Class: Wmap::SiteTracker::WpTracker

Inherits:
Wmap::SiteTracker show all
Includes:
Utils, Utils::WpDetect
Defined in:
lib/wmap/site_tracker/wp_tracker.rb

Constant Summary

Constants included from Utils::UrlMagic

Utils::UrlMagic::Max_http_timeout, Utils::UrlMagic::User_agent

Constants included from Utils::DomainRoot

Utils::DomainRoot::File_ccsld, Utils::DomainRoot::File_cctld, Utils::DomainRoot::File_gtld, Utils::DomainRoot::File_tld

Instance Attribute Summary collapse

Attributes inherited from Wmap::SiteTracker

#known_sites, #sites_file

Instance Method Summary collapse

Methods included from Utils::WpDetect

#is_wp?, #wp_css?, #wp_gen?, #wp_load_styles?, #wp_login?, #wp_meta?, #wp_readme?, #wp_rpc?, #wp_ver, #wp_ver_generator, #wp_ver_load_styles, #wp_ver_login, #wp_ver_meta, #wp_ver_readme

Methods included from Utils

#cidr_2_ips, #file_2_hash, #file_2_list, #get_nameserver, #get_nameservers, #host_2_ip, #host_2_ips, #is_cidr?, #is_fqdn?, #is_ip?, #list_2_file, #reverse_dns_lookup, #sort_ips, #valid_dns_record?, #zone_transferable?

Methods included from Utils::Logger

#wlog

Methods included from Utils::UrlMagic

#create_absolute_url_from_base, #create_absolute_url_from_context, #host_2_url, #is_site?, #is_ssl?, #is_url?, #landing_location, #make_absolute, #normalize_url, #open_page, #redirect_location, #response_code, #response_headers, #url_2_host, #url_2_path, #url_2_port, #url_2_site, #urls_on_same_domain?

Methods included from Utils::DomainRoot

#get_domain_root, #get_domain_root_by_ccsld, #get_domain_root_by_cctld, #get_domain_root_by_tlds, #get_sub_domain, #is_domain_root?, #print_ccsld, #print_cctld, #print_gtld

Methods inherited from Wmap::SiteTracker

#bulk_delete, #bulk_refresh, #count, #delete, #file_add, #file_delete, #file_refresh, #get_ext_sites, #get_int_sites, #get_ip_sites, #get_prim_uniq_sites, #get_redirection_url, #get_redirection_urls, #get_ssl_sites, #get_uniq_sites, #load_site_stores_from_file, #print_all_sites, #print_ext_sites, #print_int_sites, #print_ip_sites, #print_site, #print_ssl_sites, #print_uniq_sites, #refresh_all, #refresh_ip_sites, #refresh_uniq_sites, #resolve_ip_sites, #save_sites_to_file!, #save_uniq_sites, #save_uniq_sites_xml, #search, #site_check, #site_ip_known?, #site_known?

Constructor Details

#initialize(params = {}) ⇒ WpTracker

WordPress checker instance default variables



24
25
26
27
28
29
30
31
32
33
34
35
 
# File 'lib/wmap/site_tracker/wp_tracker.rb', line 24

def initialize (params = {})
  @verbose=params.fetch(:verbose, false)
  @data_dir=params.fetch(:data_dir, File.dirname(__FILE__)+'/../../../data/')
  Dir.mkdir(@data_dir) unless Dir.exist?(@data_dir)
   @sites_wp=params.fetch(:sites_wp, @data_dir+"wp_sites")
  @http_timeout=params.fetch(:http_timeout, 5000)
  @max_parallel=params.fetch(:max_parallel, 40)
  Dir.mkdir(@data_dir) unless Dir.exist?(@data_dir)
  @log_file=@data_dir + "wp_checker.log"
  File.write(@sites_wp, "") unless File.exist?(@sites_wp)
   load_from_file(@sites_wp)
end

Instance Attribute Details

#data_dir ⇒ Object

include Singleton



20
21
22
 
# File 'lib/wmap/site_tracker/wp_tracker.rb', line 20

def data_dir
  @data_dir
end

#http_timeout ⇒ Object

include Singleton



20
21
22
 
# File 'lib/wmap/site_tracker/wp_tracker.rb', line 20

def http_timeout
  @http_timeout
end

#known_wp_sites ⇒ Object (readonly)

Returns the value of attribute known_wp_sites.



21
22
23
 
# File 'lib/wmap/site_tracker/wp_tracker.rb', line 21

def known_wp_sites
  @known_wp_sites
end

#max_parallel ⇒ Object

include Singleton



20
21
22
 
# File 'lib/wmap/site_tracker/wp_tracker.rb', line 20

def max_parallel
  @max_parallel
end

#sites_wp ⇒ Object

include Singleton



20
21
22
 
# File 'lib/wmap/site_tracker/wp_tracker.rb', line 20

def sites_wp
  @sites_wp
end

#verbose ⇒ Object

include Singleton



20
21
22
 
# File 'lib/wmap/site_tracker/wp_tracker.rb', line 20

def verbose
  @verbose
end

Instance Method Details

#add(url, use_cache = true) ⇒ Object

Add wordpress entry to the cache one at a time



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
 
# File 'lib/wmap/site_tracker/wp_tracker.rb', line 86

def add(url, use_cache=true)
  puts "Add entry to the local cache table: #{url}" if @verbose
   site=url_2_site(url)
  if use_cache && @known_wp_sites.key?(site)
    puts "Site is already exist. Skipping: #{site}"
  else
    record=Hash.new
    redirection = landing_location(site)
    if not [nil, ''].include?(redirection)
      if is_wp?(redirection)
        version = wp_ver(redirection)
        record['site'] = site
        record['version'] = version
        record['redirection'] = redirection
        @known_wp_sites[site]=record
        puts "Entry added: #{record}"
      end
    else
      if is_wp?(site)
        version = wp_ver(site)
        record['version'] = version
        record['redirection'] = redirection
        @known_wp_sites[site]=record
        puts "Entry added: #{record}"
      end
    end
  end
   return record
rescue => ee
  puts "Exception on method #{__method__}: #{ee}: #{url}" if @verbose
end

