Class: Hongkong::News::Scrapers::AppleDailyScraper

Inherits:
Object
  • Object
show all
Includes:
PhantomScraper
Defined in:
lib/hongkong/news/scrapers/apple_daily_scraper.rb

Instance Method Summary collapse

Methods included from PhantomScraper

#html, #screenshot_data

Instance Method Details

#news(url) ⇒ Object

Extract article from page from Apple Daily



22
23
24
25
26
27
28
29
30
31
32
33
34
 
# File 'lib/hongkong/news/scrapers/apple_daily_scraper.rb', line 22

def news(url)
  visit url

  document = Document.new
  document.source = 'appledaily'
  document.title = first("#articleContent h1").text.strip
  document.url = url
  document.html = html
  document.content = page.evaluate_script("HongKongNews.getInnerText('#masterContent')")
  document.screenshot_data = screenshot_data

  document
end

#news_links ⇒ Object

Extract all news links from Apple Daily



10
11
12
13
14
15
16
17
18
19
 
# File 'lib/hongkong/news/scrapers/apple_daily_scraper.rb', line 10

def news_links
  visit "http://hk.apple.nextmedia.com/"

  all("#article_ddl option").collect do |option|
    link = Link.new
    link.title = option.text
    link.url = option["value"]
    link
  end
end