Class: Shnork::Shnork

Inherits:
Object
  • Object
show all
Includes:
Rack::Utils
Defined in:
lib/shnork.rb

Instance Attribute Summary collapse

Instance Method Summary collapse

Constructor Details

#initialize ⇒ Shnork

Returns a new instance of Shnork.



24
25
26
27
28
29
30
31
32
33
34
35
36
37
 
# File 'lib/shnork.rb', line 24

def initialize
  set :env, :production
  set :run, false
  @app = Sinatra::application
  @request = Rack::MockRequest.new(Sinatra.build_application)
  @reached = []
  @found   = []
  
  @internal_depth = {}
  @internal_pred  = {}
  @depth   = :all
  
  @results = ShnorkResults.new
end

Instance Attribute Details

#depth ⇒ Object

Returns the value of attribute depth.



22
23
24
 
# File 'lib/shnork.rb', line 22

def depth
  @depth
end

#page ⇒ Object (readonly)

Returns the value of attribute page.



21
22
23
 
# File 'lib/shnork.rb', line 21

def page
  @page
end

#reached ⇒ Object (readonly)

Returns the value of attribute reached.



21
22
23
 
# File 'lib/shnork.rb', line 21

def reached
  @reached
end

Instance Method Details

#analyze(page) ⇒ Object 



39
40
41
42
43
44
45
46
47
48
 
# File 'lib/shnork.rb', line 39

def analyze page
  @found << [page, nil]
  while @found.size != 0
    link = @found.shift
    next if @reached.include? link.first
    retrieve link
  end
  
  return @results
end

#less_than_max_depth(d) ⇒ Object 



54
55
56
57
 
# File 'lib/shnork.rb', line 54

def less_than_max_depth d
  return true if @depth == :all 
  d <= @depth
end

#links(current_page) ⇒ Object 



72
73
74
75
 
# File 'lib/shnork.rb', line 72

def links current_page
  return [] unless @page
  Nokogiri::HTML(@page.body).xpath('//a').select { |link| link.attributes["href"] =~ /^\// }.map { |link| [link.attributes["href"], current_page] }
end

#record(page) ⇒ Object 



50
51
52
 
# File 'lib/shnork.rb', line 50

def record page
  @reached = @reached | [page]
end

#retrieve(page_from) ⇒ Object 



59
60
61
62
63
64
65
66
67
68
69
70
 
# File 'lib/shnork.rb', line 59

def retrieve page_from
  page = page_from.first
  from = page_from.last
  @internal_depth[page] = from ? @internal_depth[from] + 1 : 0
  @internal_pred[page]  = from
  return unless less_than_max_depth @internal_depth[page]
    @page = @request.get(page)
    @page = @request.get(@page.headers["Location"]) if @page.status >= 300 and @page.status < 400
  @results.add @page.status, page
  @found = @found | links(page)
  record page
end