Class: WebpageArchivist::Fetcher::WebpageRequest

Inherits:
Object
  • Object
show all
Defined in:
lib/webpage-archivist/fetcher/webpage_request.rb

Overview

Requesting a webpage.

Instance Attribute Summary collapse

Instance Method Summary collapse

Constructor Details

#initialize(webpage, fetcher_watcher) ⇒ WebpageRequest

Create a request

webpage

the Webpage we want to fetch

fetcher_watcher

to be notified when the request is over



15
16
17
18
19
20
21
22
23
24
25
26
 
# File 'lib/webpage-archivist/fetcher/webpage_request.rb', line 15

def initialize webpage, fetcher_watcher
  @webpage = webpage
  @fetcher_watcher = fetcher_watcher

  @uri = Addressable::URI.parse(webpage.uri)

  @waiting_requests = 0

  @status = :fetching

  @plumber = RequestsPlumber.new self
end

Instance Attribute Details

#instance ⇒ Object (readonly)

Returns the value of attribute instance.



10
11
12
 
# File 'lib/webpage-archivist/fetcher/webpage_request.rb', line 10

def instance
  @instance
end

#result_code ⇒ Object (readonly)

Returns the value of attribute result_code.



10
11
12
 
# File 'lib/webpage-archivist/fetcher/webpage_request.rb', line 10

def result_code
  @result_code
end

#status ⇒ Object (readonly)

Returns the value of attribute status.



10
11
12
 
# File 'lib/webpage-archivist/fetcher/webpage_request.rb', line 10

def status
  @status
end

#uri ⇒ Object (readonly)

Returns the value of attribute uri.



10
11
12
 
# File 'lib/webpage-archivist/fetcher/webpage_request.rb', line 10

def uri
  @uri
end

#webpage ⇒ Object (readonly)

Returns the value of attribute webpage.



10
11
12
 
# File 'lib/webpage-archivist/fetcher/webpage_request.rb', line 10

def webpage
  @webpage
end

Instance Method Details

#after_requests ⇒ Object

Process the response once all the elements have been fetched



162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
 
# File 'lib/webpage-archivist/fetcher/webpage_request.rb', line 162

def after_requests
  ::WebpageArchivist.debug "After requests #{@uri}" if ::WebpageArchivist.log
  if @modified

    # replace elements with the local uris of the elements
    @content.each_stylesheet do |stylesheet|
      if e = @plumber[stylesheet['href']].andand.element
        stylesheet['href'] = e.file_name
      end
    end

    @content.each_script do |script|
      if e = @plumber[script['src']].andand.element
        script['src'] = e.file_name
      end
    end

    @content.each_image do |img|
      if e = @plumber[img['src']].andand.element
        img['src'] = e.file_name
      end
    end
    webpage.save_content @content.to_html
  end

  commit_timestamp = DateTime.now.strftime('%Y/%m/%d %H:%M:%S')
  # store the content, clean the repo and commit
  webpage.update_repo_commit_changes @plumber.requests_files, commit_timestamp
  @instance = WebpageArchivist::Instance.create(:webpage => webpage, :commit_timestamp => commit_timestamp)

  @fetcher_watcher.end_request self, true
end

#make_absolute_if_modified(element, link_property) ⇒ Object

Make an element’s uri absolute

element

the element

link_property

the property holding the uri



142
143
144
145
146
147
148
 
# File 'lib/webpage-archivist/fetcher/webpage_request.rb', line 142

def make_absolute_if_modified element, link_property
  if @modified
    element[link_property] = uri.absolutize(element[link_property])
  else
    element[link_property]
  end
end

#process_content ⇒ Object

Process the content



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
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
 
# File 'lib/webpage-archivist/fetcher/webpage_request.rb', line 77

def process_content
  if @modified
    charset = @plumber.response_charset @http
    @content = WebpageArchivist::HtmlDocument.new(@http.response, @uri, charset)
  else
    @content = WebpageArchivist::HtmlDocument.new(webpage.last_content, @uri, webpage.last_encoding)
  end

  # Elements from the stylesheets
  already = Set.new
  @content.each_stylesheet do |stylesheet|
    uri = make_absolute_if_modified(stylesheet, 'href')
    unless already.include? uri
      already.add uri
      @plumber.request_element self, WebpageArchivist::Stylesheet, uri
      @waiting_requests += 1
    end
  end

  # Elements from the scripts
  already = Set.new
  @content.each_script do |script|
    uri = make_absolute_if_modified(script, 'src')
    unless already.include? uri
      already.add uri
      @plumber.request_element self, WebpageArchivist::Script, uri
      @waiting_requests += 1
    end
  end

  # Elements from the images
  already = Set.new
  @content.each_image do |img|
    uri = make_absolute_if_modified(img, 'src')
    unless already.include? uri
      already.add uri
      @plumber.request_element self, WebpageArchivist::Image, uri
      @waiting_requests += 1
    end
  end

  if @modified
    # Make links absolutes
    @content.each_link do |link|
      make_absolute_if_modified(link, 'href')
    end

    @webpage.update(
        :last_modified => @plumber.last_modified(@http),
        :last_content => @content.to_html,
        :last_charset => @content.charset)
  end

  @status = :fetching_requests

  # No external resource -> end here
  if @waiting_requests == 0
    after_requests
  end

end

#process_response ⇒ Object

Process the response



58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
 
# File 'lib/webpage-archivist/fetcher/webpage_request.rb', line 58

def process_response
  @result_code = @http.response_header.status
  ::WebpageArchivist.debug "[#{@uri}] returned #{@result_code}" if ::WebpageArchivist.log
  if [304, 408, 0].include? @result_code
    # Not changed
    ::WebpageArchivist.debug "[#{@uri}] not modified" if ::WebpageArchivist.log
    @modified = false
    process_content
  elsif result_code == 200
    ::WebpageArchivist.debug "[#{@uri}] modified" if ::WebpageArchivist.log
    @modified = true
    process_content
  else
    ::WebpageArchivist.debug "Error #{@uri} #{@result_code}" if ::WebpageArchivist.log
    @fetcher_watcher.end_request self, false
  end
end

#request_over(uri) ⇒ Object

Called by a request when it is over

uri

the request uri



152
153
154
155
156
157
158
159
 
# File 'lib/webpage-archivist/fetcher/webpage_request.rb', line 152

def request_over uri
  @waiting_requests -= 1

  ::WebpageArchivist.debug "Request over for [#{@uri}] on [#{uri}], missing #{@waiting_requests}" if ::WebpageArchivist.log
  if (@status == :fetching_requests) && (@waiting_requests <= 0)
    after_requests
  end
end

#start(retries = 3) ⇒ Object

Start the request Not in initialize so we can throttle the number of connection

retries

number of retries in case of error



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
 
# File 'lib/webpage-archivist/fetcher/webpage_request.rb', line 31

def start retries = 3
  @http = EventMachine::HttpRequest.new(@uri).get :redirects => 5, :timeout => 30, :head => {'If-Modified-Since' => webpage.last_modified, 'accept-encoding' => 'gzip, compressed'}
  @http.callback do
    if ([500, 503].include? @http.response_header.status) && (retries > 0)
      start(retries - 1)
    else
      begin
        process_response
      rescue Exception => e
        ::WebpageArchivist.error e if ::WebpageArchivist.log
      end
    end
  end
  @http.errback do
    if retries > 0
      start(retries - 1)
    else
      begin
        process_response
      rescue Exception => e
        ::WebpageArchivist.error e if ::WebpageArchivist.log
      end
    end
  end
end