Class: EmbedHtml::Embeder

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

Constant Summary collapse

MAX_CONCURRENCY = 
5

Instance Attribute Summary collapse

Instance Method Summary collapse

Constructor Details

#initialize(url_or_html, logger = Logger.new($stdout), concurrency = MAX_CONCURRENCY) ⇒ Embeder

Returns a new instance of Embeder.



18
19
20
21
22
 
# File 'lib/embed_html/embeder.rb', line 18

def initialize(url_or_html, logger=Logger.new($stdout), concurrency=MAX_CONCURRENCY)
  @logger = logger
  @url_or_html = url_or_html
  @concurrency = concurrency
end

Instance Attribute Details

#base_dirname ⇒ Object

Returns the value of attribute base_dirname.



16
17
18
 
# File 'lib/embed_html/embeder.rb', line 16

def base_dirname
  @base_dirname
end

#concurrency ⇒ Object

Returns the value of attribute concurrency.



15
16
17
 
# File 'lib/embed_html/embeder.rb', line 15

def concurrency
  @concurrency
end

#logger ⇒ Object

Returns the value of attribute logger.



14
15
16
 
# File 'lib/embed_html/embeder.rb', line 14

def logger
  @logger
end

#url_or_html ⇒ Object

Returns the value of attribute url_or_html.



13
14
15
 
# File 'lib/embed_html/embeder.rb', line 13

def url_or_html
  @url_or_html
end

Instance Method Details

#process ⇒ 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
68
69
70
71
 
# File 'lib/embed_html/embeder.rb', line 24

def process
  # @logger.info "downloading url: #{@url_or_html}"
  html = (@url_or_html =~ /$http/) ? Typhoeus::Request.get(@url_or_html.to_s).body : @url_or_html
  doc = Hpricot(html)
  
  hydra = Typhoeus::Hydra.new(:max_concurrency => @concurrency)
  doc.search("//img").each do |img|                
    begin
      if img['src']=~ /^http/
        hydra.queue create_fetch_file_request(img, 'src')
      else
        fetch_file(img, 'src')
      end
    rescue StandardError => e
      @logger.error "failed download image: #{img['src']} #{e.inspect}"
    end
  end

  doc.search("//script").each do |script|                
    begin
      if script['src'] and script['src'] =~ /^http/
        hydra.queue create_fetch_file_request(script, 'src')
      elsif script['src']
        fetch_file(script, 'src')
      end
    rescue StandardError => e
      @logger.error "failed download script: #{script['src']} #{e.inspect}"
    end
  end

  doc.search("//link").each do |link|
    begin
      url = link['href']
      if url =~ /^http/
        hydra.queue create_fetch_file_request(link, 'href')
      else
        fetch_file(link, 'href')
      end
    rescue StandardError => e
      @logger.error "failed download linked resource: #{link['href']} #{e.inspect}"
    end
  end
  
  hydra.run

  # @logger.info "done"            
  doc.to_html      
end

#process_local ⇒ Object 



73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
 
# File 'lib/embed_html/embeder.rb', line 73

def process_local
  # @logger.info "downloading url: #{@url_or_html}"
  html = open(@url_or_html).read
  doc = Hpricot(html)

  doc.search("//img").each do |img|                
    begin
      fetch_file(img, 'src')
    rescue StandardError => e
      @logger.error "failed download image: #{img['src']} #{e.inspect}"
    end
  end

  doc.search("//script").each do |script|                
    begin
      fetch_file(script, 'src')
    rescue StandardError => e
      @logger.error "failed download script: #{script['src']} #{e.inspect}"
    end
  end

  doc.search("//link").each do |link|
    begin
      fetch_file(link, 'href')
    rescue StandardError => e
      @logger.error "failed download linked resource: #{link['href']} #{e.inspect}"
    end
  end

  # @logger.info "done"            
  doc.to_html      
end