A ruby script that uses nokogiri to scrape wikipedia for all the airport IATA codes, with airport names and locations too

airport_info.rb

require 'nokogiri'
require 'open-uri'

# Get all the airports from wikipedia!
def get_airport_info_from_wikipedia()
    airport_info = []
    "ABCDEFGHIJKLMNOPQRSTUVWXYZ".split("").each do |letter|
        doc = Nokogiri::HTML(open("https://en.wikipedia.org/wiki/List_of_airports_by_IATA_code:_#{letter}"))
        iata_table = doc.xpath("//table")[0]
        rows = iata_table.xpath("//tr")
        rows.each do |row|
            if not row.at_xpath('td[1]/text()').to_s.strip.empty?
                airport = {}
                [
                    [:iata, 'td[1]/text()'],
                    [:icao, 'td[2]/text()'],
                    [:name, 'td[3]/a/text()'],
                ].each do |name, xpath|
                    airport[name] = row.at_xpath(xpath).to_s.strip
                end

                airport[:location] = (row.xpath('td[4]/a/text()').zip(row.xpath('td[4]/text()')).flatten.compact.map { |e| e.to_s.strip }).join(' ')
                airport_info.push airport
            end
        end
    end
    return airport_info
end