jekyll-import 0.12.0 → 0.13.0

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

@@ -2,7 +2,7 @@ module JekyllImport
   module Importers
     class Typo < Importer
       # This SQL *should* work for both MySQL and PostgreSQL.
-      SQL = <<-EOS
+      SQL = <<-EOS.freeze
       SELECT c.id id,
              c.title title,
              c.permalink slug,
@@ -18,42 +18,44 @@ module JekyllImport
       EOS
 
       def self.require_deps
-        JekyllImport.require_with_fallback(%w[
+        JekyllImport.require_with_fallback(%w(
           rubygems
           sequel
+          mysql2
+          pg
           fileutils
           safe_yaml
-        ])
+        ))
       end
 
       def self.specify_options(c)
-        c.option 'server', '--server TYPE', 'Server type ("mysql" or "postgres")'
-        c.option 'dbname', '--dbname DB', 'Database name'
-        c.option 'user', '--user USER', 'Database user name'
-        c.option 'password', '--password PW', "Database user's password (default: '')"
-        c.option 'host', '--host HOST', 'Database host name'
+        c.option "server", "--server TYPE", 'Server type ("mysql" or "postgres")'
+        c.option "dbname", "--dbname DB", "Database name"
+        c.option "user", "--user USER", "Database user name"
+        c.option "password", "--password PW", "Database user's password (default: '')"
+        c.option "host", "--host HOST", "Database host name"
       end
 
       def self.process(options)
-        server = options.fetch('server')
-        dbname = options.fetch('dbname')
-        user   = options.fetch('user')
-        pass   = options.fetch('password', '')
-        host   = options.fetch('host', "localhost")
+        server = options.fetch("server")
+        dbname = options.fetch("dbname")
+        user   = options.fetch("user")
+        pass   = options.fetch("password", "")
+        host   = options.fetch("host", "localhost")
 
-        FileUtils.mkdir_p '_posts'
+        FileUtils.mkdir_p "_posts"
         case server.intern
         when :postgres
-          db = Sequel.postgres(dbname, :user => user, :password => pass, :host => host, :encoding => 'utf8')
+          db = Sequel.postgres(dbname, :user => user, :password => pass, :host => host, :encoding => "utf8")
         when :mysql
-          db = Sequel.mysql(dbname, :user => user, :password => pass, :host => host, :encoding => 'utf8')
+          db = Sequel.mysql2(dbname, :user => user, :password => pass, :host => host, :encoding => "utf8")
         else
           raise "Unknown database server '#{server}'"
         end
         db[SQL].each do |post|
-          next unless post[:state] =~ /published/i
+          next unless post[:state] =~ %r!published!i
 
-          if post[:slug] == nil
+          if post[:slug].nil?
             post[:slug] = "no slug"
           end
 
@@ -62,27 +64,25 @@ module JekyllImport
             post[:body] << post[:extended]
           end
 
-          name = [ sprintf("%.04d", post[:date].year),
-                   sprintf("%.02d", post[:date].month),
-                   sprintf("%.02d", post[:date].day),
-                   post[:slug].strip ].join('-')
+          name = [ format("%.04d", post[:date].year),
+                   format("%.02d", post[:date].month),
+                   format("%.02d", post[:date].day),
+                   post[:slug].strip, ].join("-")
 
           # Can have more than one text filter in this field, but we just want
           # the first one for this.
-          name += '.' + post[:filter].split(' ')[0]
+          name += "." + post[:filter].split(" ")[0]
 
-          File.open("_posts/#{name}", 'w') do |f|
-            f.puts({ 'layout'   => 'post',
-                     'title'    => (post[:title] and post[:title].to_s.force_encoding('UTF-8')),
-                     'tags'     => (post[:keywords] and post[:keywords].to_s.force_encoding('UTF-8')),
-                     'typo_id'  => post[:id]
-                   }.delete_if { |k, v| v.nil? || v == '' }.to_yaml)
-            f.puts '---'
+          File.open("_posts/#{name}", "w") do |f|
+            f.puts({ "layout"  => "post",
+                     "title"   => (post[:title] && post[:title].to_s.force_encoding("UTF-8")),
+                     "tags"    => (post[:keywords] && post[:keywords].to_s.force_encoding("UTF-8")),
+                     "typo_id" => post[:id], }.delete_if { |_k, v| v.nil? || v == "" }.to_yaml)
+            f.puts "---"
             f.puts post[:body].delete("\r")
           end
         end
       end
-
     end
   end
 end

