Class: Matryoshka::Document::Html

Inherits:
Object
  • Object
show all
Defined in:
lib/matryoshka/document/html.rb

Defined Under Namespace

Classes: Merge

Constant Summary collapse

EXTERNAL_SEARCH_PATTERN =
'[@rel="external"]'
PROXY_SEARCH_PATTERN =
'[@rel="proxy"]'

Instance Attribute Summary collapse

Class Method Summary collapse

Instance Method Summary collapse

Constructor Details

#initialize(rack_array) ⇒ Html

Returns a new instance of Html.



124
125
126
127
128
129
130
131
132
133
134
135
# File 'lib/matryoshka/document/html.rb', line 124

def initialize(rack_array)
  self.status  = rack_array[0]
  self.headers = rack_array[1]
  
  html = rack_array[2]
  if html.respond_to?(:to_html)
    self.content = html
  else
    self.body = html
  end
  load_externals unless externals_loaded?
end

Instance Attribute Details

#bodyObject

Returns the value of attribute body.



6
7
8
# File 'lib/matryoshka/document/html.rb', line 6

def body
  @body
end

#contentObject

Returns the value of attribute content.



6
7
8
# File 'lib/matryoshka/document/html.rb', line 6

def content
  @content
end

#headersObject

this needs to update headers like ‘Content-Length’ and remove Matryoshka-specific ones Might need to make this cleaned_headers or something



159
160
161
# File 'lib/matryoshka/document/html.rb', line 159

def headers
  @headers
end

#pathObject

Returns the value of attribute path.



6
7
8
# File 'lib/matryoshka/document/html.rb', line 6

def path
  @path
end

#statusObject

Returns the value of attribute status.



6
7
8
# File 'lib/matryoshka/document/html.rb', line 6

def status
  @status
end

Class Method Details

.acceptsObject



149
150
151
# File 'lib/matryoshka/document/html.rb', line 149

def self.accepts
  ['text/html']
end

.extentionsObject



145
146
147
# File 'lib/matryoshka/document/html.rb', line 145

def self.extentions
  ['.html','.htm']
end

Instance Method Details



117
118
119
120
121
122
# File 'lib/matryoshka/document/html.rb', line 117

def absolute_link(elem,attribute,base_dir)
  relative_uri = elem.attributes[attribute]
  unless (relative_uri.include?(base_dir) or is_absolute?(relative_uri))
    elem.set_attribute(attribute, (base_dir + relative_uri))
  end    
end

#call(env) ⇒ Object



166
167
168
169
# File 'lib/matryoshka/document/html.rb', line 166

def call(env)
  # content = IO.read Matryoshka::ROOT_DIR + '/public/index.html'
  [status, headers, body]
end

#closest_directory(path) ⇒ Object



53
54
55
# File 'lib/matryoshka/document/html.rb', line 53

def closest_directory(path)
  path[/\A.+\//]
end


57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
# File 'lib/matryoshka/document/html.rb', line 57

def convert_proxy_links(proxy_base)
  (content/'a').each do |link|
    href = link.attributes['href']
          
    # Remove base of urls that we are proxying.
    href.sub!(/\A#{proxy_base}/,'')
    
    # Except for absolute links, add template path to complete relative urls
    unless true #is_absolute?(href)
      if href[/\A\//] # link begins with /
        href = '?'
      else
        href = closest_directory(@headers[:template_path]) + href
      end
      link.set_attribute('href', href)
    end
    
    case href
      when /\A#{proxy_base}/
        href.sub!(/#{proxy_base}/,'').sub!(/\A\//,'')
      when /\A#{URI.parse(proxy_base).path}/
        href.sub!(/#{URI.parse(proxy_base).path}/,'').sub!(/\A\//,'')
    end
    
    unless is_absolute?(href)
      href = closest_directory(@headers[:template_path]) + href
    end
    
    link.set_attribute('href', href)
  end
end

#externals_loaded?Boolean

Returns:

  • (Boolean)


8
9
10
# File 'lib/matryoshka/document/html.rb', line 8

def externals_loaded?
  @externals_loaded
end

#is_absolute?(uri) ⇒ Boolean

Returns:

  • (Boolean)


113
114
115
# File 'lib/matryoshka/document/html.rb', line 113

def is_absolute?(uri)
  uri[/\Ahttp:\/\//]
end

#load_externalsObject



90
91
92
93
94
95
96
97
98
99
100
101
102
# File 'lib/matryoshka/document/html.rb', line 90

def load_externals
  (content/EXTERNAL_SEARCH_PATTERN).each do |element|
    external = self.class.new(Matryoshka::External.find(element[:href]))
    external.make_links_absolute(element[:href])
    if external.status == 200
      Merge.new(element,external.content,Merge::EXTERNAL_MERGE).run
      # Cleanup external code
      element.remove_attribute(:href)
      element.remove_attribute(:rel)
    end
  end
  @externals_loaded = true
end


104
105
106
107
108
109
110
111
# File 'lib/matryoshka/document/html.rb', line 104

def make_links_absolute(base_url, replace_pairs = {'a'=>'href', 'img'=>'src'})
  base_dir = base_url[/\A.+\//]
  replace_pairs.each_pair do |search_pattern, attribute|
    (content/search_pattern).each do |link|
      absolute_link link, attribute, base_dir
    end
  end
end

#prepare_proxy(full_request_path = '') ⇒ Object



23
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
# File 'lib/matryoshka/document/html.rb', line 23

def prepare_proxy(full_request_path = '')
  element = content.at PROXY_SEARCH_PATTERN
  proxy_link = if full_request_path == @headers[:template_path]
    element[:href]
  else
    closest_directory(element[:href]) +
      full_request_path.sub(/\A#{closest_directory(@headers[:template_path])}/,'')
  end
  
  # Need to prime external with template_path
  external = self.class.new(Matryoshka::External.find(proxy_link, {:template_path => @headers[:template_path]}))
  # external.make_links_absolute(element[:href])
  base_dir = closest_directory(element[:href])
  
  # if doc is not a full (<html><head> ...) don't bother with transform? 
  
  # merge the external with a copy of the element
  # convoluted copy element (.dup is weird with Hpricot)
  element_copy = Hpricot(element.to_html).at '*'
  external.content = Merge.new(element_copy,external.content,Merge::EXTERNAL_MERGE).run
  external.convert_proxy_links(base_dir)
  external.make_links_absolute(base_dir,{'img'=>'src'})
  
  # Tag the element and the proxy doc so that they cleanly merge.
  # This is to keep the pattern established in answer.rb
  element.set_attribute('id', 'matryoshka_proxy_target')
  external.content.set_attribute('id','matryoshka_proxy_target')
  external
end

#proxy?Boolean

Returns:

  • (Boolean)


19
20
21
# File 'lib/matryoshka/document/html.rb', line 19

def proxy?
  content.at PROXY_SEARCH_PATTERN
end

#proxy_loaded?Boolean

Returns:

  • (Boolean)


12
13
14
# File 'lib/matryoshka/document/html.rb', line 12

def proxy_loaded?
  @proxy_loaded
end