jekyll-import 0.4.1 → 0.5.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 3ea4fc21b7d8bde846159c3822251b2192274fa8
-  data.tar.gz: 2ab2c876e9b3e440eba999232d45665f0ba7ff7d
+  metadata.gz: b24802c3b5f7734905bef82ba555c97b273b1139
+  data.tar.gz: 6c1da2b13f4727a49144d84f140cc1fe91151d7e
 SHA512:
-  metadata.gz: bbbb4b901182d17af3d0ed0432f6443b9051cd9d3976f95108931806c82fc98342218df4a6b3750705269a88b8bce73b4f236763d5e8713d8591366189b73666
-  data.tar.gz: 6f6a0fbcf76c5a05908e8e9aba0bc8f9e3a5c9cd832a671b20b6f32d4758f69f94ac96779aa53d0bff8c151a4923a7b32cb582b21d4d173e0456e587c99f237a
+  metadata.gz: 889cf9c39a18ff9c0903ec4e154683fb895315b3cc916a5fb44559f0ef1ef47e31a17086586f72f5a4bfc3ab409825698398577856085a71a8efe0c755181b96
+  data.tar.gz: a4432acb1e4c9ecbb47e53472fa48c7fdb411b5d399c2c91573b953250c87b50fe6817fe7b19ba4311115293e560d68b5aabedb119c01032af04e7fbc71bd73b

