rocketjob 5.4.0.beta2 → 6.0.0.rc3

data/lib/rocket_job/jobs/on_demand_job.rb

@@ -38,12 +38,11 @@
 #
 # Example: Retain output:
 #   code = <<~CODE
-#     {'value' => data['a'] * data['b']}
+#     data['result'] = data['a'] * data['b']
 #   CODE
 #
 #   RocketJob::Jobs::OnDemandJob.create!(
 #     code:           code,
-#     collect_output: true,
 #     data:           {'a' => 10, 'b' => 2}
 #   )
 #

data/lib/rocket_job/jobs/performance_job.rb

@@ -6,9 +6,11 @@ module RocketJob
       # Define the job's default attributes
       self.description         = "Performance Test"
       self.priority            = 5
-      self.slice_size          = 100
       self.destroy_on_complete = false
 
+      input_category slice_size: 100
+      output_category
+
       # No operation, just return the supplied line (record)
       def perform(line)
         line

data/lib/rocket_job/jobs/re_encrypt/relational_job.rb

@@ -1,6 +1,3 @@
-require "active_record"
-require "sync_attr"
-
 # Batch Worker to Re-encrypt all encrypted fields in MySQL that start with `encrytped_`.
 #
 # Run in Rails console:
@@ -11,116 +8,126 @@ require "sync_attr"
 # * This job will find any column in the database that starts with`encrypted_`.
 # * This means that temporary or other tables not part of the application tables will also be processed.
 # * Since it automatically finds and re-encrypts any column, new columns are handled without any manual intervention.
-module RocketJob
-  module Jobs
-    module ReEncrypt
-      class RelationalJob < RocketJob::Job
-        include RocketJob::Batch
-
-        self.slice_size               = 1000
-        self.priority                 = 30
-        self.destroy_on_complete      = false
-        self.compress                 = true
-        self.throttle_running_jobs    = 1
-        self.throttle_running_workers = 10
-
-        # Name of the table being re-encrypted
-        field :table_name, type: String
-
-        # Limit the number of records to re-encrypt in test environments
-        field :limit, type: Integer
-
-        validates_presence_of :table_name
-        before_batch :upload_records
-
-        # Returns [Hash] of table names with each entry being an array
-        # of columns that start with encrypted_
-        sync_cattr_reader :encrypted_columns do
-          h = {}
-          connection.tables.each do |table|
-            columns = connection.columns(table)
-            columns.each do |column|
-              if column.name.start_with?("encrypted_")
-                add_column = column.name
-                (h[table] ||= []) << add_column if add_column
+if defined?(ActiveRecord) && defined?(SyncAttr)
+  require "active_record"
+  require "sync_attr"
+
+  module RocketJob
+    module Jobs
+      module ReEncrypt
+        class RelationalJob < RocketJob::Job
+          include RocketJob::Batch
+
+          self.priority                 = 30
+          self.destroy_on_complete      = false
+          self.throttle_running_jobs    = 1
+          self.throttle_running_workers = 10
+
+          input_category slice_size: 1_000
+
+          # Name of the table being re-encrypted
+          field :table_name, type: String
+
+          # Limit the number of records to re-encrypt in test environments
+          field :limit, type: Integer
+
+          validates_presence_of :table_name
+          before_batch :upload_records
+
+          # Returns [Hash] of table names with each entry being an array
+          # of columns that start with encrypted_
+          sync_cattr_reader :encrypted_columns do
+            h = {}
+            connection.tables.each do |table|
+              columns = connection.columns(table)
+              columns.each do |column|
+                if column.name.start_with?("encrypted_")
+                  add_column = column.name
+                  (h[table] ||= []) << add_column if add_column
+                end
               end
             end
+            h
           end
-          h
-        end
 
-        # Re-encrypt all `encrypted_` columns in the relational database.
-        # Queues a Job for each table that needs re-encryption.
-        def self.start(**args)
-          encrypted_columns.keys.collect do |table|
-            create!(table_name: table, description: table, **args)
+          # Re-encrypt all `encrypted_` columns in the relational database.
+          # Queues a Job for each table that needs re-encryption.
+          def self.start(**args)
+            encrypted_columns.keys.collect do |table|
+              create!(table_name: table, description: table, **args)
+            end
           end
-        end
 
