Playing with getting source from URL, to get gene names from refseq-ids and converting them into R if-statements. Mostly an experiment in using Scalas parallel collections. NOTE: This should under no circumstances be used in production, then you need to be using some API instead. You have been warned.

PlayWithURLs.scala

object PlayWithURLs extends App {

    // Some ids for the BRCA1 gene
    val refseqids = List("uc010whl.1", "uc002icp.3", "uc010whp.1")

    def queryRefSeqId(refSeqId: String): String = {
        def read(url: String): String = io.Source.fromURL(url).mkString
        val document = read("http://genome.ucsc.edu/cgi-bin/hgGene?hgg_gene=" + refSeqId + "&org=human")

        val regexp = """.*Genetic Association Database:\s\<A HREF=.*\>(\w+)\<.*""".r

        regexp.findFirstMatchIn(document).get.group(1)
    }

    def convertToRCode(refSeqId: String, geneName: String): String = {
        """if(refseqid == """" + refSeqId + """") return("""" + geneName + " " + "(" + refSeqId + ")" + """")"""
    }

    val startTime = System.nanoTime()
    refseqids.par.foreach(id => println(convertToRCode(id, queryRefSeqId(id))))
    val endTime = System.nanoTime()
    println("time: " + (endTime - startTime) / 1000000000.0)
}