data/lib/jekyll-import/importers/blogger.rb

@@ -0,0 +1,251 @@
+module JekyllImport
+  module Importers
+    class Blogger < Importer
+      def self.specify_options(c)
+        c.option 'source', '--source NAME', 'The XML file (blog-MM-DD-YYYY.xml) path to import'
+        c.option 'no-blogger-info', '--no-blogger-info', 'not to leave blogger-URL info (id and old URL) in the front matter (default: false)'
+        c.option 'replace-internal-link', '--replace-internal-link', 'replace internal links using the post_url liquid tag. (default: false)'
+      end
+
+      def self.validate(options)
+        if options['source'].nil?
+          raise 'Missing mandatory option: --source'
+        elsif not File.exist?(options['source'])
+          raise Errno::ENOENT, "File not found: #{options['source']}"
+        end
+      end
+
+      def self.require_deps
+        JekyllImport.require_with_fallback(%w[
+          rexml/document
+          rexml/streamlistener
+          rexml/parsers/streamparser
+          uri
+          time
+          fileutils
+          safe_yaml
+        ])
+      end
+
+      # Process the import.
+      #
+      # source::                a local file String (or IO object for internal use purpose)..
+      # no-blogger-info::       a boolean if not leave blogger info (id and original URL).
+      # replace-internal-link:: a boolean if replace internal link
+      #
+      # Returns nothing.
+      def self.process(options)
+        source = options.fetch('source')
+
+        listener = BloggerAtomStreamListener.new
+
+        listener.leave_blogger_info = ! options.fetch('no-blogger-info', false),
+
+        File.open(source, 'r') do |f|
+          f.flock(File::LOCK_SH)
+          REXML::Parsers::StreamParser.new(f, listener).parse()
+        end
+
+        options['original-url-base'] = listener.original_url_base
+
+        postprocess(options)
+      end
+
+      # Post-process after import.
+      #
+      # replace-internal-link:: a boolean if replace internal link
+      #
+      # Returns nothing.
+      def self.postprocess(options)
+        # Replace internal link URL
+        if options.fetch('replace-internal-link', false)
+          original_url_base = options.fetch('original-url-base', nil)
+          if original_url_base
+            orig_url_pattern = Regexp.new(" href=([\"\'])(?:#{Regexp.escape(original_url_base)})?/([0-9]{4})/([0-9]{2})/([^\"\']+\.html)\\1")
+
+            Dir.glob('_posts/*.*') do |filename|
+              body = nil
+              File.open(filename, 'r') do |f|
+                f.flock(File::LOCK_SH)
+                body = f.read
+              end
+
+              body.gsub!(orig_url_pattern) do
+                # for post_url
+                quote = $1
+                post_file = Dir.glob("_posts/#{$2}-#{$3}-*-#{$4.to_s.tr('/', '-')}").first
+                raise "Could not found: _posts/#{$2}-#{$3}-*-#{$4.to_s.tr('/', '-')}" if post_file.nil?
+                " href=#{quote}{{ site.baseurl }}{% post_url #{File.basename(post_file, '.html')} %}#{quote}"
+              end
+
+              File.open(filename, 'w') do |f|
+                f.flock(File::LOCK_EX)
+                f << body
+              end
+            end
+          end
+        end
+      end
+
+      class BloggerAtomStreamListener
+        def initialize
+          # use `extend` instead of `include` to use `require_deps` instead of `require`.
+          extend REXML::StreamListener
+          extend BloggerAtomStreamListenerMethods
+
+          @leave_blogger_info = true
+        end
+      end
+
+      module BloggerAtomStreamListenerMethods
+        attr_accessor :leave_blogger_info
+        attr_reader :original_url_base
+      
+        def tag_start(tag, attrs)
+          @tag_bread = [] unless @tag_bread
+          @tag_bread.push(tag)
+      
+          case tag
+          when 'entry'
+            raise 'nest entry element' if @in_entry_elem
+            @in_entry_elem = {:meta => {}, :body => nil}
+          when 'title'
+            if @in_entry_elem
+              raise 'only <title type="text"></title> is supported' if attrs['type'] != 'text'
+            end
+          when 'category'
+            if @in_entry_elem
+              if attrs['scheme'] == 'http://www.blogger.com/atom/ns#'
+                @in_entry_elem[:meta][:category] = [] unless @in_entry_elem[:meta][:category]
+                @in_entry_elem[:meta][:category] << attrs['term']
+              elsif attrs['scheme'] == 'http://schemas.google.com/g/2005#kind'
+                kind = attrs['term']
+                kind.sub!(Regexp.new("^http://schemas\\.google\\.com/blogger/2008/kind\\#"), '')