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

@@ -1,32 +1,32 @@
 module JekyllImport
   module Importers
     class WordPress < Importer
-
       def self.require_deps
-        JekyllImport.require_with_fallback(%w[
+        JekyllImport.require_with_fallback(%w(
           rubygems
           sequel
           fileutils
           safe_yaml
           unidecode
-        ])
+        ))
       end
 
       def self.specify_options(c)
-        c.option 'dbname', '--dbname DB', 'Database name (default: "")'
-        c.option 'socket', '--socket SOCKET', 'Database socket (default: "")'
-        c.option 'user', '--user USER', 'Database user name (default: "")'
-        c.option 'password', '--password PW', "Database user's password (default: "")"
-        c.option 'host', '--host HOST', 'Database host name (default: "localhost")'
-        c.option 'table_prefix', '--table_prefix PREFIX', 'Table prefix name (default: "wp_")'
-        c.option 'site_prefix', '--site_prefix PREFIX', 'Site prefix name (default: "")'
-        c.option 'clean_entities', '--clean_entities', 'Whether to clean entities (default: true)'
-        c.option 'comments', '--comments', 'Whether to import comments (default: true)'
-        c.option 'categories', '--categories', 'Whether to import categories (default: true)'
-        c.option 'tags', '--tags', 'Whether to import tags (default: true)'
-        c.option 'more_excerpt', '--more_excerpt', 'Whether to use more excerpt (default: true)'
-        c.option 'more_anchor', '--more_anchor', 'Whether to use more anchor (default: true)'
-        c.option 'status', '--status STATUS,STATUS2', Array, 'Array of allowed statuses (default: ["publish"], other options: "draft", "private", "revision")'
+        c.option "dbname", "--dbname DB", "Database name (default: '')"
+        c.option "socket", "--socket SOCKET", "Database socket (default: '')"
+        c.option "user", "--user USER", "Database user name (default: '')"
+        c.option "password", "--password PW", "Database user's password (default: '')"
+        c.option "host", "--host HOST", "Database host name (default: 'localhost')"
+        c.option "port", "--port PORT", "Database port number (default: '')"
+        c.option "table_prefix", "--table_prefix PREFIX", "Table prefix name (default: 'wp_')"
+        c.option "site_prefix", "--site_prefix PREFIX", "Site prefix name (default: '')"
+        c.option "clean_entities", "--clean_entities", "Whether to clean entities (default: true)"
+        c.option "comments", "--comments", "Whether to import comments (default: true)"
+        c.option "categories", "--categories", "Whether to import categories (default: true)"
+        c.option "tags", "--tags", "Whether to import tags (default: true)"
+        c.option "more_excerpt", "--more_excerpt", "Whether to use more excerpt (default: true)"
+        c.option "more_anchor", "--more_anchor", "Whether to use more anchor (default: true)"
+        c.option "status", "--status STATUS,STATUS2", Array, "Array of allowed statuses (default: ['publish'], other options: 'draft', 'private', 'revision')"
       end
 
