MIRIAM XML to CSV

miriam-xml2csv.rb

#!/usr/bin/env ruby
#
# Convert MIRIAM Registry XML file (http://www.ebi.ac.uk/miriam/main/export/) to CSV
# for browsing at TogoDB (http://semantic.togodb.dbcls.jp/togodb/view/miriam)
#
# Copyright (C) 2012 Toshiaki Katayama <k@bioruby.org>
#
# Pre requirements:
#  % curl http://www.ebi.ac.uk/miriam/main/export/xml/ > resources_all.xml
#  % gem install nokogiri
#  % miriam-xml2csv.rb resources_all.xml > resources_all.csv
#

require 'rubygems'
require 'nokogiri'
require 'csv'

xml = Nokogiri::XML(ARGF)

ns = xml.namespaces

header = %w(
  MIRIAM
  Namespace
  Pattern
  Name
  Synonyms
  Definition
  URNs
  URIs
  Tags
  SBML
  PMIDs
  URLs
  Resource
  State
  Reliability
  URL
  Link
  Example
  Info
  Institution
  Location
)

puts CSV.generate_line(header)


xml.xpath('//xmlns:datatype', ns).each do |datatype|
  next if datatype["obsolete"] == "true"

  path = './/xmlns:resources/xmlns:resource'
  datatype.xpath(path, ns).each do |res|
    res_id = res["id"]
    res_state = res["state"]
    res_reliability = res["reliability"]
    res_url = res.at('dataResource').content
    res_link = res.at('dataEntry').content
    res_example = res.at('dataEntityExample').content
    res_info = res.at('dataInfo').content
    res_institution = res.at('dataInstitution').content
    res_location = res.at('dataLocation').content

    path = './/xmlns:synonyms/xmlns:synonym'
    synonyms = datatype.xpath(path, ns).collect(&:text)

    path = './/xmlns:uris/xmlns:uri[@type="URN"]'
    path2 = './/xmlns:uris/xmlns:uri[@deprecated="true"]'
    urns = datatype.xpath(path, ns).collect(&:text) - datatype.xpath(path2, ns).collect(&:text)

    path = './/xmlns:uris/xmlns:uri[@type="URL"]'
    path2 = './/xmlns:uris/xmlns:uri[@deprecated="true"]'
    uris = datatype.xpath(path, ns).collect(&:text) - datatype.xpath(path2, ns).collect(&:text)

    path = './/xmlns:tags/xmlns:tag'
    tags = datatype.xpath(path, ns).collect(&:text)

    path = './/xmlns:annotation/xmlns:format[@name="SBML"]/xmlns:elements/xmlns:element'
    sbml = datatype.xpath(path, ns).collect(&:text)

    path = './/xmlns:documentations/xmlns:documentation[@type="PMID"]'
    pmids = datatype.xpath(path, ns).collect(&:text).map {|x| x[/\d+/]}

    path = './/xmlns:documentations/xmlns:documentation[@type="URL"]'
    urls = datatype.xpath(path, ns).collect(&:text)

    array = [
      datatype["id"],
      datatype.at('namespace').content,
      datatype["pattern"],
      datatype.at('name').content,
      synonyms.join("; "),
      datatype.at('definition').content,
      urns.join("\n"),
      uris.join("\n"),
      tags.join("; "),
      sbml.join("; "),
      pmids.join("\n"),
      urls.join("\n"),
      res_id,
      res_state,
      res_reliability,
      res_url,
      res_link,
      res_example,
      res_info,
      res_institution,
      res_location,
    ]

    puts CSV.generate_line(array)
  end
end