70
71
72
73
74
75
76
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
138
139
140
141
142
143
144
145
146
|
# File 'lib/cohesion.rb', line 70
def self.site(url, options={})
errors = []
failures = []
pages = {}
options[:cache] = options[:cache].to_i if options[:cache]
crawler_options = {:cache_type => :full, :crawl_linked_external => true, :store_inbound_links => true}.merge(options)
statistics = CobwebCrawler.new(crawler_options).crawl(url) do |page|
print page[:url]
duplicate = !pages[Digest::MD5.hexdigest(page[:body])].nil?
pages[Digest::MD5.hexdigest(page[:body])] = [] unless pages[Digest::MD5.hexdigest(page[:body])]
pages[Digest::MD5.hexdigest(page[:body])] << page[:url]
if page[:status_code] == 404
page = Cobweb.new(crawler_options.merge(:cache => nil)).get(page[:url])
end
if page[:status_code] == 404 || duplicate
if duplicate
puts " [duplicate] \e[31m\u2717\e[0m"
else
puts " [#{page[:status_code]}] \e[31m\u2717\e[0m"
end
failures << page
else
puts " \e[32m\u2713\e[0m"
end
end
puts statistics.redis.namespace
puts statistics.get_statistics
total_inbound_failures = 0
total_failures = 0
issues = []
if failures.count == 0
puts "All links working!"
else
puts "Failed urls:"
failures.each do |f|
inbound_links = statistics.inbound_links_for(f[:url])
issues << {:issue => f, :inbound => inbound_links}
total_inbound_failures += inbound_links.count
total_failures += 1
puts ""
puts "#{f[:url]} [ #{f[:status_code]} ]"
inbound_links.each do |inbound_link|
puts " - #{inbound_link}"
end
end
puts ""
puts "Duplicate Content"
puts ""
pages.select{|k,v| v.count > 1}.each do |k,v|
puts "Duplicate: #{k}"
v.map{|x| puts " - #{x}" }
end
puts ""
puts "Total Failed URLs: #{total_failures}"
puts "Total Duplicates: #{pages.map{|d| d[1]}.select{|d| d.count > 1}.inject{|total, d| total + d.count}.count}"
puts "Total Inbound Failures (Pages linking to a 404): #{total_inbound_failures}"
puts ""
end
puts
return issues
end
|