-        # Re-encrypt all encrypted columns for the named table.
-        # Does not use AR models since we do not have models for all tables.
-        def perform(range)
-          start_id, end_id = range
+          # Re-encrypt all encrypted columns for the named table.
+          # Does not use AR models since we do not have models for all tables.
+          def perform(range)
+            start_id, end_id = range
 
-          columns = self.class.encrypted_columns[table_name]
-          unless columns&.size&.positive?
-            logger.error "No columns for table: #{table_name} from #{start_id} to #{end_id}"
-            return
-          end
+            columns = self.class.encrypted_columns[table_name]
+            unless columns&.size&.positive?
+              logger.error "No columns for table: #{table_name} from #{start_id} to #{end_id}"
+              return
+            end
 
-          logger.info "Processing: #{table_name} from #{start_id} to #{end_id}"
-          sql = "select id, #{columns.join(',')} from #{quoted_table_name} where id >= #{start_id} and id <= #{end_id}"
-
-          # Use AR to fetch all the records
-          self.class.connection.select_rows(sql).each do |row|
-            row     = row.unshift(nil)
-            index   = 1
-            sql     = "update #{quoted_table_name} set "
-            updates = []
-            columns.collect do |column|
-              index += 1
-              value = row[index]
-              # Prevent re-encryption
-              unless value.blank?
-                new_value = re_encrypt(value)
-                updates << "#{column} = \"#{new_value}\"" if new_value != value
+            logger.info "Processing: #{table_name} from #{start_id} to #{end_id}"
+            sql = "select id, #{columns.join(',')} from #{quoted_table_name} where id >= #{start_id} and id <= #{end_id}"
+
+            # Use AR to fetch all the records
+            self.class.connection.select_rows(sql).each do |row|
+              row.unshift(nil)
+              index   = 1
+              sql     = "update #{quoted_table_name} set "
+              updates = []
+              columns.collect do |column|
+                index += 1
+                value = row[index]
+                # Prevent re-encryption
+                unless value.blank?
+                  new_value = re_encrypt(value)
+                  updates << "#{column} = \"#{new_value}\"" if new_value != value
+                end
+              end
+              if updates.size.positive?
+                sql << updates.join(", ")
+                sql << " where id=#{row[1]}"
+                logger.trace sql
+                self.class.connection.execute sql
+              else
+                logger.trace { "Skipping empty values #{table_name}:#{row[1]}" }
               end
-            end
-            if updates.size.positive?
-              sql << updates.join(", ")
-              sql << " where id=#{row[1]}"
-              logger.trace sql
-              self.class.connection.execute sql
-            else
-              logger.trace { "Skipping empty values #{table_name}:#{row[1]}" }
             end
           end
-        end
 
-        # Returns a database connection.
-        #
-        # Override this method to support other ways of obtaining a thread specific database connection.
-        def self.connection
-          ActiveRecord::Base.connection
-        end
+          # Returns a database connection.
+          #
+          # Override this method to support other ways of obtaining a thread specific database connection.
+          def self.connection
+            ActiveRecord::Base.connection
+          end
 
-        private
+          private
 
-        def quoted_table_name
-          @quoted_table_name ||= self.class.connection.quote_table_name(table_name)
-        end
+          def quoted_table_name
+            @quoted_table_name ||= self.class.connection.quote_table_name(table_name)
+          end
 
-        def re_encrypt(encrypted_value)
-          return encrypted_value if (encrypted_value == "") || encrypted_value.nil?
+          def re_encrypt(encrypted_value)
+            return encrypted_value if (encrypted_value == "") || encrypted_value.nil?
 
