A rough script to get the Two Oceans 2024 half marathon results as a CSV

two-oceans-csv.rb

#!/usr/bin/env ruby

# frozen_string_literal: true

require 'bundler/inline'

gemfile do
  source "https://rubygems.org"

  gem "csv"
  gem "http"
  gem "nokolexbor"
end

base_path = "https://results.finishtime.co.za"
pages =

begin
  # create HTTP client with persistent connection:
  http = HTTP.persistent(base_path)

  CSV.open("results.csv", "wb") do |csv|
    page_number = 0

    # Headers
    csv << [
      "Pos",
      "Race No",
      "Country",
      "Name",
      "Time",
      "Medal",
      "Club",
      "Gender",
      "Gen Pos",
      "Category",
      "Cat Pos",
      "Start",
      "5km",
      "10.55km",
      "15km"
    ]

    loop do
      page_number += 1

      puts "Fetching page #{page_number}..."
      # issue multiple requests using same connection:
      page = http.get("/results.aspx?CId=35&RId=30353&EId=2&dt=0&PageNo=#{page_number}")

      doc = Nokolexbor::HTML(page.to_s)

      unless doc.css("table tr").any?
        break
      end

      doc.css("#ctl00_Content_Main_divGrid table tr").each_with_index do |row, index|
        # Skip header row
        next if index < 1

        row_contents = row.css('td').map do |cell|
          if cell.css('img').any?
            src = cell.css('img').first['src']

            matches = src.match(/\Aimages\/flags\/\d+\/([a-zA-Z]*)\.png\z/i)
            if matches
              matches[1]
            else
              ""
            end
          else
            cell.text
          end
        end

        csv << row_contents.values_at(2, 4, 5, 6, 8..18)
      end
    end
  end
ensure
  # close underlying connection when you don't need it anymore
  http.close if http
end