Resolve #40 - Spark job running too many times.

- Use Collection, not WasapiFile in spark job
- update collection seeding job
- update db to have account in collection table
- this should probably be a join. whatever.

app/jobs/collections_spark_job.rb

@@ -11,7 +11,7 @@ def after_perform
 
   def perform(user_id, collection_id)
     spark_shell = ENV['SPARK_SHELL']
-    WasapiFile.where('user_id = ? AND collection_id = ?', user_id, collection_id).each do |c|
+    Collection.where('user_id = ? AND collection_id = ?', user_id, collection_id).each do |c|
       collection_path = ENV['DOWNLOAD_PATH'] +
                         '/' + c.account.to_s +
                         '/' + c.collection_id.to_s + '/'

app/jobs/wasapi_files_populate_job.rb

@@ -64,7 +64,7 @@ def perform(user)
         break if paginate.blank?
       end
     end
-    WasapiFile.distinct.select('user_id, collection_id').each do |cid|
+    WasapiFile.distinct.select('user_id, collection_id', 'account').each do |cid|
       collection_api_request_url = AI_COLLECTION_API_URL + cid.collection_id.to_s
       collection_api_request_code = HTTP.get(collection_api_request_url).code
       if collection_api_request_code == 200
@@ -73,6 +73,7 @@ def perform(user)
         Collection.find_or_create_by!(
           collection_id: cid.collection_id,
           user_id: cid.user_id,
+          account: cid.account,
           title: collection_api_results['name'],
           public: collection_api_results['publicly_visible']
         )

db/migrate/20180110143134_create_collections.rb

@@ -3,6 +3,7 @@ def change
     create_table :collections, id: false, force: true do |t|
       t.integer :collection_id, null: false
       t.integer :user_id
+      t.integer :account
       t.string :title
       t.boolean :public
 

db/schema.rb

@@ -15,6 +15,7 @@
   create_table "collections", id: false, force: :cascade do |t|
     t.integer "collection_id", null: false
     t.integer "user_id"
+    t.integer "account"
     t.string "title"
     t.boolean "public"
     t.datetime "created_at", null: false