Class: Archaeo::Snapshot

Inherits:
Object
  • Object
show all
Defined in:
lib/archaeo/snapshot.rb

Overview

A single CDX Server API record representing an archived document.

Maps the seven standard CDX fields and provides the computed archive URL via the ArchiveUrl model.

Constant Summary collapse

FIELDS = 
%i[urlkey timestamp original_url
mimetype status_code digest length].freeze
BLOCKED_STATUS = 
-1

Instance Method Summary collapse

Constructor Details

#initialize(urlkey:, timestamp:, original_url:, mimetype: nil, status_code: nil, digest: nil, length: nil) ⇒ Snapshot

Returns a new instance of Snapshot.



16
17
18
19
20
21
22
23
24
25
26
 
# File 'lib/archaeo/snapshot.rb', line 16

def initialize(urlkey:, timestamp:, original_url:,
               mimetype: nil, status_code: nil,
               digest: nil, length: nil)
  @urlkey = urlkey.to_s
  @timestamp = Timestamp.coerce(timestamp)
  @original_url = original_url.to_s
  @mimetype = mimetype.to_s
  @status_code = status_code.to_i
  @digest = digest.to_s
  @length = length.to_i
end

Instance Method Details

#==(other) ⇒ Object Also known as: eql? 



126
127
128
 
# File 'lib/archaeo/snapshot.rb', line 126

def ==(other)
  other.is_a?(self.class) && to_a == other.to_a
end

#age ⇒ Object 



61
62
63
 
# File 'lib/archaeo/snapshot.rb', line 61

def age
  Time.now - @timestamp.to_time
end

#archive_url ⇒ Object 



28
29
30
 
# File 'lib/archaeo/snapshot.rb', line 28

def archive_url
  ArchiveUrl.new(original_url, timestamp: @timestamp).to_s
end

#as_json ⇒ Object 



114
115
116
117
118
119
120
121
122
123
124
 
# File 'lib/archaeo/snapshot.rb', line 114

def as_json(*)
  {
    urlkey: @urlkey,
    timestamp: @timestamp.to_s,
    original_url: @original_url,
    mimetype: @mimetype,
    status_code: @status_code,
    digest: @digest,
    length: @length,
  }
end

#blocked? ⇒ Boolean

Returns:

  • (Boolean) 


37
38
39
 
# File 'lib/archaeo/snapshot.rb', line 37

def blocked?
  @status_code == BLOCKED_STATUS
end

#client_error? ⇒ Boolean

Returns:

  • (Boolean) 


49
50
51
 
# File 'lib/archaeo/snapshot.rb', line 49

def client_error?
  @status_code.between?(400, 499)
end

#duplicate_of?(other) ⇒ Boolean

Returns:

  • (Boolean) 


81
82
83
 
# File 'lib/archaeo/snapshot.rb', line 81

def duplicate_of?(other)
  same_content_as?(other) && timestamp != other.timestamp
end

#error? ⇒ Boolean

Returns:

  • (Boolean) 


57
58
59
 
# File 'lib/archaeo/snapshot.rb', line 57

def error?
  client_error? || server_error?
end

#fetch(client: HttpClient.new, identity: false) ⇒ Object 



85
86
87
88
89
 
# File 'lib/archaeo/snapshot.rb', line 85

def fetch(client: HttpClient.new, identity: false)
  Fetcher.new(client: client).fetch(
    original_url, timestamp: @timestamp, identity: identity
  )
end

#fetch_with_assets(client: HttpClient.new) ⇒ Object 



91
92
93
94
95
 
# File 'lib/archaeo/snapshot.rb', line 91

def fetch_with_assets(client: HttpClient.new)
  Fetcher.new(client: client).fetch_page_with_assets(
    original_url, timestamp: @timestamp
  )
end

#hash ⇒ Object 



131
132
133
 
# File 'lib/archaeo/snapshot.rb', line 131

def hash
  to_a.hash
end

#identity_url ⇒ Object 



32
33
34
35
 
# File 'lib/archaeo/snapshot.rb', line 32

def identity_url
  ArchiveUrl.new(original_url, timestamp: @timestamp,
                               identity: true).to_s
end

#inspect ⇒ Object 



135
136
137
138
 
# File 'lib/archaeo/snapshot.rb', line 135

def inspect
  "#<#{self.class.name} #{timestamp} " \
    "#{original_url} status=#{status_code}>"
end

#newer_than?(seconds) ⇒ Boolean

Returns:

  • (Boolean) 


69
70
71
 
# File 'lib/archaeo/snapshot.rb', line 69

def newer_than?(seconds)
  age <= seconds
end

#older_than?(seconds) ⇒ Boolean

Returns:

  • (Boolean) 


65
66
67
 
# File 'lib/archaeo/snapshot.rb', line 65

def older_than?(seconds)
  age > seconds
end

#redirect? ⇒ Boolean

Returns:

  • (Boolean) 


45
46
47
 
# File 'lib/archaeo/snapshot.rb', line 45

def redirect?
  @status_code.between?(300, 399)
end

#same_content_as?(other) ⇒ Boolean

Returns:

  • (Boolean) 


73
74
75
76
77
78
79
 
# File 'lib/archaeo/snapshot.rb', line 73

def same_content_as?(other)
  return false unless other.is_a?(self.class)
  return false if digest.nil? || digest.empty?
  return false if other.digest.nil? || other.digest.empty?

  digest == other.digest
end

#server_error? ⇒ Boolean

Returns:

  • (Boolean) 


53
54
55
 
# File 'lib/archaeo/snapshot.rb', line 53

def server_error?
  @status_code.between?(500, 599)
end

#success? ⇒ Boolean

Returns:

  • (Boolean) 


41
42
43
 
# File 'lib/archaeo/snapshot.rb', line 41

def success?
  @status_code == 200
end

#to_a ⇒ Object 



97
98
99
100
 
# File 'lib/archaeo/snapshot.rb', line 97

def to_a
  [@urlkey, @timestamp, @original_url, @mimetype,
   @status_code, @digest, @length]
end

#to_h ⇒ Object 



102
103
104
105
106
107
108
109
110
111
112
 
# File 'lib/archaeo/snapshot.rb', line 102

def to_h
  {
    urlkey: @urlkey,
    timestamp: @timestamp,
    original_url: @original_url,
    mimetype: @mimetype,
    status_code: @status_code,
    digest: @digest,
    length: @length,
  }
end