       # Main migrator function. Call this to perform the migration.
@@ -35,6 +35,7 @@ module JekyllImport
       # user::    The database user name
       # pass::    The database user's password
       # host::    The address of the MySQL database host. Default: 'localhost'
+      # port::    The port number of the MySQL database. Default: '3306'
       # socket::  The database socket's path
       # options:: A hash table of configuration options.
       #
@@ -74,28 +75,29 @@ module JekyllImport
       #
       def self.process(opts)
         options = {
-          :user           => opts.fetch('user', ''),
-          :pass           => opts.fetch('password', ''),
-          :host           => opts.fetch('host', 'localhost'),
-          :socket         => opts.fetch('socket', nil),
-          :dbname         => opts.fetch('dbname', ''),
-          :table_prefix   => opts.fetch('table_prefix', 'wp_'),
-          :site_prefix    => opts.fetch('site_prefix', nil),
-          :clean_entities => opts.fetch('clean_entities', true),
-          :comments       => opts.fetch('comments', true),
-          :categories     => opts.fetch('categories', true),
-          :tags           => opts.fetch('tags', true),
-          :more_excerpt   => opts.fetch('more_excerpt', true),
-          :more_anchor    => opts.fetch('more_anchor', true),
-          :extension      => opts.fetch('extension', 'html'),
-          :status         => opts.fetch('status', ['publish']).map(&:to_sym) # :draft, :private, :revision
+          :user           => opts.fetch("user", ""),
+          :pass           => opts.fetch("password", ""),
+          :host           => opts.fetch("host", "localhost"),
+          :port           => opts.fetch("port", "3306"),
+          :socket         => opts.fetch("socket", nil),
+          :dbname         => opts.fetch("dbname", ""),
+          :table_prefix   => opts.fetch("table_prefix", "wp_"),
+          :site_prefix    => opts.fetch("site_prefix", nil),
+          :clean_entities => opts.fetch("clean_entities", true),
+          :comments       => opts.fetch("comments", true),
+          :categories     => opts.fetch("categories", true),
+          :tags           => opts.fetch("tags", true),
+          :more_excerpt   => opts.fetch("more_excerpt", true),
+          :more_anchor    => opts.fetch("more_anchor", true),
+          :extension      => opts.fetch("extension", "html"),
+          :status         => opts.fetch("status", ["publish"]).map(&:to_sym) # :draft, :private, :revision
         }
 
         if options[:clean_entities]
           begin
-            require 'htmlentities'
+            require "htmlentities"
           rescue LoadError
-            STDERR.puts "Could not require 'htmlentities', so the " +
+            STDERR.puts "Could not require 'htmlentities', so the " \
                         ":clean_entities option is now disabled."
             options[:clean_entities] = false
           end
@@ -104,8 +106,13 @@ module JekyllImport
         FileUtils.mkdir_p("_posts")
         FileUtils.mkdir_p("_drafts") if options[:status].include? :draft
 
-        db = Sequel.mysql2(options[:dbname], :user => options[:user], :password => options[:pass],
-                          :socket => options[:socket], :host => options[:host], :encoding => 'utf8')
+        db = Sequel.mysql2(options[:dbname],
+            :user     => options[:user],
+            :password => options[:pass],
+            :socket   => options[:socket],
+            :host     => options[:host],
+            :port     => options[:port],
+            :encoding => "utf8")
 
         px = options[:table_prefix]
         sx = options[:site_prefix]
@@ -122,12 +129,12 @@ module JekyllImport
            WHERE posts.post_type = 'page'"
 
         db[page_name_query].each do |page|
-          if !page[:slug] or page[:slug].empty?
+          if !page[:slug] || page[:slug].empty?
             page[:slug] = sluggify(page[:title])
           end
           page_name_list[ page[:id] ] = {
             :slug   => page[:slug],
-            :parent => page[:parent]
+            :parent => page[:parent],
           }
         end
 
@@ -152,13 +159,13 @@ module JekyllImport
              LEFT JOIN #{px}users AS `users`
                ON posts.post_author = users.ID"
 
-        if options[:status] and not options[:status].empty?
+        if options[:status] && !options[:status].empty?
           status = options[:status][0]
           posts_query << "
-           WHERE posts.post_status = '#{status.to_s}'"
+           WHERE posts.post_status = '#{status}'"
           options[:status][1..-1].each do |status|
             posts_query << " OR
-             posts.post_status = '#{status.to_s}'"
+             posts.post_status = '#{status}'"
           end
         end
 