+                @in_entry_elem[:meta][:kind] = kind
+              end
+            end
+          when 'content'
+            if @in_entry_elem
+              @in_entry_elem[:meta][:content_type] = attrs['type']
+            end
+          when 'link'
+            if @in_entry_elem
+              if attrs['rel'] == 'alternate' && attrs['type'] == 'text/html'
+                @in_entry_elem[:meta][:original_url] = attrs['href']
+              elsif attrs['rel'] == 'replies' && attrs['type'] == 'text/html'
+                unless @in_entry_elem[:meta][:original_url]
+                  @in_entry_elem[:meta][:original_url] = attrs['href'].sub(/\#comment-form$/, '')
+                end
+              end
+            end
+          when 'media:thumbnail'
+            if @in_entry_elem
+              @in_entry_elem[:meta][:thumbnail] = attrs['url']
+            end
+          end
+        end
+      
+        def text(text)
+          if @in_entry_elem
+            case @tag_bread.last
+            when 'id'
+              @in_entry_elem[:meta][:id] = text
+            when 'published'
+              @in_entry_elem[:meta][:published] = text
+            when 'updated'
+              @in_entry_elem[:meta][:updated] = text
+            when 'title'
+              @in_entry_elem[:meta][:title] = text
+            when 'content'
+              @in_entry_elem[:body] = text
+            when 'name'
+              if @tag_bread[-2..-1] == %w[author name]
+                @in_entry_elem[:meta][:author] = text
+              end
+            when 'app:draft'
+              if @tag_bread[-2..-1] == %w[app:control app:draft]
+                @in_entry_elem[:meta][:draft] = true if text == 'yes'
+              end
+            end
+          end
+        end 
+      
+        def tag_end(tag)
+          case tag
+          when 'entry'
+            raise 'nest entry element' unless @in_entry_elem
+      
+            if @in_entry_elem[:meta][:kind] == 'post'
+              post_data = get_post_data_from_in_entry_elem_info
+
+              if post_data
+                target_dir = '_posts'
+                target_dir = '_drafts' if @in_entry_elem[:meta][:draft]
+
+                FileUtils.mkdir_p(target_dir)
+      
+                File.open(File.join(target_dir, "#{post_data[:filename]}.html"), 'w') do |f|
+                  f.flock(File::LOCK_EX)
+      
+                  f << post_data[:header].to_yaml
+                  f << "---\n\n"
+                  f << post_data[:body]
+                end
+              end
+            end
+      
+            @in_entry_elem = nil
+          end
+      
+          @tag_bread.pop
+        end
+
+        def get_post_data_from_in_entry_elem_info
+          if (@in_entry_elem.nil? || ! @in_entry_elem.has_key?(:meta) || ! @in_entry_elem[:meta].has_key?(:kind))
+            nil
+          elsif @in_entry_elem[:meta][:kind] == 'post'
+            if @in_entry_elem[:meta][:original_url]
+              original_uri = URI.parse(@in_entry_elem[:meta][:original_url])
+              original_path = original_uri.path.to_s
+              filename = "%s-%s" %
+                [Time.parse(@in_entry_elem[:meta][:published]).strftime('%Y-%m-%d'),
+                 File.basename(original_path, File.extname(original_path))]
+
+              @original_url_base = "#{original_uri.scheme}://#{original_uri.host}"
+            else
+              raise 'Original URL is missing'
+            end
+        
+            header = {
+              'layout' => 'post',
+              'title' => @in_entry_elem[:meta][:title],
+              'date' => @in_entry_elem[:meta][:published],
+              'author' => @in_entry_elem[:meta][:author],
+              'tags' => @in_entry_elem[:meta][:category],
+            }
+            header['modified_time'] = @in_entry_elem[:meta][:updated] if @in_entry_elem[:meta][:updated] && @in_entry_elem[:meta][:updated] != @in_entry_elem[:meta][:published]
+            header['thumbnail'] = @in_entry_elem[:meta][:thumbnail] if @in_entry_elem[:meta][:thumbnail]
+            header['blogger_id'] = @in_entry_elem[:meta][:id] if @leave_blogger_info
+            header['blogger_orig_url'] = @in_entry_elem[:meta][:original_url] if @leave_blogger_info
+        
+            body = @in_entry_elem[:body]
+
+            # body escaping associated with liquid
+            if body =~ /{{/
+              body.gsub!(/{{/, '{{ "{{" }}')
+            end
+            if body =~ /{%/
+              body.gsub!(/{%/, '{{ "{%" }}')
+            end
+  
+            { :filename => filename, :header => header, :body => body }
+          else
+            nil
+          end
+        end
+
+      end
+
+    end
+  end
+end

