bengaluru-food-census/parse_listing.rb

require 'nokogiri'
require 'csv'
require 'date'

restaurants = []

CSV.open("data/#{Date.today.to_s}.csv", 'wb') do |csv|
  csv << ['url', 'title', 'location', 'address', 'cuisine']
  Dir.glob('html/restaurants-*.html') do |file|
    page = Nokogiri::HTML(open(file))

    page.css('.plr10').each do |div|
      links = div.css('a')
      spans = div.css('span')


      title = links[0].text
      location = links[1].text
      address = spans[1].text
      cuisine = spans[0].text
      url = links[0]['href']

      csv << [url, title, location, address, cuisine]

      puts url
    end
  end
end