@@ -167,7 +174,6 @@ module JekyllImport
         end
       end
 
-
       def self.process_post(post, db, options, page_name_list)
         px = options[:table_prefix]
         sx = options[:site_prefix]
@@ -179,13 +185,12 @@ module JekyllImport
         end
 
         slug = post[:slug]
-        if !slug or slug.empty?
+        if !slug || slug.empty?
           slug = sluggify(title)
         end
 
         date = post[:date] || Time.now
-        name = "%02d-%02d-%02d-%s.%s" % [date.year, date.month, date.day,
-                                         slug, extension]
+        name = format("%02d-%02d-%02d-%s.%s", date.year, date.month, date.day, slug, extension)
         content = post[:content].to_s
         if options[:clean_entities]
           content = clean_entities(content)
@@ -193,17 +198,17 @@ module JekyllImport
 
         excerpt = post[:excerpt].to_s
 
-        more_index = content.index(/<!-- *more *-->/)
+        more_index = content.index(%r/<!-- *more* -->/)
         more_anchor = nil
         if more_index
-          if options[:more_excerpt] and
-              (post[:excerpt].nil? or post[:excerpt].empty?)
+          if options[:more_excerpt] &&
+              (post[:excerpt].nil? || post[:excerpt].empty?)
             excerpt = content[0...more_index]
           end
           if options[:more_anchor]
-            more_link = "more"
-            content.sub!(/<!-- *more *-->/,
-                         "<a id=\"more\"></a>" +
+            more_anchor = "more"
+            content.sub!(%r/<!-- *more *-->/,
+                         "<a id=\"more\"></a>" \
                          "<a id=\"more-#{post[:id]}\"></a>")
           end
         end
@@ -211,7 +216,7 @@ module JekyllImport
         categories = []
         tags = []
 
-        if options[:categories] or options[:tags]
+        if options[:categories] || options[:tags]
 
           cquery =
             "SELECT
@@ -227,13 +232,13 @@ module JekyllImport
                terms.term_id = ttax.term_id"
 
           db[cquery].each do |term|
-            if options[:categories] and term[:type] == "category"
+            if options[:categories] && term[:type] == "category"
               if options[:clean_entities]
                 categories << clean_entities(term[:name])
               else
                 categories << term[:name]
               end
-            elsif options[:tags] and term[:type] == "post_tag"
+            elsif options[:tags] && term[:type] == "post_tag"
               if options[:clean_entities]
                 tags << clean_entities(term[:name])
               else
@@ -245,7 +250,7 @@ module JekyllImport
 
         comments = []
 
-        if options[:comments] and post[:comment_count].to_i > 0
+        if options[:comments] && post[:comment_count].to_i > 0
           cquery =
             "SELECT
                comment_ID           AS `id`,
@@ -260,9 +265,7 @@ module JekyllImport
                comment_post_ID = '#{post[:id]}' AND
                comment_approved != 'spam'"
 
-
           db[cquery].each do |comment|
-
             comcontent = comment[:content].to_s
             if comcontent.respond_to?(:force_encoding)
               comcontent.force_encoding("UTF-8")
@@ -276,50 +279,50 @@ module JekyllImport
             end
 
             comments << {
-              'id'           => comment[:id].to_i,
-              'author'       => comauthor,
-              'author_email' => comment[:author_email].to_s,
-              'author_url'   => comment[:author_url].to_s,
-              'date'         => comment[:date].to_s,
-              'date_gmt'     => comment[:date_gmt].to_s,
-              'content'      => comcontent,
+              "id"           => comment[:id].to_i,
+              "author"       => comauthor,
+              "author_email" => comment[:author_email].to_s,
+              "author_url"   => comment[:author_url].to_s,
+              "date"         => comment[:date].to_s,
+              "date_gmt"     => comment[:date_gmt].to_s,
+              "content"      => comcontent,
             }
           end
 
-          comments.sort!{ |a,b| a['id'] <=> b['id'] }
+          comments.sort! { |a, b| a["id"] <=> b["id"] }
         end
 