-          SymmetricEncryption.encrypt(SymmetricEncryption.decrypt(encrypted_value))
-        end
+            SymmetricEncryption.encrypt(SymmetricEncryption.decrypt(encrypted_value))
+          end
 
-        # Upload range to re-encrypt all rows in the specified table.
-        def upload_records
-          start_id          = self.class.connection.select_value("select min(id) from #{quoted_table_name}").to_i
-          last_id           = self.class.connection.select_value("select max(id) from #{quoted_table_name}").to_i
-          self.record_count = last_id.positive? ? (input.upload_integer_range_in_reverse_order(start_id, last_id) * slice_size) : 0
+          # Upload range to re-encrypt all rows in the specified table.
+          def upload_records
+            start_id          = self.class.connection.select_value("select min(id) from #{quoted_table_name}").to_i
+            last_id           = self.class.connection.select_value("select max(id) from #{quoted_table_name}").to_i
+            self.record_count =
+              if last_id.positive?
+                input.upload_integer_range_in_reverse_order(start_id, last_id) * input_category.slice_size
+              else
+                0
+              end
+          end
         end
       end
     end

data/lib/rocket_job/jobs/upload_file_job.rb

@@ -19,7 +19,7 @@ module RocketJob
       field :properties, type: Hash, default: {}, user_editable: true
 
       # File to upload
-      field :upload_file_name, type: String, user_editable: true
+      field :upload_file_name, type: IOStreams::Path, user_editable: true
 
       # The original Input file name.
       # Used by #upload to extract the IOStreams when present.
@@ -33,10 +33,11 @@ module RocketJob
       validate :job_is_a_rocket_job
       validate :job_implements_upload
       validate :file_exists
+      validate :job_has_properties
 
       # Create the job and upload the file into it.
       def perform
-        job    = job_class.new(properties)
+        job    = job_class.from_properties(properties)
         job.id = job_id if job_id
         upload_file(job)
         job.save!
@@ -66,7 +67,10 @@ module RocketJob
         elsif job.respond_to?(:full_file_name=)
           job.full_file_name = upload_file_name
         else
-          raise(ArgumentError, "Model #{job_class_name} must implement '#upload', or have attribute 'upload_file_name' or 'full_file_name'")
+          raise(
+            ArgumentError,
+            "Model #{job_class_name} must implement '#upload', or have attribute 'upload_file_name' or 'full_file_name'"
+          )
         end
       end
 
@@ -85,17 +89,49 @@ module RocketJob
         klass = job_class
         return if klass.nil? || klass.instance_methods.any? { |m| VALID_INSTANCE_METHODS.include?(m) }
 
-        errors.add(:job_class_name, "#{job_class} must implement any one of: :#{VALID_INSTANCE_METHODS.join(' :')} instance methods")
+        errors.add(:job_class_name,
+                   "#{job_class} must implement any one of: :#{VALID_INSTANCE_METHODS.join(' :')} instance methods")
       end
 
       def file_exists
-        return if upload_file_name.nil?
+        # Only check for file existence when it is a local file
+        return unless upload_file_name.is_a?(IOStreams::Paths::File)
+        return errors.add(:upload_file_name, "Upload file name can't be blank.") if upload_file_name.to_s == ""
 
-        uri = URI.parse(upload_file_name)
-        return unless uri.scheme.nil? || uri.scheme == "file"
-        return if File.exist?(upload_file_name)
+        return if upload_file_name.exist?
 
         errors.add(:upload_file_name, "Upload file: #{upload_file_name} does not exist.")
+      rescue NotImplementedError
+        nil
+      end
+
+      def job_has_properties
+        klass = job_class
+        return unless klass
+
+        properties.each_pair do |k, _v|
+          next if klass.public_method_defined?("#{k}=".to_sym)
+
+          if %i[output_categories input_categories].include?(k)
+            category_class = k == :input_categories ? RocketJob::Category::Input : RocketJob::Category::Output
+            properties[k].each do |category|
+              category.each_pair do |key, _value|
+                next if category_class.public_method_defined?("#{key}=".to_sym)
+
+                errors.add(
+                  :properties,
+                  "Unknown Property in #{k}: Attempted to set a value for #{key}.#{k} which is not allowed on the job #{job_class_name}"
+                )
+              end
+            end
+            next
+          end
+
+          errors.add(
+            :properties,
+            "Unknown Property: Attempted to set a value for #{k.inspect} which is not allowed on the job #{job_class_name}"
+          )
+        end
       end
     end
   end

