikidata, LIPIDMAPS, lipidbank, swisslipids, plantfa, 
#' kegg_glycan, rhea-comp, polymer
#'
#' The COMETS data format, which uses HMDB_ID and PUBCHEM as column names, is also supported.
#'
#' @param dataFrame a data.frame object where the column names are prefixes that will be prepended to each identifier 
#' in that column. Please specify either this argument or filePath. Supported prefixes can be found with 
#' getPrefixesFromAnalytes()
#' @param filePath a string containing the file path to a csv that will be read and converted to RaMP data input 
#' format. Column names should be prefixes that will be prepended to each identifier in that column.
#' @param db a RaMP database object, if not specified a new one is created with RaMP::RaMP()
#' @param ... additional arguments that will be passed to readr::read_csv() or readxl::read_excel()
#'
#' @returns A character() object including each identifier in the input prepended to a supported prefix in the 
#' format "prefix:identifier". This result is designed to be input for several RaMP functions.
#'
#' @examples
#'\dontrun{
#' # Retrieve supported ID source lists for metabolites and genes
#' RaMP::getPrefixesFromAnalytes(analyteType = 'metabolite')
#'
#' RaMP::getPrefixesFromAnalytes(analyteType = 'gene')
#'
#' # Example use with demo dataframe data
#' df <- data.frame(ensembl = c('ENSG00000135679', 'ENSG00000141510'), hmdb = c('HMDB0000064', NA), 
#'      fake_ID = c('123', NA))
#'
#' RaMPInput <- createRaMPInput(dataFrame = df)
#'
#' getPathwayFromAnalyte(RaMPInput)
#'
#' # Example use with demo csv data
#'
#' dir <- system.file("extdata", package="RaMP", mustWork=TRUE)
#' ExampleRaMPInputPath <- file.path(dir,"ExampleRaMPInput.csv")
#'
#' RaMPInput <- createRaMPInput(filePath = ExampleRaMPInputPath)
#'
#' getPathwayFromAnalyte(RaMPInput)
#'}
#' @export
createRaMPInput <- function ( dataFrame = NULL, filePath = NULL, db = RaMP(), ... ) {

  #Produce dataframe of user data from csv path if not provided initially. User must specify exactly one of these.
  if (!is.null(filePath) & is.null(dataFrame)) {
    if (substr(filePath, (nchar(filePath) - 3), nchar(filePath)) == '.csv') {
      dataFrame <- readr::read_csv(filePath, ...)
    }
    else if (substr(filePath, (nchar(filePath) - 4), nchar(filePath)) == '.xlsx') {
      dataFrame <- readxl::read_excel(filePath, ...)
    }

  }

  else if (!is.null(filePath) & !is.null(dataFrame)) {
    stop('You have entered values for both the filePath and dataFrame parameters. Please only specify one of these.')
  }

  else if (is.null(filePath) & is.null(dataFrame)) {
    stop('You have not entered any input data to convert. Please specify either the filePath or dataFrame parameter.')
  }

  dataFrame <- tibble::tibble(dataFrame)

  #Get acceptable prefixes from RaMP and convert them to character vector.
  prefix_c <- strsplit(getPrefixesFromAnalytes(db = db, analyteType = 'metabolite')$idTypes[1], ',')[[1]]

  prefix_c <- append(prefix_c, strsplit(getPrefixesFromAnalytes(db = db, analyteType = 'gene')$idTypes[1], ',')[[1]])

  prefix_c <- gsub(' ', '', prefix_c)

  prefix_c <- unique(prefix_c)

  #Map from COMETS column names to RaMP prefixes
  COMETS_convert_df <- data.frame(COMETS = c('HMDB_ID', 'PUBCHEM'), RaMP = c('hmdb', 'pubchem'))

  for (i in colnames(dataFrame)) {

    if (i %in% COMETS_convert_df$COMETS) {
      colnames_c <- colnames(dataFrame)
      colnames_c[which(colnames_c == i)] <- COMETS_convert_df$RaMP[[which(COMETS_convert_df$COMETS == i)]]
      colnames(dataFrame) <- colnames_c
    }
  }


  #Parse dataFrame into RaMP prefix:ID input
  final_id_c <- c()

  for (i in colnames(dataFrame)) {

    if (i %in% prefix_c) {
      ids <- dplyr::pull(dataFrame[,i])
      ids <- as.character(ids)
      ids <- ids[ids != 'NA']
      ids <- ids[!is.na(ids)]
      ids <- unlist(strsplit(ids, split = ';'))

      if (length(ids) > 0) {
        final_id_c <- append(final_id_c, paste0(i, ':', ids))
      }
    }

    else {
      warning('Column "', i, '" not processed because identifier name not recognized. Valid names are: ', paste(prefix_c, collapse = ', '))
    }
  }

  return(final_id_c)
}