         # Get the relevant fields as a hash, delete empty fields and
         # convert to YAML for the header.
         data = {
-          'layout'        => post[:type].to_s,
-          'status'        => post[:status].to_s,
-          'published'     => post[:status].to_s == 'draft' ? nil : (post[:status].to_s == 'publish'),
-          'title'         => title.to_s,
-          'author'        => {
-            'display_name'=> post[:author].to_s,
-            'login'       => post[:author_login].to_s,
-            'email'       => post[:author_email].to_s,
-            'url'  => post[:author_url].to_s,
+          "layout"        => post[:type].to_s,
+          "status"        => post[:status].to_s,
+          "published"     => post[:status].to_s == "draft" ? nil : (post[:status].to_s == "publish"),
+          "title"         => title.to_s,
+          "author"        => {
+            "display_name" => post[:author].to_s,
+            "login"        => post[:author_login].to_s,
+            "email"        => post[:author_email].to_s,
+            "url"          => post[:author_url].to_s,
           },
-          'author_login'  => post[:author_login].to_s,
-          'author_email'  => post[:author_email].to_s,
-          'author_url'    => post[:author_url].to_s,
-          'excerpt'       => excerpt,
-          'more_anchor'   => more_anchor,
-          'wordpress_id'  => post[:id],
-          'wordpress_url' => post[:guid].to_s,
-          'date'          => date.to_s,
-          'date_gmt'      => post[:date_gmt].to_s,
-          'categories'    => options[:categories] ? categories : nil,
-          'tags'          => options[:tags] ? tags : nil,
-          'comments'      => options[:comments] ? comments : nil,
-        }.delete_if { |k,v| v.nil? || v == '' }.to_yaml
-
-        if post[:type] == 'page'
+          "author_login"  => post[:author_login].to_s,
+          "author_email"  => post[:author_email].to_s,
+          "author_url"    => post[:author_url].to_s,
+          "excerpt"       => excerpt,
+          "more_anchor"   => more_anchor,
+          "wordpress_id"  => post[:id],
+          "wordpress_url" => post[:guid].to_s,
+          "date"          => date.to_s,
+          "date_gmt"      => post[:date_gmt].to_s,
+          "categories"    => options[:categories] ? categories : nil,
+          "tags"          => options[:tags] ? tags : nil,
+          "comments"      => options[:comments] ? comments : nil,
+        }.delete_if { |_k, v| v.nil? || v == "" }.to_yaml
+
+        if post[:type] == "page"
           filename = page_path(post[:id], page_name_list) + "index.#{extension}"
           FileUtils.mkdir_p(File.dirname(filename))
-        elsif post[:status] == 'draft'
+        elsif post[:status] == "draft"
           filename = "_drafts/#{slug}.md"
         else
           filename = "_posts/#{name}"
@@ -333,8 +336,7 @@ module JekyllImport
         end
       end
 
-
-      def self.clean_entities( text )
+      def self.clean_entities(text)
         if text.respond_to?(:force_encoding)
           text.force_encoding("UTF-8")
         end
@@ -346,27 +348,25 @@ module JekyllImport
         text.gsub!("&gt;", ">")
         text.gsub!("&quot;", '"')
         text.gsub!("&apos;", "'")
-        text.gsub!("/", "&#47;")
+        text.gsub!("&#47;", "/")
         text
       end
 
-
-      def self.sluggify( title )
-        title = title.to_ascii.downcase.gsub(/[^0-9A-Za-z]+/, " ").strip.gsub(" ", "-")
+      def self.sluggify(title)
+        title.to_ascii.downcase.gsub(%r![^0-9A-Za-z]+!, " ").strip.tr(" ", "-")
       end
 
-      def self.page_path( page_id, page_name_list )
+      def self.page_path(page_id, page_name_list)
         if page_name_list.key?(page_id)
           [
-            page_path(page_name_list[page_id][:parent],page_name_list),
+            page_path(page_name_list[page_id][:parent], page_name_list),
             page_name_list[page_id][:slug],
-            '/'
+            "/",
           ].join("")
         else
           ""
         end
       end
-
     end
   end
 end