Scrapes an HTML string for image URLs using Oga + regex.

html_image_url_scraper.rb

require 'oga'

# Scrapes an HTML string for image URLs using Oga + regex.
class HtmlImageUrlScraper

  EXTNAMES = ['.jpeg','.jpg','.gif','.png','.bmp','.svg','.tif','.tiff','.ai','.apng','.bpg','.cgm','.dxf','.eps','.flif','.hdp','.hdr','.heic','.heif','.ico','.iff','.jp2','.jpx','.jxr','.lbm','.pbm','.pgm','.pnm','.ppm','.wdp','.webp']

  attr_accessor :extnames, :regexes

  def initialize(extnames: EXTNAMES, regexes: nil)
    @extnames,@regexes=extnames,regexes

    unless !!@regexes
      extnames_str = extnames.map { |n| n.sub('.','') }.join('|')
      @regexes = [
        /"([^"]+\.(?:{{EXTNAMES}}))"/i, 
        /'([^']+\.(?:{{EXTNAMES}}))'/i
      ].map { |frmt| Regexp.new(frmt.to_s.gsub('{{EXTNAMES}}',extnames_str)) }
    end
  end

  # @return [Array<String>]
  def scrape(string)
    doc      = Oga::parse_html(string.force_encoding('UTF-8'))

    # ignores extnames; assumes any img src value is an image URL
    img_srcs = doc.css('img').map { |img| img.get('src') }.select {|src| !!src}
    
    # relies on extnames
    a_hrefs = doc.css('a').map { |a| a.get('href') }.select do |href|
      !!href && @extnames.include?(File.extname(href).downcase)
    end

    via_regex = @regexes.map { |regex| string.scan(regex) }.flatten

    return (img_srcs + a_hrefs + via_regex).uniq
  end
end