Artist check within local, mb and lastfm databases

Gemfile.lock

@@ -1,6 +1,6 @@
 GIT
   remote: http://github.com/magnolia-fan/lastfm-client.git
-  revision: fe5614565eebb5b0b019c103211bba3698848daa
+  revision: 26fec825943d0bb77606bf058386e786ee97572b
   specs:
     lastfm-client (0.0.2)
       json (>= 1.4.6)
@@ -59,7 +59,7 @@ GEM
     haml (3.1.3)
     hike (1.2.1)
     i18n (0.6.0)
-    json (1.6.0)
+    json (1.6.1)
     libv8 (3.3.10.2)
     mail (2.3.0)
       i18n (>= 0.4.0)
@@ -67,7 +67,7 @@ GEM
       treetop (~> 1.4.8)
     mime-types (1.16)
     multi_json (1.0.3)
-    musicbrainz (0.4.5)
+    musicbrainz (0.4.7)
       nokogiri
     nokogiri (1.5.0)
     pg (0.11.0)

app/controllers/import_controller.rb

@@ -23,6 +23,13 @@ class ImportController < ApplicationController
       ap e.backtrace
     end
     
+    if brainz_artist.nil?
+      artist.mbid = nil
+      artist.status = 2
+      artist.save
+      return 2
+    end
+
     begin
       # Save artist
       artist.desc = lastfm_artist['artist']['bio']['summary']

db/migrate/20110919003651_add_approved_flag_to_artist.rb

@@ -0,0 +1,5 @@
+class AddApprovedFlagToArtist < ActiveRecord::Migration
+  def change
+    add_column :artists, :approved, :integer
+  end
+end

db/schema.rb

@@ -11,7 +11,7 @@
 #
 # It's strongly recommended to check this file into your version control system.
 
-ActiveRecord::Schema.define(:version => 20110915220228) do
+ActiveRecord::Schema.define(:version => 20110919003651) do
 
   create_table "albums", :force => true do |t|
     t.string   "name"
@@ -36,6 +36,7 @@ ActiveRecord::Schema.define(:version => 20110915220228) do
     t.string   "mbid"
     t.integer  "status"
     t.integer  "listeners"
+    t.integer  "approved"
   end
 
   create_table "delayed_jobs", :force => true do |t|

lib/tasks/artist_correct.rake

@@ -0,0 +1,47 @@
+require 'cgi'
+
+namespace :artist do
+  desc 'Find and fix uncorrect mbid'
+  task :correct => :environment do
+    Artist.where( :approved => nil ).each do |artist|
+      puts "Artist: "+artist.name
+      bh_albums = artist.albums.map{ |a| a.name.scan(/[a-zA-Z0-9]/).join() }
+      tmp = LastFM::Artist.get_top_albums( :artist => artist.name )['topalbums']
+      if tmp['album'].nil?
+        tmp['album'] = []
+      elsif tmp['album'].is_a? Hash
+        tmp['album'] = [tmp['album']]
+      end
+      lastfm_albums = tmp['album'].map do |a|
+        a['name'].scan(/[a-zA-Z0-9]/).join()
+      end
+      if (bh_albums & lastfm_albums).empty?
+        puts "Result: No intersection"
+        best_candidate = nil
+        best_intersection = 0
+        MusicBrainz::Artist.search(artist.name.gsub(/\!/, '')).each do |candidate|
+          candidate = MusicBrainz::Artist.find(candidate[:mbid])
+          candidate_albums = candidate.release_groups.map{ |a| a.title.scan(/[a-zA-Z0-9]/).join() }
+          if (candidate_albums & lastfm_albums).length > best_intersection
+            best_intersection = (candidate_albums & lastfm_albums).length
+            best_candidate = candidate.id
+          end
+        end
+        if best_candidate.nil?
+          puts "No better candidate found (now: "+artist.mbid.to_s+")\n\n"
+          artist.approved = 0
+          artist.save
+        else
+          puts "Better candidate found: "+best_candidate+" (now: "+artist.mbid.to_s+")\n\n"
+          artist.approved = 2
+          artist.save
+        end
+      else
+        artist.approved = 1
+        artist.save
+        puts "Result: OK\n\n"
+      end
+      sleep 1
+    end
+  end
+end

lib/tasks/db_artist.rake

@@ -6,14 +6,15 @@ namespace :db do
         artist.albums.each do |album|
           album.destroy
         end
-        puts artist.name+(artist.mbid.nil? or artist.mbid.empty? ? '' : ' ('+artist.mbid+')')
+        puts 'Artist:'+artist.name
+        puts 'MBID:  '+artist.mbid unless artist.mbid.nil?
         res = ImportController.importArtist(artist.name)
         if res == 1
-          puts 'OK'
+          puts 'Status: OK'
         elsif res == 2
-          puts 'FAIL'
+          puts 'Status: FAILED'
         elsif res == 3
-          puts 'SKIP'
+          puts 'Status: SKIP'
         end
         puts ''
       end