view MassFiledbConn.R @ 3:f61ce21ed17c draft

planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit 476a081c0da66822f4e77070f5ce59d9f14511f4-dirty

# LCMS File db.
# In this type of database, a single file is provided in CSV format. Default separator is tabulation.
# Each line is a MS peak measure, .
# The file contains molecule and spectrum information. Each spectrum has an accession id.

# TODO Rename setField into setFieldName + addNewField, and setMsMode into setMsModeValue

#############
# CONSTANTS #
#############

# Default database fields
.BIODB.DFT.DB.FIELDS <- list()
for (f in c(BIODB.ACCESSION, BIODB.NAME, BIODB.FULLNAMES, BIODB.COMPOUND.ID, BIODB.MSMODE, BIODB.PEAK.MZEXP, BIODB.PEAK.MZTHEO, BIODB.PEAK.COMP, BIODB.PEAK.ATTR, BIODB.CHROM.COL, BIODB.CHROM.COL.RT, BIODB.FORMULA, BIODB.MASS))
	.BIODB.DFT.DB.FIELDS[[f]] <- f

#####################
# CLASS DECLARATION #
#####################

MassFiledbConn <- methods::setRefClass("MassFiledbConn", contains = "MassdbConn", fields = list(.file = "character", .file.sep = "character", .file.quote = "character", .field.multval.sep = 'character', .db = "ANY", .db.orig.colnames = "character", .fields = "list", .ms.modes = "character"))

###############
# CONSTRUCTOR #
###############

MassFiledbConn$methods( initialize = function(file = NA_character_, file.sep = "\t", file.quote = "\"", ...) {

	# Check file
	(! is.null(file) && ! is.na(file)) || stop("You must specify a file database to load.")
	file.exists(file) || stop(paste0("Cannot locate the file database \"", file ,"\"."))

	# Set fields
	.db <<- NULL
	.db.orig.colnames <<- NA_character_
	.file <<- file
	.file.sep <<- file.sep
	.file.quote <<- file.quote
	.fields <<- .BIODB.DFT.DB.FIELDS
	.field.multval.sep <<- ';'
	.ms.modes <<- c(BIODB.MSMODE.NEG, BIODB.MSMODE.POS)
	names(.self$.ms.modes) <- .self$.ms.modes

	callSuper(...)
})

######################
# Is valid field tag #
######################

MassFiledbConn$methods( isValidFieldTag = function(tag) {
	return (tag %in% names(.self$.fields))
})

###########
# INIT DB #
###########

MassFiledbConn$methods( .init.db = function() {

	if (is.null(.self$.db)) {

		# Load database
		.db <<- read.table(.self$.file, sep = .self$.file.sep, .self$.file.quote, header = TRUE, stringsAsFactors = FALSE, row.names = NULL)

		# Save column names
		.db.orig.colnames <<- colnames(.self$.db)
	}
})

#############
# Set field #
#############

MassFiledbConn$methods( setField = function(tag, colname) {

	( ! is.null(tag) && ! is.na(tag)) || stop("No tag specified.")
	( ! is.null(colname) && ! is.na(colname)) || stop("No column name specified.")

	# Load database file
	.self$.init.db()

	# Check that this field tag is defined in the fields list
	.self$isValidFieldTag(tag) || stop(paste0("Database field tag \"", tag, "\" is not valid."))

	# Check that columns are defined in database file
	all(colname %in% names(.self$.db)) || stop(paste0("One or more columns among ", paste(colname, collapse = ", "), " are not defined in database file."))

	# Set new definition
	if (length(colname) == 1)
		.fields[[tag]] <<- colname
	else {
		new.col <- paste(colname, collapse = ".")
		.self$.db[[new.col]] <- vapply(seq(nrow(.self$.db)), function(i) { paste(.self$.db[i, colname], collapse = '.') }, FUN.VALUE = '')
		.fields[[tag]] <<- new.col
	}

	# Update data frame column names
	colnames(.self$.db) <- vapply(.self$.db.orig.colnames, function(c) if (c %in% .self$.fields) names(.self$.fields)[.self$.fields %in% c] else c, FUN.VALUE = '')
})

######################################
# SET FIELD MULTIPLE VALUE SEPARATOR #
######################################

MassFiledbConn$methods( setFieldMultValSep = function(sep) {
	.field.multval.sep <<- sep
})

################
# SET MS MODES #
################

MassFiledbConn$methods( setMsMode = function(mode, value) {
	.self$.ms.modes[[mode]] <- value
})

##########################
# GET ENTRY CONTENT TYPE #
##########################

MassFiledbConn$methods( getEntryContentType = function(type) {
	return(BIODB.DATAFRAME)
})

################
# CHECK FIELDS #
################

MassFiledbConn$methods( .check.fields = function(fields) {

	if (length(fields) ==0 || (length(fields) == 1 && is.na(fields)))
		return

	# Check if fields are known
	unknown.fields <- names(.self$.fields)[ ! fields %in% names(.self$.fields)]
	if (length(unknown.fields) > 0)
		stop(paste0("Field(s) ", paste(fields, collapse = ", "), " is/are unknown."))