data/lib/jekyll-import/importers/drupal6.rb

@@ -1,9 +1,3 @@
-# NOTE: This converter requires Sequel and the MySQL gems.
-# The MySQL gem can be difficult to install on OS X. Once you have MySQL
-# installed, running the following commands should work:
-# $ sudo gem install sequel
-# $ sudo gem install mysql -- --with-mysql-config=/usr/local/mysql/bin/mysql_config
-
 module JekyllImport
   module Importers
     class Drupal6 < Importer

data/lib/jekyll-import/importers/drupal7.rb

@@ -1,9 +1,3 @@
-# NOTE: This converter requires Sequel and the MySQL gems.
-# The MySQL gem can be difficult to install on OS X. Once you have MySQL
-# installed, running the following commands should work:
-# $ sudo gem install sequel
-# $ sudo gem install mysql -- --with-mysql-config=/usr/local/mysql/bin/mysql_config
-
 module JekyllImport
   module Importers
     class Drupal7 < Importer

data/lib/jekyll-import/importers/easyblog.rb

@@ -1,10 +1,3 @@
-# NOTE: This migrator is made for Joomla 2.5 database and EasyBlog 3.8.
-# NOTE: This converter requires Sequel and the MySQL gems.
-# The MySQL gem can be difficult to install on OS X. Once you have MySQL
-# installed, running the following commands should work:
-# $ sudo gem install sequel
-# $ sudo gem install mysql -- --with-mysql-config=/usr/local/mysql/bin/mysql_config
-
 module JekyllImport
   module Importers
     class Easyblog < Importer
@@ -15,7 +8,7 @@ module JekyllImport
           end
         end
       end
-      
+
       def self.specify_options(c)
         c.option 'dbname', '--dbname', 'Database name'
         c.option 'user', '--user', 'Database user name'
@@ -24,7 +17,7 @@ module JekyllImport
         c.option 'section', '--section', 'Table prefix name'
         c.option 'prefix', '--prefix', 'Table prefix name'
       end
-      
+
       def self.require_deps
         JekyllImport.require_with_fallback(%w[
                                            rubygems
@@ -33,7 +26,7 @@ module JekyllImport
                                           safe_yaml
                                           ])
       end
-      
+
       def self.process(options)
         dbname  = options.fetch('dbname')
         user    = options.fetch('user')
@@ -41,20 +34,20 @@ module JekyllImport
         host    = options.fetch('host', "localhost")
         section = options.fetch('section', '1')
         table_prefix = options.fetch('prefix', "jos_")
-        
+
         db = Sequel.mysql(dbname, :user => user, :password => pass, :host => host, :encoding => 'utf8')
-        
+
         FileUtils.mkdir_p("_posts")
-        
+
         # Reads a MySQL database via Sequel and creates a post file for each
         # post in wp_posts that has post_status = 'publish'. This restriction is
         # made because 'draft' posts are not guaranteed to have valid dates.
 
         query = "
-        select 
+        select
 	  ep.`title`, `permalink` as alias, concat(`intro`, `content`) as content, ep.`created`, ep.`id`, ec.`title` as category, tags
-        from 
-          #{table_prefix}easyblog_post ep 
+        from
+          #{table_prefix}easyblog_post ep
           left join #{table_prefix}easyblog_category ec on (ep.category_id = ec.id)
           left join (
             select
@@ -66,7 +59,7 @@ module JekyllImport
             group by
               ept.post_id) x on (ep.id = x.post_id);
         "
-        
+
         db[query].each do |post|
           # Get required fields and construct Jekyll compatible name.
           title = post[:title]
@@ -74,10 +67,10 @@ module JekyllImport
           date = post[:created]
           content = post[:content]
           category = post[:category]
-          tags = post[:tags]    
+          tags = post[:tags]
           name = "%02d-%02d-%02d-%s.markdown" % [date.year, date.month, date.day,
                                                  slug]
-          
+
           # Get the relevant fields as a hash, delete empty fields and convert
           # to YAML for the header.
           data = {
@@ -86,10 +79,10 @@ module JekyllImport
             'joomla_id' => post[:id],
             'joomla_url' => post[:alias],
             'category' => post[:category],
-            'tags' => post[:tags],     
+            'tags' => post[:tags],
             'date' => date
           }.delete_if { |k,v| v.nil? || v == '' }.to_yaml
-          
+
           # Write out the data and content to file
           File.open("_posts/#{name}", "w") do |f|
             f.puts data

data/lib/jekyll-import/importers/enki.rb

@@ -1,6 +1,3 @@
-# Adapted by Rodrigo Pinto <rodrigopqn@gmail.com>
-# Based on typo.rb by Toby DiPasquale
-
 module JekyllImport
     module Importers
     class Enki < Importer

data/lib/jekyll-import/importers/ghost.rb

@@ -1,4 +1,3 @@
-# NOTE: This migrator is made for Ghost sqlite databases.
 module JekyllImport
   module Importers
     class Ghost < Importer
@@ -38,7 +37,7 @@ module JekyllImport
         # detect if the post is a draft
         draft = post[:status].eql?('draft')
 
