z4.rb

require 'rubygems'
require 'parslet'

class Address < Parslet::Parser
  # Single character rules
  rule(:comma)      { str(',') >> space? }
  rule(:space)      { match('\s').repeat }
  rule(:space?)     { space.maybe }
  rule(:newline)    { str("\n").repeat }
  rule(:digit)      { match('[0-9]') }

  # Things
  rule(:line1)      { (match('[A-z0-9 ]').repeat).as(:address1) }
  rule(:line2)      { newline >> (match('[A-z0-9 ]').repeat).as(:address2) }
  rule(:word)       { match('[A-z]').repeat }
  rule(:capword)    { space >> match['A-Z'] >> match['a-z'].repeat(1) }

  # Address Parts
  rule(:address)    { line1 >> line2 }
  rule(:state)      { (str("CA") | str("CO")).as(:state) }
  rule(:zip)        { digit.repeat(5,5).as(:zip) }
  rule(:city)       { (word >> capword.maybe).as(:city) }
  rule(:sz)         { state >> space.maybe >> zip }
  rule(:csz)        { city >> space >> state >> space >> zip.maybe }
  rule(:all)        { (line1 >> newline >> csz) | (address >> newline >> csz) }

  root :all
end

def run(rule, string)
  begin
    parser = Address.new
    eval "p parser.#{rule}.parse('#{string}')"
  rescue Parslet::ParseFailed => e
    puts e, parser.root.error_tree
  end
end

run(  "comma", ",")
run(  "comma", "x")
run(  "digit", "5")
run(  "comma", ",")
run(  "digit", "5")
run(    "zip", "94022")
run(  "state", "CA")
run(  "state", "CO")
run(     "sz", "CA 94022")
run(    "csz", " CA 94022")
run(    "csz", "Mountain CA 94022")
run(    "csz", "Mountain View CA 94022")
run(    "csz", "SF CA 94022")
run(    "csz", "SF CA")
run("address", "1523 Broker Way")
run("address", "1523 Broker Way\nApartment 22")
run(    "all", "1523 Broker Way\nMountain View CA 94022")
run(    "all", "1523 Broker Way\nApartment 22\nMountain View CA 94022")