	# Init db
	.self$.init.db()

	# Check if fields are defined in file database
	undefined.fields <- colnames(.self$.db)[ ! fields %in% colnames(.self$.db)]
	if (length(undefined.fields) > 0)
		stop(paste0("Column(s) ", paste(fields), collapse = ", "), " is/are undefined in file database.")
})

##########
# SELECT #
##########

# Select data from database
MassFiledbConn$methods( .select = function(cols = NULL, mode = NULL, compound.ids = NULL, drop = FALSE, uniq = FALSE, sort = FALSE, max.rows = NA_integer_) {

	x <- NULL

	# Init db
	.self$.init.db()

	# Get db
	db <- .self$.db

	# Filter db on mode
	if ( ! is.null(mode) && ! is.na(mode)) {

		# Check mode value
		mode %in% names(.self$.ms.modes) || stop(paste0("Unknown mode value '", mode, "'."))
		.self$.check.fields(BIODB.MSMODE)

		# Filter on mode
		db <- db[db[[unlist(.self$.fields[BIODB.MSMODE])]] %in% .self$.ms.modes[[mode]], ]
	}

	# Filter db on compound ids
	# TODO

	if ( ! is.null(cols) && ! is.na(cols))
		.self$.check.fields(cols)

	# Get subset
	if (is.null(cols) || is.na(cols))
		x <- db
	else
		x <- db[, unlist(.self$.fields[cols]), drop = drop]

	# Rearrange
	if (drop && is.vector(x)) {
		if (uniq)
			x <- x[ ! duplicated(x)]
		if (sort)
			x <- sort(x)
	}

	# Cut
	if ( ! is.na(max.rows))
		x <- if (is.vector(x)) x[1:max.rows] else x[1:max.rows, ]

	return(x)
})

#################
# GET ENTRY IDS #
#################

MassFiledbConn$methods( getEntryIds = function(type) {

	ids <- NA_character_

	if (type %in% c(BIODB.SPECTRUM, BIODB.COMPOUND))
		ids <- as.character(.self$.select(cols = if (type == BIODB.SPECTRUM) BIODB.ACCESSION else BIODB.COMPOUND.ID, drop = TRUE, uniq = TRUE, sort = TRUE))

	return(ids)
})

##################
# GET NB ENTRIES #
##################

MassFiledbConn$methods( getNbEntries = function(type) {
	return(length(.self$getEntryIds(type)))
})

###############################
# GET CHROMATOGRAPHIC COLUMNS #
###############################

# Inherited from MassdbConn.
MassFiledbConn$methods( getChromCol = function(compound.ids = NULL) {

	# Extract needed columns
	db <- .self$.select(cols = c(BIODB.COMPOUND.ID, BIODB.CHROM.COL))

	# Filter on molecule IDs
	if ( ! is.null(compound.ids))
		db <- db[db[[BIODB.COMPOUND.ID]] %in% compound.ids, ]

	# Get column names
	cols <- db[[BIODB.CHROM.COL]]

	# Remove duplicates
	cols <- cols[ ! duplicated(cols)]

	# Make data frame
	chrom.cols <- data.frame(cols, cols, stringsAsFactors = FALSE)
	colnames(chrom.cols) <- c(BIODB.ID, BIODB.TITLE)

	return(chrom.cols)
})

#################
# GET MZ VALUES #
#################

# Inherited from MassdbConn.
MassFiledbConn$methods( getMzValues = function(mode = NULL, max.results = NA_integer_) {

	# Get mz values
	mz <- .self$.select(cols = BIODB.PEAK.MZ, mode = mode, drop = TRUE, uniq = TRUE, sort = TRUE, max.rows = max.results)

	return(mz)
})

################
# GET NB PEAKS #
################

# Inherited from MassdbConn.
MassFiledbConn$methods( getNbPeaks = function(mode = NULL, compound.ids = NULL) {

	# Get peaks
	peaks <- .self$.select(cols = BIODB.PEAK.MZTHEO, mode = mode, compound.ids = compound.ids)

	return(length(peaks))
})