Class: BillboardScraper

Inherits:
Object
  • Object
show all
Defined in:
lib/top_100/billboard_scraper.rb

Class Method Summary collapse

Class Method Details

.scrape_from_artist_bio_page(url) ⇒ Object 



25
26
27
28
29
30
31
32
33
34
35
36
37
38
 
# File 'lib/top_100/billboard_scraper.rb', line 25

def self.scrape_from_artist_bio_page(url)
  bio = Nokogiri::HTML(open(url))
  location_date = bio.css('dl.facts > dd').text
  info = bio.css('article.bio_content').text
  artist = {
    name: bio.css('h1.title').text,
    # missing information for some artists, need to catch that by checking if nokogiri returned empty strings.
    location: location_date.empty? ? "Not Specified" : location_date.split("  ")[0].strip,
    date: location_date.empty? ? "Not Specified" : location_date.match(/\d+/)[0],
    #issues with .empty? for bio due to return  values of long whitespace strings, so using regex to filter those out instead.
    # calling split on info to separate bio aside and main text which does not have a different grouping outside 'article.bio_content'
    bio: info.match(/\A\s*\z/) ? "Not Specified" : info.split("      ")[-1].strip
  }
end

.scrape_from_chart_page ⇒ Object

will only use class methods, no need to create instances as there’s nothing unique between these scrapers and we only need one.



4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
 
# File 'lib/top_100/billboard_scraper.rb', line 4

def self.scrape_from_chart_page
  billboard_page = Nokogiri::HTML(open('http://www.billboard.com/charts/hot-100'))
  rss = RSS::Parser.parse(open('http://www.billboard.com/rss/charts/hot-100'))
  rss.items.each_with_index do |song, index|
    #account for any song titles that might happen to have ': ' in their title.
    name = song.title.split(": ")[1..-1].join(": ")
    rank = song.title.split(": ")[0]
    song_hash = {
      rank: rank,
      name: name,
      #ISSUE: RSS Parser feed doesn't seem to offer an artist value, requiring us to extract artist name from description instead. Alternatively, Nokogiri::XML can also open up the RSS information.
      # Regardless, the HTML page has to be opened for the artist_url information, and as a result, the gem is suspectible to breaking due to front-end design changes.
      #ISSUE: Collaborators saved as part of the artist name, which causes some difficulties with the search term for looking up the artist biography.
      # One possible solution would be to compare artist_name and the artist_url slug value in order to determine where the artist name begins and ends and then save that value to song instead.
      artist_name: song.description.split("#{name} by ")[1].split(" ranks ##{rank}")[0],
      artist_url: billboard_page.css('a.chart-row__artist')[index].attribute('href').value + '/biography',
    }
    Song.new(song_hash)
  end
end