Class: Eatable::Scraper

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

Constant Summary collapse

NUTS = 
/nut|almond|cashew|macadamia|pecan|pigñolia|pistachio|praline|pesto|filbert/
SHELL_FISH = 
/shrimp|clam|mussel|lobster|crab|prawn/

Class Method Summary collapse

Class Method Details

.city_url ⇒ Object 



62
63
64
 
# File 'lib/eatable/scraper.rb', line 62

def self.city_url
  @@city_url
end

.filter_menus(menus_array) ⇒ Object 



41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
 
# File 'lib/eatable/scraper.rb', line 41

def self.filter_menus(menus_array)
  valid_restaurants = []

  menus_array[0..10].each do |menu|
    menupg = Nokogiri::HTML(open(menu))
    menu_body = (menupg.css('div #restaurant-menu').text.gsub(/\r\n/, "").gsub(/\u00A0/, "")).downcase
    # --------------- Adjust tolerence level for number of search terms -----------------------------
    if (menu_body.scan(NUTS).length <= 2) && (menu_body.scan(SHELL_FISH).length <= 2) && (menu_body.length > 20)
      restaurant = {
        "name" => menupg.css('div h1.title1respage').text,
        "address" => "#{menupg.css('span.addr').text}",
        "phone" => menupg.css('li.phonenew').text.gsub(/[\\rn\s]/, "")
        }

      valid_restaurants << restaurant        
    end
  end
  valid_restaurants
end

.full_neighborhood_url ⇒ Object 



74
75
76
 
# File 'lib/eatable/scraper.rb', line 74

def self.full_neighborhood_url
  @@full_neighborhood_url
end

.full_neighborhood_url=(neighborhood_url) ⇒ Object 



66
67
68
69
70
71
72
 
# File 'lib/eatable/scraper.rb', line 66

def self.full_neighborhood_url=(neighborhood_url)
  if @@city_url == "http://www.menupages.com"
    @@full_neighborhood_url = neighborhood_url
  else
    @@full_neighborhood_url = @@city_url + neighborhood_url
  end
end

.neighborhood_scrape(city_name) ⇒ Object 



11
12
13
14
15
16
17
18
19
20
21
22
 
# File 'lib/eatable/scraper.rb', line 11

def self.neighborhood_scrape(city_name)
  neighborhoods = {}
  @@city_url = 'http://' + city_name + '.menupages.com'
  citypg = Nokogiri::HTML(open(@@city_url))

  citypg.css('#image-map area').each do |a|
    neighborhood_url = a.attr('href')
    name = a.attr('alt')
    neighborhoods[name] = neighborhood_url
  end
  neighborhoods
end

.restaurant_scrape(neighborhood_url) ⇒ Object 



25
26
27
28
29
30
31
32
33
34
35
36
37
 
# File 'lib/eatable/scraper.rb', line 25

def self.restaurant_scrape(neighborhood_url)
  menus_array = []

  self.full_neighborhood_url=(neighborhood_url)
  neighb_pg = Nokogiri::HTML(open(self.full_neighborhood_url))
  rest_table = neighb_pg.css('tbody tr td a')
  rest_table.each do|listing| 
    menus_array << listing.attr('href') unless ((listing.attr('href')).include?('grubhub') || (listing.attr('href')).include?('seamless'))
  end
  menus_array.map! {|link| self.city_url + link + "menu"}

  self.filter_menus(menus_array)
end