refactor logic of keywords into a view-based model.

app/controllers/pages_controller.rb

@@ -43,7 +43,7 @@ def imprint
   # This provide a first look to our metadata and give a hint about our data
   def data
     validate_sort_params
-    @tags = Dataset.tag_counts.order("tags.name")
+    @tags = DatasetTag.tag_counts
     @datasets = Dataset.joins_datafile_and_freeformats(params[:workbook]).select("datasets.id, title, 
       GREATEST(datasets.updated_at, max(freeformats.updated_at)) as last_update,
       count(datafiles.id)").order("#{params[:sort]} #{params[:direction]}")

app/controllers/tags_controller.rb

@@ -9,7 +9,8 @@ class TagsController < ApplicationController
   end
 
   def index
-    @tags = Dataset.tag_counts.where("name iLike ?", "%#{params[:q]}%").order("tags.name")
+    @tags = DatasetTag.tag_counts.where("name iLike ?", "%#{params[:q]}%")
+
     respond_to do |format|
       format.html
       format.json { render :json => @tags}
@@ -19,7 +20,10 @@ def index
 
   def show
     @tag = ActsAsTaggableOn::Tag.find(params[:id])
-    @datasets = Dataset.tag_usage.select("datasets.*").where("tags.id = ?", @tag.id).order("datasets.title")
+    @datasets = Dataset.joins(:dataset_tags)
+                       .select("datasets.id, title")
+                       .where(['dataset_tags.tag_id = ?', @tag.id])
+                       .order("lower(datasets.title)")
 
     respond_to do |format|
       format.html
@@ -30,7 +34,7 @@ def show
   end
 
   def manage
-    @all_tags = Dataset.tag_counts.where("name iLike ?", "%#{params[:search]}%").order("tags.name")
+    @all_tags = DatasetTag.tag_counts
   end
 
   def delete

app/models/dataset.rb

@@ -54,6 +54,9 @@ class Dataset < ActiveRecord::Base
   has_many :paperproposals, :through => :dataset_paperproposals
   has_many :proposers, :through => :paperproposals, :source => :author, :uniq => true
 
+  has_many :dataset_tags
+  has_many :all_tags, :through => :dataset_tags, :source => :tag, :order => 'lower(name)'
+
   validates :title, :presence => true, :uniqueness => { case_sensitive: false }
   ACCESS_CODES = {
     private: 0,
@@ -258,43 +261,18 @@ def to_csv (separate_category_columns = false)
     end
   end
 
-  def all_tags
-    Dataset.tag_usage.select("tags.*").where("dataset_id =  #{self.id}").order("tags.name")
-  end
-
-  # This method returns similar datasets which are sorted by similarity in descending order
+  # This method returns similar datasets which share keywords with current dataset.
+  # datasets are sorted by similarity in descending order
   def find_related_datasets
-    tags = self.all_tags.map(&:id)
+    tags = self.all_tags.pluck(:id)
     return [] if tags.empty?
-    datasets = Dataset.tag_usage.select("datasets.*,count(tags.*) as count").
-                    where(["tags.id in (?) and datasets.id <> ?", tags, self.id]).
-                    group("datasets.id").order("count(tags.*) desc")
+    datasets = Dataset.joins(:dataset_tags)
+                      .select("datasets.*, count(tag_id) as count")
+                      .where(["tag_id in (?) and datasets.id <> ?", tags, self.id])
+                      .group("datasets.id").order("count(tag_id) desc")
     return(datasets)
   end
 
-  def self.tag_counts
-    Dataset.tag_usage.select("tags.*, count(datasets.id) as count").group("tags.id")
-  end
-  def self.tag_usage
-    # Return a ActiveRecord::Relation object that can be reused by other methods
-    Dataset.joins("
-      join
-      (
-      select taggable_id as dataset_id, tag_id
-      from taggings
-      where taggable_type = 'Dataset'
-      union
-      select distinct d.dataset_id, g.tag_id
-      from taggings g join datacolumns d
-      on g.taggable_id = d.id
-      where g.taggable_type = 'Datacolumn'
-      ) c
-      on datasets.id = c.dataset_id
-      join tags
-      on tags.id = c.tag_id
-    ")
-  end
-
   def self.joins_datafile_and_freeformats(workbook = nil)
     rel = self.joins("
       left join freeformats on freeformats.freeformattable_id = datasets.id AND freeformats.freeformattable_type='Dataset'

