Class: CachedWeb

Inherits:
Object
  • Object
show all
Defined in:
lib/cached_web.rb,
lib/cached_web/version.rb

Constant Summary collapse

VERSION = 
"0.0.1"
@@min_wait_times = 
{}
@@last_request_times = 
{}

Class Method Summary collapse

Instance Method Summary collapse

Class Method Details

.fetch_cache(params, &block) ⇒ Object 



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

def self.fetch_cache(params, &block)
  CachedWeb.new.fetch_cache(params, &block)
end

.get(params) ⇒ Object 



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

def self.get(params)
  CachedWeb.new.get(params)
end

Instance Method Details

#escape_key(key) ⇒ Object 



8
9
10
11
12
13
14
 
# File 'lib/cached_web.rb', line 8

def escape_key(key)
  if key.size > 140
    Digest::MD5.hexdigest(key)
  else
    key.gsub(/[^a-z0-9]/i,"_")
  end
end

#fetch_cache(params) ⇒ Object 



58
59
60
61
62
63
64
65
66
67
68
69
70
71
 
# File 'lib/cached_web.rb', line 58

def fetch_cache(params)
  ret_val = nil
  begin
    ret_val = get_cache(params)
  rescue
    #puts "didn't find data in cache: #{$!}"
  end
  unless ret_val
    ret_val = yield self
    set_cache(params.merge(:page=>ret_val))
  end
  
  return ret_val
end

#get(params) ⇒ Object 



20
21
22
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
52
 
# File 'lib/cached_web.rb', line 20

def get(params)
  url = params[:url].gsub(" ","%20").gsub("%off", "%25off")
  page, redirect_url, headers = get_cache(:url=>url, :expires_in=>params[:expires_in], :details=>true) rescue [nil, nil, nil]
  
  if params[:cache_only]
    return [nil, nil, nil] if page.blank?
    return [page, redirect_url, headers]
  end

  if page.blank?
    uri = URI.parse(url)
    domain = uri.host.downcase
    #puts "Not in cache - getting: #{params[:url]}"
    # Don't hit a site more than once per 5 seconds
    time_to_wait = (@@min_wait_times[domain] || 5) -
      (Time.now - (@@last_request_times[domain] || Time.at(0)))
    if time_to_wait > 0  
      puts "Waiting for #{time_to_wait} on domain #{domain}"
      sleep time_to_wait
    end
    
    @agent = Mechanize.new unless @agent
    page = @agent.get(url) 
    @@last_request_times[uri.host]
    content = page.body
    redirect_url = page.uri.to_s
    headers = page.header
    set_cache(:url=>url, :redirect_url=>redirect_url, :headers=>headers, :content=>content)
    [content, redirect_url, headers]
  else
    [page, redirect_url, headers]
  end
end

#get_cache(params) ⇒ Object 



116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
 
# File 'lib/cached_web.rb', line 116

def get_cache(params)
  key = escape_key(params[:url].present? ? params[:url] : params[:key])
  details = params[:details]
  expires_in = params[:expires_in]

  timestamp = nil

  path = "/tmp/cachedweb/#{key}"
  timestamp = File.ctime(path)

  if expires_in and timestamp.nil?
    #puts "There is no timestamp, force a get of this URL so we can add a timestamp"
    throw "There is no timestamp, force a get of this URL so we can add a timestamp"
  end

  if expires_in and timestamp and timestamp < (Time.now - expires_in)
    #puts "Cache is older than desired, saved on #{timestamp}"
    throw "Cache is older than desired, saved on #{timestamp}"
  end

  h = YAML::load_file(path)
  content = h[:content]
  redirect_url = h[:redirect_url]
  if details
    [content, redirect_url, nil]
  else
    content
  end
end

#post(url, params) ⇒ Object 



73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
 
# File 'lib/cached_web.rb', line 73

def post(url, params)
  puts "Not in cache - POSTing to: #{url}"
  @agent = Mechanize.new unless @agent
  begin
    page = @agent.post(url, params)
    uri = URI.parse(url)
    @@last_request_times[uri.host]
    content = page.body
    redirect_url = page.uri.to_s
    headers = page.header
    set_cache(:url=>url, :redirect_url=>redirect_url, :headers=>headers, :page=>content)
    [content, redirect_url, headers]
  rescue Exception => e
    puts e
    puts e.backtrace
    Toadhopper(AIRBRAKE_KEY).post!(e)
    ["", url, {}]
  end
end

#set_cache(params) ⇒ Object 



93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
 
# File 'lib/cached_web.rb', line 93

def set_cache(params)
  key = escape_key(params[:url].present? ? params[:url] : params[:key])
  content = params[:content]
  redirect_url = params[:redirect_url]

  h = {:content => content, :redirect_url=>redirect_url}


  path = "/tmp/cachedweb/#{key}"
  #puts "Saving file to #{path}"
  begin
    File.open(path, 'w') do |out|
       YAML.dump(h, out)
    end
  rescue Exception
    `mkdir -p /tmp/cachedweb`
    File.open(path, 'w') do |out|
       YAML.dump(h, out)
    end
  end
  true
end