arachnid 0.3.0 → 0.3.1
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- data/lib/arachnid.rb +22 -1
- metadata +1 -1
data/lib/arachnid.rb
CHANGED
|
@@ -15,6 +15,7 @@ class Arachnid
|
|
|
15
15
|
@split_url_at_hash = options[:split_url_at_hash] ? options[:split_url_at_hash] : false
|
|
16
16
|
@exclude_urls_with_hash = options[:exclude_urls_with_hash] ? options[:exclude_urls_with_hash] : false
|
|
17
17
|
@exclude_urls_with_images = options[:exclude_urls_with_images] ? options[:exclude_urls_with_images] : false
|
|
18
|
+
@proxy_list = options[:proxy_list] ? options[:proxy_list] : nil
|
|
18
19
|
|
|
19
20
|
@debug = options[:debug] ? options[:debug] : false
|
|
20
21
|
end
|
|
@@ -26,6 +27,8 @@ class Arachnid
|
|
|
26
27
|
#defaults to -1 so it will always keep running until it runs out of urls
|
|
27
28
|
max_urls = options[:max_urls] ? options[:max_urls] : nil
|
|
28
29
|
|
|
30
|
+
|
|
31
|
+
|
|
29
32
|
@hydra = Typhoeus::Hydra.new(:max_concurrency => threads)
|
|
30
33
|
@global_visited = BloomFilter::Native.new(:size => 1000000, :hashes => 5, :seed => 1, :bucket => 8, :raise => false)
|
|
31
34
|
@global_queue = []
|
|
@@ -38,7 +41,11 @@ class Arachnid
|
|
|
38
41
|
temp_queue.each do |q|
|
|
39
42
|
|
|
40
43
|
begin
|
|
41
|
-
|
|
44
|
+
ip,port,user,pass = grab_proxy
|
|
45
|
+
|
|
46
|
+
request = Typhoeus::Request.new(q, :timeout => 10000, :follow_location => true) if ip == nil
|
|
47
|
+
request = Typhoeus::Request.new(q, :timeout => 10000, :follow_location => true, :proxy => "#{ip}:#{port}") if ip != nil && user == nil
|
|
48
|
+
request = Typhoeus::Request.new(q, :timeout => 10000, :follow_location => true, :proxy => "#{ip}:#{port}", :proxy_username => user, :proxy_password => pass) if user != nil
|
|
42
49
|
|
|
43
50
|
request.on_complete do |response|
|
|
44
51
|
|
|
@@ -95,6 +102,20 @@ class Arachnid
|
|
|
95
102
|
end
|
|
96
103
|
end
|
|
97
104
|
|
|
105
|
+
def internal_link?(url, effective_url)
|
|
106
|
+
|
|
107
|
+
absolute_url = make_absolute(url, effective_url)
|
|
108
|
+
|
|
109
|
+
parsed_url = parse_domain(absolute_url)
|
|
110
|
+
|
|
111
|
+
def grab_proxy
|
|
112
|
+
|
|
113
|
+
return nil unless @proxy_list
|
|
114
|
+
|
|
115
|
+
return @proxy_list.sample.split(':')
|
|
116
|
+
|
|
117
|
+
end
|
|
118
|
+
|
|
98
119
|
def internal_link?(url, effective_url)
|
|
99
120
|
|
|
100
121
|
absolute_url = make_absolute(url, effective_url)
|