-        # Ghost saves the time in an weird format with 3 more numbers. 
+        # Ghost saves the time in an weird format with 3 more numbers.
         # But the time is correct when we remove the last 3 numbers.
         date = Time.at(post[:created_at].to_i.to_s[0..-4].to_i)
 

data/lib/jekyll-import/importers/joomla.rb

@@ -1,10 +1,3 @@
-# NOTE: This migrator is made for Joomla 1.5 databases.
-# NOTE: This converter requires Sequel and the MySQL gems.
-# The MySQL gem can be difficult to install on OS X. Once you have MySQL
-# installed, running the following commands should work:
-# $ sudo gem install sequel
-# $ sudo gem install mysql -- --with-mysql-config=/usr/local/mysql/bin/mysql_config
-
 module JekyllImport
   module Importers
     class Joomla < Importer

data/lib/jekyll-import/importers/jrnl.rb

@@ -1,5 +1,3 @@
-# Author: Aniket Pant <me@aniketpant.com>
-
 module JekyllImport
   module Importers
     class Jrnl < Importer

data/lib/jekyll-import/importers/mephisto.rb

@@ -1,13 +1,3 @@
-# Quickly hacked together my Michael Ivey
-# Based on mt.rb by Nick Gerakines, open source and publically
-# available under the MIT license. Use this module at your own risk.
-
-# NOTE: This converter requires Sequel and the MySQL gems.
-# The MySQL gem can be difficult to install on OS X. Once you have MySQL
-# installed, running the following commands should work:
-# $ sudo gem install sequel
-# $ sudo gem install mysql -- --with-mysql-config=/usr/local/mysql/bin/mysql_config
-
 module JekyllImport
   module Importers
     class Mephisto < Importer

data/lib/jekyll-import/importers/mt.rb

@@ -1,13 +1,3 @@
-# Created by Nick Gerakines, open source and publically available under the
-# MIT license. Use this module at your own risk.
-# I'm an Erlang/Perl/C++ guy so please forgive my dirty ruby.
-
-# NOTE: This converter requires Sequel and the MySQL gems.
-# The MySQL gem can be difficult to install on OS X. Once you have MySQL
-# installed, running the following commands should work:
-# $ sudo gem install sequel
-# $ sudo gem install mysql -- --with-mysql-config=/usr/local/mysql/bin/mysql_config
-
 module JekyllImport
   module Importers
     class MT < Importer

data/lib/jekyll-import/importers/posterous.rb

@@ -1,12 +1,3 @@
-# ruby -r './lib/jekyll/migrators/posterous.rb' -e 'Jekyll::Posterous.process(email, pass, api_key)'
-# Other arguments are optional; the default values are:
-# * :include_imgs => false         # should images be downloaded as well?
-# * :blog         => 'primary'     # blog, if you have more than one.
-# * :base_path    => '/'           # for image, if they will be served from a different host for eg.
-
-# For example, to download images as well as your posts, use the above command with
-# ....process(email, pass, api_key, :include_imgs => true)
-
 module JekyllImport
   module Importers
     class Posterous < Importer

data/lib/jekyll-import/importers/rss.rb

@@ -1,13 +1,3 @@
-# Created by Kendall Buchanan (https://github.com/kendagriff) on 2011-12-22.
-# Use at your own risk. The end.
-#
-# Usage:
-#   (URL)
-#   ruby -r 'jekyll/jekyll-import/rss' -e "JekyllImport::RSS.process(:source => 'http://yourdomain.com/your-favorite-feed.xml')"
-#
-#   (Local file)
-#   ruby -r 'jekyll/jekyll-import/rss' -e "JekyllImport::RSS.process(:source => './somefile/on/your/computer.xml')"
-
 module JekyllImport
   module Importers
     class RSS < Importer

data/lib/jekyll-import/importers/s9y.rb

@@ -1,10 +1,3 @@
-# Migrator to import entries from an Serendipity (S9Y) blog
-#
-# Entries can be exported from http://blog.example.com/rss.php?version=2.0&all=1
-#
-# Usage:
-# ruby -r './s9y_rss.rb' -e 'Jekyll::S9Y.process("http://blog.example.com/rss.php?version=2.0&all=1")'
-
 module JekyllImport
   module Importers
     class S9Y < Importer

data/lib/jekyll-import/importers/textpattern.rb

@@ -1,9 +1,3 @@
-# NOTE: This converter requires Sequel and the MySQL gems.
-# The MySQL gem can be difficult to install on OS X. Once you have MySQL
-# installed, running the following commands should work:
-# $ sudo gem install sequel
-# $ sudo gem install mysql -- --with-mysql-config=/usr/local/mysql/bin/mysql_config
-
 module JekyllImport
   module Importers
     class TextPattern < Importer

