concurrency outline

gossamer_forums.rb

@@ -14,6 +14,8 @@ require 'fileutils'
 require 'csv'
 require 'time'
 
+require 'concurrent'
+
 require File.expand_path("../../../../config/environment", __FILE__)
 # require_relative '../base'
 require File.expand_path("../../../../script/import_scripts/base", __FILE__)
@@ -925,6 +927,50 @@ class GossamerForumsImporter < ImportScripts::Base
     result ? result['post_views'] : 0
   end
 
+  #THREADING OUTLINE HERE --------------------------------------------
+
+  def threaded_topic_import
+    # Get list of IDS that have no parent ID - SELECT post_id FROM gforum_Post WHERE post_root_id = 0;
+    parent_post_ids = execute_query("SELECT post_id FROM gforum_Post WHERE post_root_id = 0")
+
+    parent_post_count = parent_post_ids.count
+    batch_size = 100 #set our batch size
+    current_post_batch = 0 #set our current batch number
+    is_complete = false
+
+    until is_complete
+      # Query in batches, create pool, wait for termination, do it again
+    # SELECT post_id FROM gforum_Post WHERE post_root_id = 0 ORDER BY post_id  
+      current_post_batch_max = current_post_batch + batch_size
+      pool = Concurrent::FixedThreadPool.new(Concurrent.processor_count) #create thread pool that is bounded by processors avaialable
+
+      while current_post_batch < current_post_batch_max
+        post_id = parent_post_ids[current_post_batch]
+        pool.post do
+          topic_import_job(post_id)
+        end
+        current_post_batch += 1
+      end
+      
+      pool.shutdown
+      pool.wait_for_termination
+
+      if current_post_batch >= parent_post_count
+        is_complete = true
+      end
+    end
+  end
+
+  def topic_import_job(post_id)
+    #Here is where you can import the entire topic
+    #Get post -- SELECT post_id, user_id_fk, forum_id_fk, post_root_id, post_subject, post_time, post_message, post_father_id, post_replies FROM gforum_Post WHERE post_id = post_id
+    #check if exists, create if not
+    #get children, create -- SELECT post_id, user_id_fk, forum_id_fk, post_root_id, post_subject, post_time, post_message, post_father_id, post_replies FROM gforum_Post WHERE post_root_id = post_id
+    #this parts needs to be synchronously to avoid race conditions
+  end
+
+  #-------------------------------------------------------------------
+
   # Import topics and posts from Gossamer Forums to Discourse
   def import_topics_and_posts_with_attachments
     puts "Importing topics and posts with attachments..."
@@ -946,7 +992,7 @@ class GossamerForumsImporter < ImportScripts::Base
 #   Attachment example:  highest_processed_post_id = 1359862 
 
     # Execute the query to get all posts ordered by post_id
-    execute_query("SELECT post_id, user_id_fk, forum_id_fk, post_root_id, post_subject, post_time, post_message, post_likes, post_father_id, post_replies FROM gforum_Post ORDER BY post_id").each do |row|
+    execute_query("SELECT post_id, user_id_fk, forum_id_fk, post_root_id, post_subject, post_time, post_message, post_father_id, post_replies FROM gforum_Post ORDER BY post_id").each do |row|
       post_id = row['post_id'].to_i
 
      # Skip posts that have already been processed