#!/usr/bin/env ruby
require_relative '../../lazar/lib/lazar.rb'

sdfs = []
results = []
read_result = false
@sdf = ""
File.readlines(ARGV[0]).each do |line|
  if line.match %r{\$\$\$\$}
    @sdf << line
    sdfs << @sdf
    @sdf = ""
  elsif line.match ">  <Ames test categorisation>"
    read_result = true
  else
      #p line
    if read_result
      if line.chomp == "mutagen"
        results << 1
      elsif line.chomp == "nonmutagen"
        results << 0
      else
        results << line.chomp
      end
      read_result = false
    else
      @sdf << line
    end
  end
end

obconversion = OpenBabel::OBConversion.new
obconversion.set_in_and_out_formats "sdf","can"
obmol = OpenBabel::OBMol.new

puts "SMILES,Activity"
sdfs.each_with_index do |sdf,i|
  obconversion.read_string obmol,sdf
  s = obconversion.write_string(obmol).split.first
  puts [s,results[i]].join ","
end