sitemap_checker 0.0.2 → 0.0.3

data/README.md

@@ -1,6 +1,6 @@
 # SitemapChecker
 
-Takes a url pointing to an xml or xml.gz sitemap or siteindex file and returns array of status messages for urls contained within.
+Takes a url pointing to an xml or xml.gz sitemap or siteindex file and returns array of urls contained within.
 
 ## Installation
 
@@ -18,8 +18,13 @@ Or install it yourself as:
 
 ## Usage
 
+Get list of urls from xml or xml.gz sitemap url
+    
     SitemapChecker::Checker.new(url)
 
+Get status of url
+
+    SitemapChecker::Checker.get_status(url)
 
 ## Contributing
 

data/lib/sitemap_checker.rb

@@ -5,14 +5,23 @@ require 'zlib'
 
 module SitemapChecker
   class Checker
-    attr_reader :status_list
+    attr_reader :url_list
 
     def initialize(url,schema='')
       @url = url
+      @url_list = Array.new
       @status_list = Array.new
       process_xml
     end
 
+    def self.get_status(url)
+      begin
+        status = [url.content,open(url).status[0]]
+      rescue OpenURI::HTTPError => e
+        status = [url.content,e.io.status[0]]
+      end
+    end
+
     private
 
     def get_xml_from_url(url)
@@ -28,12 +37,12 @@ module SitemapChecker
       ixsd = Nokogiri::XML::Schema(open('http://www.sitemaps.org/schemas/sitemap/0.9/siteindex.xsd'))
       xml = get_xml_from_url(@url)
       if mxsd.valid?(xml)
-        @status_list = get_status_list(urls(xml))
+        @url_list = urls(xml)
       elsif ixsd.valid?(xml)
         maps = urls(xml)
         maps.each do |map|
           xml = get_xml_from_url(map)
-          @status_list += get_status_list(urls(xml))
+          @url_list = urls(xml)
         end
       else raise 'Invalid Schema'
         false

data/lib/sitemap_checker/version.rb

@@ -1,3 +1,3 @@
 module SitemapChecker
-  VERSION = "0.0.2"
+  VERSION = "0.0.3"
 end

data/spec/sitemap_checker_spec.rb

@@ -16,32 +16,35 @@ describe SitemapChecker do
     stub_request(:get, "http://www.sitemaps.org/schemas/sitemap/0.9/siteindex.xsd").to_return(:status => 200, :body => File.read(@dir + 'fixtures/siteindex.xsd'), :headers => {})
   end
 
-  it "accepts xml and gzipped siteindexes" do
+  it "accepts xml siteindexes" do
     @xml_sitemap = SitemapChecker::Checker.new('http://www.github.com/siteindex.xml')
+    @xml_sitemap.url_list.size.should eq(2)
+  end
+
+  it "accepts gzipped siteindexes" do
     @gz_sitemap = SitemapChecker::Checker.new('http://www.github.com/siteindex.xml.gz')
-    @xml_sitemap.status_list.size.should eq(4)
-    @gz_sitemap.status_list.size.should eq(4)
+    @gz_sitemap.url_list.size.should eq(2)
+  end
+
+  it "accepts xml sitemaps" do
+    @xml_sitemap = SitemapChecker::Checker.new('http://www.github.com/sitemap.xml')
+    @xml_sitemap.url_list.size.should eq(2)
   end
 
   it "accepts xml and gzipped sitemaps" do
     @xml_sitemap = SitemapChecker::Checker.new('http://www.github.com/sitemap.xml')
     @gz_sitemap = SitemapChecker::Checker.new('http://www.github.com/sitemap.xml.gz')
-    @xml_sitemap.status_list.size.should eq(2)
-    @gz_sitemap.status_list.size.should eq(2)
+    @xml_sitemap.url_list.size.should eq(2)
+    @gz_sitemap.url_list.size.should eq(2)
   end
 
   it "Errors if input doc does not match sitemap schema" do
     lambda {SitemapChecker::Checker.new('http://www.github.com')}.should raise_error(RuntimeError, 'Invalid Schema')
   end
 
-  it "returns list of urls with responses from sitemap" do
+  it "returns status if given a url" do
     @sitemap = SitemapChecker::Checker.new('http://www.github.com/sitemap.xml')
-    @sitemap.status_list.should eq([['http://www.github.com','200'], ['http://www.github.com/404','404']])
-  end
-
-  it "returns list of urls with responses from siteindex" do
-    @siteindex = SitemapChecker::Checker.new('http://www.github.com/siteindex.xml')
-    @siteindex.status_list.should eq([['http://www.github.com','200'], ['http://www.github.com/404','404'], ['http://www.github.com','200'], ['http://www.github.com/404','404']])
+    SitemapChecker::Checker.get_status(@sitemap.url_list.first).should eq(['http://www.github.com','200'])
   end
 
 end

metadata

@@ -1,13 +1,13 @@
 --- !ruby/object:Gem::Specification 
 name: sitemap_checker
 version: !ruby/object:Gem::Version 
-  hash: 1317335842608397244
+  hash: 1889055196096400351
   prerelease: 
   segments: 
   - 0
   - 0
-  - 2
-  version: 0.0.2
+  - 3
+  version: 0.0.3
 platform: ruby
 authors: 
 - Gerlando Piro
@@ -15,7 +15,7 @@ autorequire:
 bindir: bin
 cert_chain: []
 
-date: 2012-07-31 00:00:00 Z
+date: 2012-08-22 00:00:00 Z
 dependencies: 
 - !ruby/object:Gem::Dependency 
   name: nokogiri