app/models/dataset_tag.rb

@@ -0,0 +1,15 @@
+# This is a view-based model
+# The view is defined in migration file: 20131116163151_create_dataset_tags_view.rb
+# More info about view-based model can be found in Chapter 11 in book 'Enterprise Rails'.
+
+class DatasetTag < ActiveRecord::Base
+  belongs_to :dataset
+  belongs_to :tag, class_name: 'ActsAsTaggableOn::Tag'
+
+  def self.tag_counts
+    ActsAsTaggableOn::Tag.joins('left join dataset_tags on tags.id = dataset_tags.tag_id')
+                         .select('tags.*, count(dataset_id) as count')
+                         .group('tags.id')
+                         .order('lower(tags.name)')
+  end
+end

app/models/user.rb

@@ -141,7 +141,7 @@ def change_avatar_file_name
   end
 
   def add_protocol_to_url
-    unless self.url.blank?
+    if self.url.present? and self.url_changed?
       /^http/.match(self.url) ? self.url : self.url = "http://#{url}"
     end
   end

app/models/workbook.rb

@@ -155,7 +155,7 @@ def datemax
     date_string.to_i > 2000 ? Date.new(date_string.to_i) : Date.today
   end
 
-  # This generate a hash in form of {header: columnnr}. columnnr is 0-based
+  # This generate a hash in form of {header: columnnr}. columnnr is 0-based.
   # When a column is saved later, its info will be stored into values.
   # then, this hash becomes {header: [columnnr, datacolumn_id, data_type_id]}
   def header_info_lookup

app/views/datasets/show.html.haml

@@ -16,7 +16,7 @@
   - content_for :info do
     - unless @tags.blank?
       %h3= link_to "Keywords", keywords_dataset_path(@dataset)
-      = raw  @tags.sort_by(&:name).map{|t| link_to t.name, keyword_path(t.id)}.join(", ")
+      = raw  @tags.map{|t| link_to t.name, keyword_path(t.id)}.join(", ")
 
     %h3
       Last update

db/migrate/20131116163151_create_dataset_tags_view.rb

@@ -0,0 +1,23 @@
+class CreateDatasetTagsView < ActiveRecord::Migration
+  def up
+    execute <<-SQL
+      CREATE OR REPLACE view dataset_tags AS
+        (
+          select taggable_id as dataset_id, tag_id
+          from taggings
+          where taggable_type = 'Dataset'
+        union
+          select distinct d.dataset_id, g.tag_id
+          from taggings g join datacolumns d
+          on g.taggable_id = d.id
+          where g.taggable_type = 'Datacolumn'
+        );
+    SQL
+  end
+
+  def down
+    execute <<-SQL
+      drop view dataset_tags;
+    SQL
+  end
+end

db/non_schema_sql.sql

@@ -171,4 +171,21 @@ CREATE OR REPLACE FUNCTION update_date_category_datasets(category_id integer) RE
          set updated_at = now()
         from sheetcells sc inner join datacolumns dc on sc.datacolumn_id = dc.id
         where category_id = $1 and datasets.id = dc.dataset_id
-    returning true$_$;
+    returning true$_$;
+
+
+
+--
+--- Define a view
+--
+CREATE OR REPLACE view dataset_tags AS
+  (
+    select taggable_id as dataset_id, tag_id
+    from taggings
+    where taggable_type = 'Dataset'
+  union
+    select distinct d.dataset_id, g.tag_id
+    from taggings g join datacolumns d
+    on g.taggable_id = d.id
+    where g.taggable_type = 'Datacolumn'
+  );

db/schema.rb

@@ -11,7 +11,7 @@
 #
 # It's strongly recommended to check this file into your version control system.
 
-ActiveRecord::Schema.define(:version => 20131111135126) do
+ActiveRecord::Schema.define(:version => 20131116163151) do
 
   create_table "author_paperproposals", :force => true do |t|
     t.integer  "paperproposal_id"