data/lib/rocket_job/lookup_collection.rb

@@ -0,0 +1,69 @@
+module RocketJob
+  class LookupCollection < Mongo::Collection
+    # Rapidly upload individual records in batches.
+    #
+    # Operates directly on a Mongo Collection to avoid the overhead of creating Mongoid objects
+    # for each and every row.
+    #
+    # Example:
+    #   lookup_collection(:my_lookup).upload do |io|
+    #     io << {id: 123, data: "first record"}
+    #     io << {id: 124, data: "second record"}
+    #   end
+    #
+    #   input_category(:my_lookup).find(id: 123).first
+    def upload(batch_size: 10_000, &block)
+      BatchUploader.upload(batch_size: batch_size, &block)
+    end
+
+    # Looks up the value at the specified id.
+    # Returns [nil] if no record was found with the supplied id.
+    def lookup(id)
+      find(id: id).first
+    end
+
+    # Internal class for uploading records in batches
+    class BatchUploader
+      attr_reader :record_count
+
+      def self.upload(collection, **args)
+        writer = new(collection, **args)
+        yield(writer)
+        writer.record_count
+      ensure
+        writer&.close
+      end
+
+      def initialize(collection, batch_size:)
+        @batch_size   = batch_size
+        @record_count = 0
+        @batch_count  = 0
+        @documents    = []
+        @collection   = collection
+      end
+
+      def <<(record)
+        raise(ArgumentError, "Record must be a Hash") unless record.is_a?(Hash)
+
+        unless record.key?(:id) || record.key?("id") || record.key?("_id")
+          raise(ArgumentError, "Record must include an :id key")
+        end
+
+        @documents << record
+        @record_count += 1
+        @batch_count  += 1
+        if @batch_count >= @batch_size
+          @collection.insert_many(@documents)
+          @documents.clear
+          @batch_count = 0
+        end
+
+        self
+      end
+
+      def close
+        @collection.insert_many(@documents) unless @documents.empty?
+      end
+    end
+  end
+end

data/lib/rocket_job/plugins/job/model.rb

@@ -37,12 +37,10 @@ module RocketJob
           # arrives, then the current job will complete the current slices and process
           # the new higher priority job
           field :priority, type: Integer, default: 50, class_attribute: true, user_editable: true, copy_on_restart: true
+          validates_inclusion_of :priority, in: 1..100
 
           # When the job completes destroy it from both the database and the UI
-          field :destroy_on_complete, type: Boolean, default: true, class_attribute: true, copy_on_restart: true
-
-          # Whether to store the results from this job
-          field :collect_output, type: Boolean, default: false, class_attribute: true
+          field :destroy_on_complete, type: Mongoid::Boolean, default: true, class_attribute: true, copy_on_restart: true
 
           # Run this job no earlier than this time
           field :run_at, type: Time, user_editable: true
@@ -54,14 +52,15 @@ module RocketJob
           # Can be used to reduce log noise, especially during high volume calls
           # For debugging a single job can be logged at a low level such as :trace
           #   Levels supported: :trace, :debug, :info, :warn, :error, :fatal
-          field :log_level, type: Symbol, class_attribute: true, user_editable: true, copy_on_restart: true
+          field :log_level, type: Mongoid::StringifiedSymbol, class_attribute: true, user_editable: true, copy_on_restart: true
+          validates_inclusion_of :log_level, in: SemanticLogger::LEVELS + [nil]
 
           #
           # Read-only attributes
           #
 
           # Current state, as set by the state machine. Do not modify this value directly.
-          field :state, type: Symbol, default: :queued
+          field :state, type: Mongoid::StringifiedSymbol, default: :queued
 
           # When the job was created
           field :created_at, type: Time, default: -> { Time.now }
@@ -89,17 +88,12 @@ module RocketJob
           # Store the last exception for this job
           embeds_one :exception, class_name: "RocketJob::JobException"
 
