Updating with more rake tweaks and some post customizations

README.rdoc

@@ -1,8 +1,10 @@
 = Wordpress-import
 
-Fork of Marc Remolt's Refinerycms-wordpress-import ( https://github.com/mremolt/refinerycms-wordpress-import )
+This little project is an importer for WordPress XML dumps into Rails.
 
-This little project is an importer for WordPress XML dumps into Rails
+It's been somewhat customized for one particular project; you probably want to fork this and modify it to fit your app's schema.
+
+It's a fork of Marc Remolt's Refinerycms-wordpress-import ( https://github.com/mremolt/refinerycms-wordpress-import )
 
 You can find the source code on github: https://github.com/zyphlar/wordpress-import
 
@@ -11,11 +13,6 @@ If your site (blog) structure uses new urls, the links WILL break! For example,
 the popular WP blog url structure "YYYY-MM/slug", be warned that Refinery just uses "blog/slug". 
 So your inner site links will point to the old WP url. 
 
-== TODO
-- tags
-- attachments
-  - what about wordpress suffixes like -300x290 ?
-
 == Prerequisites
 
 TODO
@@ -29,7 +26,7 @@ Just add the gem to your projects Gemfile:
 
 Or if you want to stay on the bleeding edge: 
 
-  gem 'wordpress-import', :git => 'git://github.com/zyphlarz/wordpress-import.git'
+  gem 'wordpress-import', :git => 'git://github.com/zyphlar/wordpress-import.git'
 
 and run
 

lib/tasks/wordpress.rake

@@ -13,9 +13,9 @@ namespace :wordpress do
   end
 
   desc "import blog data from a WordPressImport XML dump"
-  task :import_blog, :file_name do |task, params|
+  task :import_blog, :file_name, :blog_slug do |task, params|
     Rake::Task["environment"].invoke
-    p "Loading XML from #{params[:file_name]} ..."
+    p "Loading XML from #{params[:file_name]} (using blog #{params[:blog_slug]}) ..."
     dump = WordPressImport::Dump.new(params[:file_name])
 
     p "Importing #{dump.authors.count} authors ..."
@@ -31,14 +31,14 @@ namespace :wordpress do
       p "(export ONLY_PUBLISHED=true to import only published posts)"
     end
 
-    dump.posts(only_published).each(&:to_rails)
+    dump.posts(only_published).each{|p| p.to_rails(params[:blog_slug]) }
   end
 
   desc "reset blog tables and then import blog data from a WordPressImport XML dump"
-  task :reset_and_import_blog, :file_name do |task, params|
+  task :reset_and_import_blog, :file_name, :blog_slug do |task, params|
     Rake::Task["environment"].invoke
     Rake::Task["wordpress:reset_blog"].invoke
-    Rake::Task["wordpress:import_blog"].invoke(params[:file_name])
+    Rake::Task["wordpress:import_blog"].invoke(params[:file_name], params[:blog_slug])
   end
 
 
@@ -129,19 +129,12 @@ namespace :wordpress do
   end
 
   desc "reset and import all data (see the other tasks)"
-  task :full_import, :file_name do |task, params|
+  task :full_import, :file_name, :blog_slug do |task, params|
     Rake::Task["environment"].invoke
-    Rake::Task["wordpress:reset_and_import_blog"].invoke(params[:file_name])
+    Rake::Task["wordpress:reset_and_import_blog"].invoke(params[:file_name],params[:blog_slug])
     #Rake::Task["wordpress:reset_and_import_pages"].invoke(params[:file_name])
     #Rake::Task["wordpress:reset_import_and_replace_media"].invoke(params[:file_name])
     Rake::Task["wordpress:import_and_replace_media"].invoke(params[:file_name])
   end
 
-
-  desc "Local First master import (no resets)"
-  task :lfa_import, :file_name do |task, params|
-    Rake::Task["environment"].invoke
-    Rake::Task["wordpress:import_blog"].invoke(params[:file_name])
-    Rake::Task["wordpress:import_and_replace_media"].invoke(params[:file_name])
-  end
 end

lib/wordpress/author.rb

@@ -38,6 +38,9 @@ module WordPressImport
         user.password_confirmation = 'password'
       end
       user.save
+
+      puts "User #{login} imported."
+
       user
     end
   end

lib/wordpress/page.rb

@@ -13,6 +13,10 @@ module WordPressImport
       "WordPress::Page(#{post_id}): #{title}"     
     end
 
+    def link
+      node.xpath("link").text
+    end
+
     def title
       node.xpath("title").text
     end

lib/wordpress/post.rb

@@ -29,7 +29,8 @@ module WordPressImport
       end
     end
 
-    def to_rails
+    # blog_slug is used to identify which blog this import is from
+    def to_rails(blog_slug)
 
       user = ::User.find_by_wp_username(creator)
 
@@ -37,34 +38,24 @@ module WordPressImport
         raise "User with wp_username #{creator} not found"
       end
 
-      post = ::Post.find_or_initialize_by(:id => post_id, :slug => post_name)
-
-      post.assign_attributes( 
+      post = ::Post.create({ 
+        :wp_post_id => post_id, :slug => post_name,
         :user_id => user.id, :title => title, 
         :created_at => post_date, 
-        :published_at => publish_date)
-      # :body => content_formatted taken care of by translation below
-
-      if post.translations.blank?
-        translation = post.translations.build
-      else
-        translation = post.translations.first
-      end
-      
-      translation.locale = "en"
-      translation.title = title
-      translation.body = content_formatted
-
-      # merge the translation's category list with the wordpress post's
-      translation.category_list |= categories.collect(&:name)
-      # and tags
-      translation.category_list |= tags.collect(&:name)
-
-      translation.save
-      
-      post.save
+        :published_at => publish_date,
+        :wp_link => link,
+        :wp_blog => blog_slug,
+        :translations_attributes => { "0" => {
+            :locale => "en",
+            :title => title,
+            :body => content_formatted,
+            # merge the translation's category list with the wordpress post's
+            :category_list => categories.collect(&:name) | tags.collect(&:name) 
+          }}
+        })
 
       if post.errors.blank?
+        puts "Post #{post_name} imported."
         return post.reload
       else
         puts post.inspect