Class: Rcrawl::RobotRules

Inherits:
Object
  • Object
show all
Defined in:
lib/rcrawl/robot_rules.rb

Overview

Based on Perl’s WWW::RobotRules module, by Gisle Aas.

Instance Method Summary collapse

Constructor Details

#initialize(user_agent) ⇒ RobotRules

Returns a new instance of RobotRules.



15
16
17
18
19
 
# File 'lib/rcrawl/robot_rules.rb', line 15

def initialize( user_agent )
  @user_agent = user_agent.scan(/\S+/).first.sub(%r{/.*},
															"").downcase
  @rules      = Hash.new { |rules, rule| rules[rule] = Array.new }
end

Instance Method Details

#allowed?(text_uri) ⇒ Boolean

Returns:

  • (Boolean) 


72
73
74
75
76
77
78
79
80
 
# File 'lib/rcrawl/robot_rules.rb', line 72

def allowed?( text_uri )
  uri      = URI.parse(text_uri)
  location = "#{uri.host}:#{uri.port}"
  path     = uri.path

  return true unless %w{http https}.include?(uri.scheme)

  not @rules[location].any? { |rule| path.index(rule) == 0 }
end

#parse(text_uri, robots_data) ⇒ Object 



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
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
 
# File 'lib/rcrawl/robot_rules.rb', line 21

def parse( text_uri, robots_data )
  uri      = URI.parse(text_uri)
  location = "#{uri.host}:#{uri.port}"
  @rules.delete(location)

  rules      = robots_data.split(/[\015\012]+/).
									map { |rule| rule.sub(/\s*#.*$/, "") }
  anon_rules = Array.new
  my_rules   = Array.new
  current    = anon_rules
  rules.each do |rule|
	 case rule
	 when /^\s*User-Agent\s*:\s*(.+?)\s*$/i
		break unless my_rules.empty?

		current = if $1 == "*"
		  anon_rules
		elsif $1.downcase.index(@user_agent)
		  my_rules
		else
		  nil
		end
	 when /^\s*Disallow\s*:\s*(.*?)\s*$/i
		next if current.nil?

		if $1.empty?
		  current << nil
		else
		  disallow = URI.parse($1)

		  next unless disallow.scheme.nil? or disallow.scheme == uri.scheme
		  next unless disallow.port.nil?   or disallow.port == uri.port
		  next unless disallow.host.nil?   or
						  disallow.host.downcase == uri.host.downcase

		  disallow = disallow.path
		  disallow = "/"            if disallow.empty?
		  disallow = "/#{disallow}" unless disallow[0] == ?/

		  current << disallow
		end
	 end
  end

  @rules[location] = if my_rules.empty?
	 anon_rules.compact
  else
	 my_rules.compact
  end
end