data/lib/jekyll-import/importers/typo.rb

@@ -1,4 +1,3 @@
-# Author: Toby DiPasquale <toby@cbcg.net>
 module JekyllImport
   module Importers
     class Typo < Importer

data/lib/jekyll-import/importers/wordpress.rb

@@ -1,12 +1,3 @@
-# NOTE: This converter requires Sequel and the MySQL gems.
-# The MySQL gem can be difficult to install on OS X. Once you have MySQL
-# installed, running the following commands should work:
-# $ sudo gem install sequel
-# $ sudo gem install mysql -- --with-mysql-config=/usr/local/mysql/bin/mysql_config
-#
-# If you are running MacPorts, it the mysql2 gem can be installed like this:
-# $ gem install mysql2 -- --with-mysql-lib=/opt/local/lib/mysql56/mysql/ --with-mysql-include=/opt/local/include/mysql56/mysql
-
 module JekyllImport
   module Importers
     class WordPress < Importer

data/lib/jekyll-import/importers/wordpressdotcom.rb

@@ -1,6 +1,4 @@
-# coding: utf-8
-# This importer takes a wordpress.xml file, which can be exported from your
-# wordpress.com blog (/wp-admin/export.php).
+# encoding: UTF-8
 
 module JekyllImport
   module Importers
@@ -8,7 +6,6 @@ module JekyllImport
       def self.require_deps
         JekyllImport.require_with_fallback(%w[
           rubygems
-          sequel
           fileutils
           safe_yaml
           hpricot

data/lib/jekyll-import/version.rb

@@ -1,3 +1,3 @@
 module JekyllImport
-  VERSION = '0.4.1'
+  VERSION = '0.5.0'
 end

data/lib/jekyll/commands/import.rb

@@ -7,6 +7,7 @@ module Jekyll
     class Import < Command
 
       IMPORTERS = {
+        :blogger => 'Blogger',
         :behance => 'Behance',
         :csv => 'CSV',
         :drupal6 => 'Drupal6',
@@ -53,7 +54,11 @@ module Jekyll
           if IMPORTERS.keys.include?(migrator.to_sym)
             if JekyllImport::Importers.const_defined?(IMPORTERS[migrator.to_sym])
               klass = JekyllImport::Importers.const_get(IMPORTERS[migrator.to_sym])
-              klass.run(options.__hash__)
+              if options.respond_to?(:__hash__)
+                klass.run(options.__hash__)
+              else
+                klass.run(options)
+              end
             end
           else
             abort_on_invalid_migrator(migrator)
@@ -61,11 +66,9 @@ module Jekyll
         end
 
         def abort_on_invalid_migrator(migrator)
-          msg = "Sorry, '#{migrator}' isn't a valid migrator. Valid choices:\n"
-          IMPORTERS.keys.each do |k, v|
-            msg += "* #{k}\n"
-          end
-          abort msg
+          $stderr.puts "Sorry, '#{migrator}' isn't a valid migrator. Valid choices:"
+          IMPORTERS.keys.each { |k| $stderr.puts "* #{k}" }
+          raise RuntimeError.new("'#{migrator}' is not a valid migrator.")
         end
 
       end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: jekyll-import
 version: !ruby/object:Gem::Version
-  version: 0.4.1
+  version: 0.5.0
 platform: ruby
 authors:
 - Tom Preston-Werner
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2014-08-01 00:00:00.000000000 Z
+date: 2014-09-19 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: jekyll
@@ -304,6 +304,7 @@ files:
 - lib/jekyll-import/importer.rb
 - lib/jekyll-import/importers.rb
 - lib/jekyll-import/importers/behance.rb
+- lib/jekyll-import/importers/blogger.rb
 - lib/jekyll-import/importers/csv.rb
 - lib/jekyll-import/importers/drupal6.rb
 - lib/jekyll-import/importers/drupal7.rb