A rake task to scan Craigslist periodically and email any new listings that match a set of parameters.

craigslist.rake

require 'open-uri'

def in_bounds?(lat, lng)
  bounds = [
    {:north => 41.87849, :east => -87.623917, :south => 41.872041, :west => -87.633787}
  ]

  bounds.any?{|b| lat >= b[:south] && lat <= b[:north] && lng <= b[:east] && lng >= b[:west]}
end

def email(body, to = "[email protected]")
  Pony.mail({
    :to => to,
    :via => :smtp,
    :via_options => {
      :address => 'smtp.gmail.com',
      :port => '587',
      :enable_starttls_auto => true,
      :user_name => "[email protected]",
      :password => "password",
      :authentication => :plain,
      :domain => 'localhost'
    },
    :subject => 'Cragslist Watcher',
    :from => "[email protected]",
    :body => body
  })
end

def email_listings(good)
  body = good.map{|l| "#{l[:name]}\n#{l[:url]}"}.join("\n\n\n")
  email(body)
end

def email_error(error)
  email("#{error.message}\n\n#{error.backtrace.join("\n")}")
end

task :cl_watcher do
  begin
    query = "http://chicago.craigslist.org/search/apa?zoomToPosting=&altView=&query=loop&srchType=A&minAsk=1200&maxAsk=2000&addTwo=purrr&bedrooms="
    file = "/tmp/cl_watcher.txt"
    dupes = []

    `touch #{file}`

    File.open(file) do |f|
      while line = f.gets
        dupes << line.chop
      end
    end

    result = Nokogiri::HTML(open(query))
    good_listings = []

    result.css('p.row').each do |row|
      begin
        next unless row.attributes['data-longitude'] && row.attributes['data-latitude']
        lat = row.attributes["data-latitude"].value.to_f
        lng = row.attributes["data-longitude"].value.to_f
        link = row.css('a')[1]
        url = 'http://chicago.craigslist.org' + link.attributes["href"].value
        name = link.children.first.text

        if !dupes.include?(name) && in_bounds?(lat, lng)
          good_listings << {:name => name, :url => url}
          dupes << name
          File.open(file, 'a'){|f| f.puts name}
        end
      rescue Exception => e
        email_error(e)
      end
    end

    email_listings(good_listings) unless good_listings.empty?
  rescue Exception => e
    email_error(e)
  end
end