web_stat 0.3.6 → 0.3.11

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 2ed82b635d751fd0e40eeeef77348ebfc02636f1a9f2d86ca0018c572c4b1f2d
-  data.tar.gz: 5034dcc4ba993f6d084228eb3e7e12fe03b92e8ff9e0119d3db8f7ad81521f4a
+  metadata.gz: 7a94017f641fb1f84d67ea5d650c094df5a1d738ccbbf880c112f95da4a2792b
+  data.tar.gz: 7383ba299f9f02bae998a104e84302ea99a11a28c77ad42fa23637eec9e2a922
 SHA512:
-  metadata.gz: d1a20f17519854baf5cea62ff55236b0ea63f9ab61b3c97df2f276ad7e64694d7d6387824957bee717c50a042734b37dcd26a6761591fa3d6550de8266454ac6
-  data.tar.gz: 1669e18b93b126bc3d382eba1e25de554fc23751f0d51635ec42c89fa8c05392d61901bb4fa510e27b1048133e96dba153e213cab3c3de90a2af1461cca88bf2
+  metadata.gz: 4c7d593118d75755f3db68a9f7995de60c63c5c47ba92c30a4772bc6ade5cefda14c8b44a87475ebc6cb3ab1a0badefcd3df1f6d2038aec714af654a07c89ccd
+  data.tar.gz: 2aa7dc288acd6de6207ee8d3e2833ba2b480b99a6d3ffeed651cce6982bbb557adfe173dfd3af1d10345bb0ad4710f1e4b434272039bb8a988beabb35811c5b4

data/Gemfile.lock

@@ -1,7 +1,7 @@
 PATH
   remote: .
   specs:
-    web_stat (0.3.5)
+    web_stat (0.3.10)
       bundler (>= 2.0.2)
       cld (>= 0.8.0)
       mechanize (>= 2.7)

data/lib/helpers/web_drive_helper.rb

@@ -5,14 +5,29 @@ module WebStat
       # @param [String] url
       # @param [Integer] delay
       def get_last_url(url, delay=nil)
+        driver = get_driver(url, delay)
+        last_url = driver.current_url
+        driver.quit
+        last_url
+      end
+      # Get source of html
+      # @param [String] url
+      # @param [Integer] delay
+      def get_source(url, delay=nil)
+        driver = get_driver(url, delay)
+        source = driver.page_source
+        driver.quit
+        source
+      end
+      
+      private
+      def get_driver(url, delay=nil)
         Selenium::WebDriver.logger.output = File.join("/tmp", "selenium.log")
         Selenium::WebDriver.logger.level = :info
         options = Selenium::WebDriver::Chrome::Options.new(args: [
           'headless',
           'no-sandbox',
-          'disable-gpu',
-          'start-maximized',
-          'window-size=1920,1080'
+          'disable-gpu'
           ])
         driver = Selenium::WebDriver.for(:chrome, options: options)
         driver.manage.timeouts.implicit_wait = 10
@@ -21,9 +36,10 @@ module WebStat
         if delay.is_a?(Integer)
           sleep delay
         end
-        last_url = driver.current_url
+        driver
+      rescue => e
         driver.quit
-        last_url
+        raise e
       end
     end
   end

data/lib/web_stat/fetch.rb

@@ -47,12 +47,9 @@ module WebStat
           break
         end
       end
-      if path.nil?
+      if path.nil? || path.empty? || @nokogiri.at('body').xpath('//img').first
         path = @nokogiri.at('body').xpath('//img').first.attr('src')
       end
-      if path.nil?
-        Readability::Document.new(@nokogiri.at('body')).content
-      end
       if ! path.nil? && path.match(/^\//)
         "#{URI.parse(@url).scheme}://#{URI.parse(@url).host}#{path}"
       else
@@ -81,17 +78,25 @@ module WebStat
     # @param [String] url
     # @param [String] body
     def get_url(url)
-      agent = Mechanize.new { |_agent| _agent.user_agent = WebStat::Configure.get["user_agent"] }
+      mech = Mechanize.new { |_mech| _mech.user_agent = WebStat::Configure.get["user_agent"] }
       # Enable to read Robots.txt
-      agent.robots = true
+      mech.robots = true
       begin
-        document = agent.get(url, [], nil, { 'Accept-Language' => 'ja'})
-        if document.class == Mechanize::File
-          body = document.body
+        if mech.agent.robots_disallowed?(url)
+          raise Mechanize::RobotsDisallowedError.new(url)
+        end
+        if WebStat::Configure.get["use_chromedirver"]
+          body = WebStat::WebDriverHelper.get_source(url)
+          @status = 200
         else
-          body = document.body.encode('UTF-8', document.encoding)
+          document = mech.get(url, [], nil, { 'Accept-Language' => 'ja'})
+          if document.class == Mechanize::File
+            body = document.body
+          else
+            body = document.body.encode('UTF-8', document.encoding)
+          end
+          @status = document.code
         end
-        @status = document.code
       rescue Mechanize::ResponseCodeError => e
         body = e.page.body
         @status = e.page.code

data/lib/web_stat/version.rb

@@ -1,3 +1,3 @@
 module WebStat
-  VERSION = "0.3.6"
+  VERSION = "0.3.11"
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: web_stat
 version: !ruby/object:Gem::Version
-  version: 0.3.6
+  version: 0.3.11
 platform: ruby
 authors:
 - yusuke abe
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2020-06-14 00:00:00.000000000 Z
+date: 2020-06-21 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler