Class: EmailCrawler::Runner

Inherits:
Object
  • Object
show all
Defined in:
lib/email_crawler.rb

Instance Method Summary collapse

Constructor Details

#initialize(google_website) ⇒ Runner 



14
15
16
17
18
19
20
21
22
 
# File 'lib/email_crawler.rb', line 14

def initialize(google_website)
  @google_website = google_website

  log_file = File.join(ENV["HOME"], "email-crawler.log")
  file = File.open(log_file, File::WRONLY | File::APPEND | File::CREAT)
  @logger = ::Logger.new(file).tap do |logger|
    logger.level = ENV["DEBUG"] ? Logger::INFO : Logger::ERROR
  end
end

Instance Method Details

#run(q, max_links = PageLinks::MAX_LINKS) ⇒ Object 



24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
 
# File 'lib/email_crawler.rb', line 24

def run(q, max_links = PageLinks::MAX_LINKS)
  urls = Scraper.new(@google_website).top_ten_urls_for(q)
  urls.each { |url, links| @logger.info "#{url}" }

  threads = (1..urls.length).map do |i|
    Thread.new(i, urls[i-1]) do |i, url|
      @logger.info "[Thread ##{i}] grabbing page links for '#{url}'.."
      Thread.current[:url] = url
      Thread.current[:links] = PageLinks.for(url, max_links)
    end
  end

  threads.each(&:join)
  threads.each { |thread| @logger.info "#{thread[:url]} (#{thread[:links].length} links)" }
  links_by_url = Hash[threads.map { |thread| [thread[:url], thread[:links]] }]

  threads = (links_by_url).map.with_index do |arr, i|
    Thread.new(i+1, arr.first, arr.last) do |i, url, links|
      @logger.info "[Thread ##{i}] scanning for emails on page '#{url}' (#{links.length} links)"
      Thread.current[:url] = url
      Thread.current[:emails] = EmailScanner.new(url).scan(links)
    end
  end

  threads.each(&:join)

  read_emails = Set.new
  CSV.generate do |csv|
    csv << %w(Email Domain URL)
    csv << []

    threads.each do |thread|
      email_count = thread[:emails].inject(0) { |sum, arr| sum += arr.last.length }
      @logger.info "#{thread[:url]} (#{email_count} emails)"

      url = thread[:url]
      thread[:emails].each do |link, emails|
        emails.each do |email|
          csv << [email, url, link] if read_emails.add?(email)
        end
      end
    end
  end
end