Merge 8cbd87a into 39f405c

config/neo4j/add_classnames.yml

@@ -0,0 +1 @@
+# This file is used as the source for adding _classname properties to nodes and relationships.# Within the relationship array, the from/to labels are optional but "type" is not.## nodes:#   add:        [Lesson, Student, Exam]#   overwrite:  [Teacher]# relationships:#   add:#     StudentLesson: { from: 'Student', to: 'Lesson', type: 'enrolled_in' }#   overwrite:#     ExamLesson:    { type: 'has_this_thing' }

lib/neo4j/migration.rb

@@ -55,15 +55,7 @@ def add_ids_to(model)
 
           new_ids = to_set.times.map { new_id_for(model) }
           begin
-            last_time_taken = Benchmark.realtime do
-              Neo4j::Session.query("MATCH (n:`#{label}`) WHERE NOT has(n.#{property})
-                with COLLECT(n) as nodes, {new_ids} as ids
-                FOREACH(i in range(0,#{to_set - 1})| 
-                  FOREACH(node in [nodes[i]]|
-                    SET node.#{property} = ids[i]))
-                RETURN distinct(true)
-                LIMIT #{to_set}", new_ids: new_ids)
-            end
+            last_time_taken = id_batch_set(label, property, new_ids, to_set)
           rescue Neo4j::Server::CypherResponse::ResponseError, Faraday::TimeoutError
             new_max_per_batch = (max_per_batch * 0.8).round
             puts "Error querying #{max_per_batch} nodes.  Trying #{new_max_per_batch}"
@@ -72,6 +64,20 @@ def add_ids_to(model)
         end
       end
 
+      def id_batch_set(label, property, new_ids, to_set)
+        Benchmark.realtime do
+          Neo4j::Transaction.run do
+            Neo4j::Session.query("MATCH (n:`#{label}`) WHERE NOT has(n.#{property})
+              with COLLECT(n) as nodes, #{new_ids} as ids
+              FOREACH(i in range(0,#{to_set - 1})|
+                FOREACH(node in [nodes[i]]|
+                  SET node.#{property} = ids[i]))
+              RETURN distinct(true)
+              LIMIT #{to_set}")
+          end
+        end
+      end
+
       def default_max_per_batch
         900
       end
@@ -84,5 +90,96 @@ def new_id_for(model)
         end
       end
     end
+
+    class AddClassnames < Neo4j::Migration
+      attr_reader :classnames_filename, :classnames_filepath
+
+      def initialize
+        @classnames_filename = 'add_classnames.yml'
+        @classnames_filepath = File.join(Rails.root.join('db', 'neo4j-migrate'), classnames_filename)
+      end
+
+      def migrate
+        puts "Adding classnames. This make take some time."
+        execute(true)
+      end
+
+      def test
+        puts "TESTING! No queries will be executed."
+        execute(false)
+      end
+
+      def setup
+        puts "Creating file #{classnames_filepath}. Please use this as the migration guide."
+        FileUtils.mkdir_p("db/neo4j-migrate")
+        unless File.file?(@classnames_filepath)
+          source = File.join(File.dirname(__FILE__), "..", "..", "config", "neo4j", classnames_filename)
+          FileUtils.copy_file(source, classnames_filepath)
+        end
+      end
+
+      private
+
+      def execute(migrate = false)
+        file_init
+        map = []
+        map.push :nodes         if @model_map[:nodes]
+        map.push :relationships if @model_map[:relationships]
+        map.each do |type|
+          @model_map[type].each do |action, labels|
+            do_classnames(action, labels, type, migrate)
+          end
+        end
+      end
+
+      def do_classnames(action, labels, type, migrate = false)
+        method = type == :nodes ? :node_cypher : :rel_cypher
+        labels.each do |label|
+          puts cypher = self.send(method, label, action)
+          execute_cypher(cypher) if migrate
+        end
+      end
+
+      def file_init
+        @model_map = ActiveSupport::HashWithIndifferentAccess.new(YAML.load_file(classnames_filepath)) 
+      end
+
+      def node_cypher(label, action)
+        where, phrase_start = action_variables(action, 'n')
+        puts "#{phrase_start} _classname '#{label}' on nodes with matching label:"
+        "MATCH (n:`#{label}`) #{where} SET n._classname = '#{label}' RETURN COUNT(n) as modified"
+      end
+
+      def rel_cypher(hash, action)
+        label = hash[0]
+        value = hash[1]
+        from = value[:from]
+        raise "All relationships require a 'type'" unless value[:type]
+
+        from_cypher = from ? "(from:`#{from}`)" : "(from)"
+        to = value[:to]
+        to_cypher = to ? "(to:`#{to}`)" : "(to)"
+        type = "[r:`#{value[:type]}`]"
+        where, phrase_start = action_variables(action, 'r')
+        puts "#{phrase_start} _classname '#{label}' where type is '#{value[:type]}' using cypher:"
+        "MATCH #{from_cypher}-#{type}->#{to_cypher} #{where} SET r._classname = '#{label}' return COUNT(r) as modified"
+      end
+
+      def execute_cypher(query_string)
+        puts "Modified #{Neo4j::Session.query(query_string).first.modified} records"
+        puts ""
+      end
+
+      def action_variables(action, identifier)
+        case action
+        when 'overwrite'
+          ['', 'Overwriting']
+        when 'add'
+          ["WHERE NOT HAS(#{identifier}._classname)", 'Adding']
+        else
+          raise "Invalid action #{action} specified"
+        end
+      end
+    end
   end
 end