#bulk_add(list, num = @max_parallel, use_cache = true) ⇒ Object Also known as: adds

Method to load wp sites in parallel



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
147
148
149
150
 
# File 'lib/wmap/site_tracker/wp_tracker.rb', line 119

def bulk_add(list,num=@max_parallel,use_cache=true)
  puts "Add entries to the local wp_site store from list:\n #{list}"
  results=Hash.new
  list = list - [nil,""]
  if list.size > 0
    puts "Start parallel adding on the sites:\n #{list}"
    Parallel.map(list, :in_processes => num) { |target|
      add(target,use_cache)
    }.each do |process|
      if process.nil?
        next
      elsif process.empty?
        next #do nothing
      else
        results[process['site']]=Hash.new
        results[process['site']]=process
      end
    end
    @known_wp_sites.merge!(results)
  else
    puts "Error: no entry is added. Please check your list and try again."
  end
  puts "Done adding site entries."
  if results.size>0
    puts "New entries added: #{results}"
  else
    puts "No new entry added. "
  end
  return results
rescue => ee
  puts "Exception on method #{__method__}: #{ee}" if @verbose
end

#load_from_file(file = @sites_wp, lc = true) ⇒ Object

‘setter’ to load the known wordpress sites into an instance variable



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
 
# File 'lib/wmap/site_tracker/wp_tracker.rb', line 38

def load_from_file (file=@sites_wp, lc=true)
  puts "Loading trusted file: #{file}" if @verbose
  @known_wp_sites=Hash.new
  f_wp_sites=File.open(file, 'r')
  f_wp_sites.each_line do |line|
    puts "Processing line: #{line}" if @verbose
    line=line.chomp.strip
    next if line.nil?
    next if line.empty?
    next if line =~ /^\s*#/
    line=line.downcase if lc==true
    entry=line.split(',')
    site = entry[0].strip()
    next if site.nil?
    if @known_wp_sites.key?(site)
      next
    else
      @known_wp_sites[site] = Hash.new
      @known_wp_sites[site]['site'] = site
      @known_wp_sites[site]['version'] = entry[1].strip()
      @known_wp_sites[site]['redirection'] = entry[2].strip()
    end
  end
  f_wp_sites.close
  return @known_wp_sites
rescue => ee
  puts "Exception on method #{__method__}: #{ee}" if @verbose
  return Hash.new
end

#refresh(target, use_cache = false) ⇒ Object

Refresh one site entry then update the instance variable (cache)



154
155
156
 
# File 'lib/wmap/site_tracker/wp_tracker.rb', line 154

def refresh (target,use_cache=false)
  return add(target,use_cache)
end

#refreshs(num = @max_parallel, use_cache = false) ⇒ Object

Refresh wordpress site entries within the sitetracker list



159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
 
# File 'lib/wmap/site_tracker/wp_tracker.rb', line 159

def refreshs (num=@max_parallel,use_cache=false)
  puts "Add entries to the local cache table from site tracker: " if @verbose
  results=Hash.new
  wps=@known_wp_sites.keys
  if wps.size > 0
    Parallel.map(wps, :in_processes => num) { |target|
      refresh(target,use_cache)
    }.each do |process|
      if process.nil?
        next
      elsif process.empty?
        #do nothing
      else
        site = process['site']
        results[site] = process
      end
    end
    @known_wp_sites.merge!(results)
    puts "Done loading entries."
    return results
  else
    puts "Error: no entry is loaded. Please check your list and try again."
  end
  wps=nil
  return results
#rescue => ee
# puts "Exception on method #{__method__}: #{ee}" if @verbose
# return Hash.new
end

#save_to_file!(file_wps = @sites_wp, wps = @known_wp_sites) ⇒ Object Also known as: save!

Save the current hash table into a file



69
70
71
72
73
74
75
76
77
78
79
80
81
82
 
# File 'lib/wmap/site_tracker/wp_tracker.rb', line 69

def save_to_file!(file_wps=@sites_wp, wps=@known_wp_sites)
  puts "Saving the current wordpress site table from memory to file: #{file_wps} ..." if @verbose
  timestamp=Time.now
  f=File.open(file_wps, 'w')
  f.write "# Local wps file created by class #{self.class} method #{__method__} at: #{timestamp}\n"
  f.write "# WP Site URL, WP Version, Redirection \n"
  (wps.keys - [nil,'']).sort.map do |key|
    f.write "#{key}, #{wps[key]['version']}, #{wps[key]['redirection']}\n"
  end
  f.close
  puts "WordPress site cache table is successfully saved: #{file_wps}"
rescue => ee
  puts "Exception on method #{__method__}: #{ee}" if @verbose
end