view ChebiCompound.R @ 0:e66bb061af06 draft

planemo upload for repository https://github.com/workflow4metabolomics/lcmsmatching.git commit 3529b25417f8e1a5836474c9adec4b696d35099d-dirty

if ( ! exists('ChebiCompound')) { # Do not load again if already loaded

	source('BiodbEntry.R')
	
	#####################
	# CLASS DECLARATION #
	#####################
	
	ChebiCompound <- setRefClass("ChebiCompound", contains = "BiodbEntry")
	
	###########
	# FACTORY #
	###########
	
	createChebiCompoundFromHtml <- function(contents, drop = TRUE) {

		library(XML)

		compounds <- list()

		# Define xpath expressions
		xpath.expr <- character()
#		xpath.expr[[RBIODB.ACCESSION]] <- "//b[starts-with(., 'CHEBI:')]"
		xpath.expr[[RBIODB.INCHI]] <- "//td[starts-with(., 'InChI=')]"
		xpath.expr[[RBIODB.INCHIKEY]] <- "//td[text()='InChIKey']/../td[2]"

		for (html in contents) {

			# Create instance
			compound <- ChebiCompound$new()
		
			# Parse HTML
			xml <-  htmlTreeParse(html, asText = TRUE, useInternalNodes = TRUE)

			# Test generic xpath expressions
			for (field in names(xpath.expr)) {
				v <- xpathSApply(xml, xpath.expr[[field]], xmlValue)
				if (length(v) > 0)
					compound$setField(field, v)
			}
		
			# Get accession
			accession <- xpathSApply(xml, "//b[starts-with(., 'CHEBI:')]", xmlValue)
			if (length(accession) > 0) {
				accession <- sub('^CHEBI:([0-9]+)$', '\\1', accession, perl = TRUE)
				compound$setField(RBIODB.ACCESSION, accession)
			}

			compounds <- c(compounds, compound)
		}

		# Replace elements with no accession id by NULL
		compounds <- lapply(compounds, function(x) if (is.na(x$getField(RBIODB.ACCESSION))) NULL else x)

		# If the input was a single element, then output a single object
		if (drop && length(contents) == 1)
			compounds <- compounds[[1]]
	
		return(compounds)
	}
}