sitemap_check 0.1.0 → 0.1.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 51d3f2f99c25a2034868ae1434c94e9adafdb50a
-  data.tar.gz: adf5452975f758257bfabf1d87e471bd2c2fb98a
+  metadata.gz: 84dc68b7fb5e1fec3814c8f6e82f94d8cf43b619
+  data.tar.gz: 37cefa64e976770366fafe82b89d5d34dc97adce
 SHA512:
-  metadata.gz: b9736873b07433315af0134f0cd5df00a39d9dc7b316fc2066bf3d00423dc38007e32cbc587320058e1360180d14549f7c9e48585744d263fdcf128461b055d6
-  data.tar.gz: 9d6411787e55f4767cf8a10224b04ed4696b3381bfc425a2010c960a2c81512658bde3299b0abf56df580b028a531af6cc09249be44ff547926f63672e15f742
+  metadata.gz: a89b4e56d7df0b4cafef3e5844c64afc5510d7dfeccf6db602bcfecd591d1a95acc15370fa10468de214d5965e0bc47e9d066fdd306a6f1827e931b089360bf2
+  data.tar.gz: 0cc65ef4219b022f7675b53e10a3a242ac3e723d75ea89b0ad38b4c5650011f70c06828286b3fe123d37929141f6a246946808a2fb4f8dcc02861a63a9c09829

data/README.md

@@ -9,3 +9,12 @@
 ```bash
 $ CHECK_URL=http://reevoo.com/sitemap_index.xml sitemap_check
 ```
+
+## Config
+
+Config can be set with enviroment variables
+
+variable     | default | description
+-------------|---------|-------------
+`CHECK_URL`  | `nil`   | The url of the sitemap or sitemap index to check
+`CONCURRENCY`| `10`    | The number of concurent threads to use when checking the sitemap

data/lib/sitemap_check/logger.rb

@@ -0,0 +1,16 @@
+class SitemapCheck
+  class Logger
+    def initialize(stream = $stdout)
+      self.stream = stream
+      self.mutex = Mutex.new
+    end
+
+    def log(message)
+      mutex.synchronize { stream.puts message }
+    end
+
+    protected
+
+    attr_accessor :stream, :mutex
+  end
+end

data/lib/sitemap_check/page.rb

@@ -13,7 +13,7 @@ class SitemapCheck
 
     def exists?
       @_exists ||= http.head(url, follow_redirect: true).ok?
-    rescue SocketError, HTTPClient::ConnectTimeoutError
+    rescue SocketError, HTTPClient::ConnectTimeoutError, Errno::ETIMEDOUT
       self.tries += 1
       if tries < 5
         sleep holdoff

data/lib/sitemap_check/sitemap.rb

@@ -1,11 +1,13 @@
 require 'httpclient'
 require 'sitemap_check/page'
+require 'sitemap_check/logger'
 require 'nokogiri'
 require 'colorize'
 
 class SitemapCheck
   class Sitemap
-    def initialize(url, http = HTTPClient.new)
+    def initialize(url, http = HTTPClient.new, logger = Logger.new)
+      self.logger = logger
       self.url = url
       self.checked = 0
       self.http = http
@@ -32,33 +34,32 @@ class SitemapCheck
 
     protected
 
-    attr_accessor :http, :doc
+    attr_accessor :http, :doc, :logger
     attr_writer :url, :checked
 
     private
 
     def concurency
-      ENV.fetch('CONCURENCY', 10)
+      ENV.fetch('CONCURRENCY', '10').to_i
     end
 
     def find_missing_pages # rubocop:disable Metrics/AbcSize
       q = Queue.new
-      mutex = Mutex.new
       pages.each { |page| q.push page }
       concurency.times.map do
         Thread.new do
           begin
             while (page = q.pop(true))
               unless page.exists?
-                puts "  missing: #{page.url}".red
+                logger.log "  missing: #{page.url}".red
                 page
               end
-              mutex.synchronize { self.checked += 1 }
             end
           rescue ThreadError # rubocop:disable Lint/HandleExceptions
           end
         end
       end.each(&:join)
+      self.checked = pages.count
       pages.reject(&:exists?)
     end
 

data/lib/sitemap_check/version.rb

@@ -1,3 +1,3 @@
 class SitemapCheck
-  VERSION = '0.1.0'
+  VERSION = '0.1.1'
 end

data/lib/sitemap_check.rb

@@ -3,12 +3,14 @@ require 'sitemap_check/sitemap'
 
 class SitemapCheck
 
+
   def self.check
     $stdout.sync = true
     new.check
   end
 
   def initialize(http = HTTPClient.new)
+    self.start_time = Time.now
     self.exit_code = 0
     puts "Expanding Sitemaps from #{ENV['CHECK_URL']}"
     self.sitemaps = Sitemap.new(ENV['CHECK_URL'], http).sitemaps
@@ -17,18 +19,36 @@ class SitemapCheck
   def check
     check_indexes
     check_pages
+    stats
     exit exit_code
   end
 
   protected
 
-  attr_accessor :sitemaps, :exit_code
+  attr_accessor :sitemaps, :exit_code, :start_time, :logger
 
   private
 
+  def stats
+    puts "checked #{sitemaps.count} sitemaps and #{checked_pages} in #{time_taken} seconds"
+    puts "thats #{pages_per_second} pages per second"
+  end
+
+  def pages_per_second
+    checked_pages / time_taken
+  end
+
+  def time_taken
+    Time.now - start_time
+  end
+
+  def checked_pages
+    sitemaps.map(&:checked).reduce(&:+)
+  end
+
   def check_indexes
     sitemaps.reject(&:exists?).each do |sitemap|
-      puts "#{sitemap.url} does not exist".red.bold
+      puts "  #{sitemap.url} does not exist".red.bold
       self.exit_code = 1
     end
     puts ''
@@ -58,14 +78,14 @@ class SitemapCheck
 
   def missing_pages(sitemap)
     self.exit_code = 1
-    puts "checked #{sitemap.checked} pages and #{sitemap.missing_pages.count} were missing".red.bold
+    puts "  checked #{sitemap.checked} pages and #{sitemap.missing_pages.count} were missing".red.bold
   end
 
   def a_ok(sitemap)
-    puts "checked #{sitemap.checked} pages and everything was ok".green.bold
+    puts "  checked #{sitemap.checked} pages and everything was ok".green.bold
   end
 
   def nothing_doing
-    puts 'this sitemap did not contain any pages'.green
+    puts '  this sitemap did not contain any pages'.green
   end
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: sitemap_check
 version: !ruby/object:Gem::Version
-  version: 0.1.0
+  version: 0.1.1
 platform: ruby
 authors:
 - Ed Robinson
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2015-04-17 00:00:00.000000000 Z
+date: 2015-04-23 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: nokogiri
@@ -155,6 +155,7 @@ files:
 - Rakefile
 - bin/sitemap_check
 - lib/sitemap_check.rb
+- lib/sitemap_check/logger.rb
 - lib/sitemap_check/page.rb
 - lib/sitemap_check/sitemap.rb
 - lib/sitemap_check/version.rb