tinyseq.rb

require 'open-uri'
require 'rubygems'
require 'hpricot'

ARGV.each do |gi|

  # Fetch the URL and pass the contents to Hpricot
  url = "http://www.ncbi.nlm.nih.gov/entrez/eutils/efetch.fcgi?db=protein&id=#{gi}&rettype=fasta&retmode=xml"
  doc = Hpricot.XML(open(url).read)
 
  # Use Hpricot to naviagte to the necessary position in the document
  # Select only the sub nodes that are elements, as carriage returns are treated as content
  # but are not what we are interested in
  elements = doc.search("TSeqSet/TSeq/*").select{|x| x.instance_of? Hpricot::Elem}

  # Iterate through each element, and pass to the injected hash map
  # Inject is a bit of Ruby magic that allows an enum to be traversed
  # but also includes the method argument in the scope of the loop
  map = elements.inject(Hash.new) do |hash,elem|

    # Split each element into a key value pair and add to the hash
    # Use gsub to remove the 'TSeq_' at the start of each key, then downcase and convert to a symbol
    hash[elem.name.gsub('TSeq_','').downcase.to_sym] = elem.inner_text
   
    # Updated object needs to be returned at the end on an inject loop
    hash
  end

  # Output the arguments to the command line
  # NOTE: This is the only part of the code that is dependent on the names of the tinyseq values
  puts "(#{map[:gi]}|#{map[:accver]}) \"#{map[:defline]}\" size:#{map[:length]} #{map[:orgname]}"
end