-          # Store the Hash result from this job if collect_output is true,
-          # and the job returned actually returned a Hash, otherwise nil
-          # Not applicable to SlicedJob jobs, since its output is stored in a
-          # separate collection
-          field :result, type: Hash
-
+          # Used when workers fetch jobs to work on.
           index({state: 1, priority: 1, _id: 1}, background: true)
+          # Used by Mission Control to display completed jobs sorted by completion.
+          index({completed_at: 1}, background: true)
 
           validates_presence_of :state, :failure_count, :created_at
-          validates :priority, inclusion: 1..100
-          validates :log_level, inclusion: SemanticLogger::LEVELS + [nil]
         end
 
         module ClassMethods
@@ -155,14 +149,8 @@ module RocketJob
 
           # Scope for queued jobs that can run now
           # I.e. Queued jobs excluding scheduled jobs
-          if Mongoid::VERSION.to_f >= 7.1
-            def queued_now
-              queued.and(RocketJob::Job.where(run_at: nil).or(:run_at.lte => Time.now))
-            end
-          else
-            def queued_now
-              queued.or({run_at: nil}, :run_at.lte => Time.now)
-            end
+          def queued_now
+            queued.and(RocketJob::Job.where(run_at: nil).or(:run_at.lte => Time.now))
           end
 
           # Defines all the fields that are accessible on the Document
@@ -183,43 +171,30 @@ module RocketJob
           #
           # @return [ Field ] The generated field
           def field(name, options)
-            if options.delete(:user_editable) == true
-              self.user_editable_fields += [name.to_sym] unless user_editable_fields.include?(name.to_sym)
+            if (options.delete(:user_editable) == true) && !user_editable_fields.include?(name.to_sym)
+              self.user_editable_fields += [name.to_sym]
             end
+
             if options.delete(:class_attribute) == true
               class_attribute(name, instance_accessor: false)
               public_send("#{name}=", options[:default]) if options.key?(:default)
               options[:default] = -> { self.class.public_send(name) }
             end
-            if options.delete(:copy_on_restart) == true
-              self.rocket_job_restart_attributes += [name.to_sym] unless rocket_job_restart_attributes.include?(name.to_sym)
+
+            if (options.delete(:copy_on_restart) == true) && !rocket_job_restart_attributes.include?(name.to_sym)
+              self.rocket_job_restart_attributes += [name.to_sym]
             end
-            super(name, options)
-          end
 
-          # DEPRECATED
-          def rocket_job
-            warn "Replace calls to .rocket_job with calls to set class instance variables. For example: self.priority = 50"
-            yield(self)
+            super(name, options)
           end
 
-          # DEPRECATED
-          def public_rocket_job_properties(*args)
-            warn "Replace calls to .public_rocket_job_properties by adding `user_editable: true` option to the field declaration in #{name} for: #{args.inspect}"
-            self.user_editable_fields += args.collect(&:to_sym)
+          # Builds this job instance from the supplied properties hash.
+          # Overridden by batch to support child objects.
+          def from_properties(properties)
+            new(properties)
           end
         end
 
-        # Returns [true|false] whether to collect nil results from running this batch
-        def collect_nil_output?
-          collect_output? ? (collect_nil_output == true) : false
-        end
-
-        # Returns [true|false] whether to collect the results from running this batch
-        def collect_output?
-          collect_output == true
-        end
-
         # Returns [Float] the number of seconds the job has taken
         # - Elapsed seconds to process the job from when a worker first started working on it
         #   until now if still running, or until it was completed
@@ -282,7 +257,6 @@ module RocketJob
         # Returns [Hash] status of this job
         def as_json
           attrs = serializable_hash(methods: %i[seconds duration])
-          attrs.delete("result") unless collect_output?
           attrs.delete("failure_count") unless failure_count.positive?
           if queued?
             attrs.delete("started_at")
@@ -319,16 +293,17 @@ module RocketJob
           h = as_json
           h.delete("seconds")
           h.dup.each_pair do |k, v|
-            if v.is_a?(Time)
+            case v
+            when Time
               h[k] = v.in_time_zone(time_zone).to_s
-            elsif v.is_a?(BSON::ObjectId)
+            when BSON::ObjectId
               h[k] = v.to_s
             end
           end
           h
         end
 
-        # Returns [Boolean] whether the worker runs on a particular server.
+        # Returns [true|false] whether the worker runs on a particular server.
         def worker_on_server?(server_name)
           return false unless worker_name.present? && server_name.present?