rocketjob 5.3.3 → 5.4.0.beta1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: d487b34621338794b4e3da43753d858bf174b45f4f0dff6c9f478e212a71db8e
-  data.tar.gz: b6c3ddf2688a54dee0ad6ae6d1c4e4fc38448674dcbb7e0052d0295e57258f2d
+  metadata.gz: 55ab6ca2b3f76cdb4ddf679a2d3e88e1d6a6f3106a69349129a267b629ef4a53
+  data.tar.gz: 4e0d07878fb4265179b4a270650cc9b89ca4bca55f1d8f9a3451cb3064062c35
 SHA512:
-  metadata.gz: 83e61b6520e71e370b2519ee8f937bdf69060050e67ca4ec3146e6a9306a6c26fd0307cb54c94189c2149eb3f8e14fbe7aa8f3fc6f9c17d6b06804745ecaf570
-  data.tar.gz: 20d83702151f6a1333152aa7ee903e7f7c4a79593291496f80670b7952c170734edec16280f0f449db01483e77e7e3af48c8100905688a5bd8998da573f74160
+  metadata.gz: fdc7ae3827d987404f431410cb81161fbfa269dfe7575a4a86a6abf362c59c6acd0d2c7e9856273ced396eeaf20e50561f1be10a7de0d5c4ae45e5648d15d083
+  data.tar.gz: 7bb9e9ac90569e78e135293efcd357a0d75037b5417f5f530f18ae3f17f44acf28d7de5b4aeef880a0315117f2992623958a4fb4c9df93ca7273369ac052e759

data/lib/rocket_job/batch/io.rb

@@ -18,7 +18,7 @@ module RocketJob
           raise "Category #{category.inspect}, must be registered in input_categories: #{input_categories.inspect}"
         end
 
-        (@inputs ||= {})[category] ||= RocketJob::Sliced::Input.new(**rocket_job_io_slice_arguments("inputs", category))
+        (@inputs ||= {})[category] ||= RocketJob::Sliced.factory(:input, category, self)
       end
 
       # Returns [RocketJob::Sliced::Output] output collection for holding output slices
@@ -34,7 +34,7 @@ module RocketJob
           raise "Category #{category.inspect}, must be registered in output_categories: #{output_categories.inspect}"
         end
 
-        (@outputs ||= {})[category] ||= RocketJob::Sliced::Output.new(**rocket_job_io_slice_arguments("outputs", category))
+        (@outputs ||= {})[category] ||= RocketJob::Sliced.factory(:output, category, self)
       end
 
       # Upload the supplied file, io, IOStreams::Path, or IOStreams::Stream.
@@ -355,8 +355,18 @@ module RocketJob
 
         return output(category).download(header_line: header_line, &block) if block
 
-        IOStreams.new(stream).writer(:line, **args) do |io|
-          output(category).download(header_line: header_line) { |record| io << record }
+        output_collection = output(category)
+
+        if output_collection.binary?
+          IOStreams.new(stream).stream(:none).writer(**args) do |io|
+            raise(ArgumenError, "A `header_line` is not supported with binary output collections") if header_line
+
+            output_collection.download { |record| io << record[:binary] }
+          end
+        else
+          IOStreams.new(stream).writer(:line, **args) do |io|
+            output_collection.download(header_line: header_line) { |record| io << record }
+          end
         end
       end
 
@@ -393,21 +403,6 @@ module RocketJob
           RocketJob::Sliced::Writer::Output.collect(self, input_slice) { |writer| writer << result }
         end
       end
-
-      private
-
-      def rocket_job_io_slice_arguments(collection_type, category)
-        collection_name = "rocket_job.#{collection_type}.#{id}"
-        collection_name << ".#{category}" unless category == :main
-
-        args = {collection_name: collection_name, slice_size: slice_size}
-        if encrypt
-          args[:slice_class] = Sliced::EncryptedSlice
-        elsif compress
-          args[:slice_class] = Sliced::CompressedSlice
-        end
-        args
-      end
     end
   end
 end

data/lib/rocket_job/batch/model.rb

@@ -44,12 +44,12 @@ module RocketJob
         # Compress uploaded records.
         # The fields are not affected in any way, only the data stored in the
         # records and results collections will compressed
-        field :compress, type: Boolean, default: false, class_attribute: true
+        field :compress, type: Object, default: false, class_attribute: true
 
         # Encrypt uploaded records.
         # The fields are not affected in any way, only the data stored in the
         # records and results collections will be encrypted
-        field :encrypt, type: Boolean, default: false, class_attribute: true
+        field :encrypt, type: Object, default: false, class_attribute: true
 
         #
         # Values that jobs can also update during processing

data/lib/rocket_job/jobs/upload_file_job.rb

@@ -40,7 +40,7 @@ module RocketJob
         job.id = job_id if job_id
         upload_file(job)
         job.save!
-      rescue StandardError => e
+      rescue Exception => e
         # Prevent partial uploads
         job&.cleanup! if job.respond_to?(:cleanup!)
         raise(e)

data/lib/rocket_job/sliced.rb

@@ -0,0 +1,91 @@
+module RocketJob
+  module Sliced
+    autoload :BZip2OutputSlice, "rocket_job/sliced/bzip2_output_slice"
+    autoload :CompressedSlice, "rocket_job/sliced/compressed_slice"
+    autoload :EncryptedSlice, "rocket_job/sliced/encrypted_slice"
+    autoload :Input, "rocket_job/sliced/input"
+    autoload :Output, "rocket_job/sliced/output"
+    autoload :Slice, "rocket_job/sliced/slice"
+    autoload :Slices, "rocket_job/sliced/slices"
+    autoload :Store, "rocket_job/sliced/store"
+
+    module Writer
+      autoload :Input, "rocket_job/sliced/writer/input"
+      autoload :Output, "rocket_job/sliced/writer/output"
+    end
+
+    # Returns [RocketJob::Sliced::Slices] for the relevant type and category.
+    #
+    # Supports compress and encrypt with [true|false|Hash] values.
+    # When [Hash] they must specify whether the apply to the input or output collection types.
+    #
+    # Example, compress both input and output collections:
+    #   class MyJob < RocketJob::Job
+    #     include RocketJob::Batch
+    #     self.compress = true
+    #   end
+    #
+    # Example, compress just the output collections:
+    #   class MyJob < RocketJob::Job
+    #     include RocketJob::Batch
+    #     self.compress = {output: true}
+    #   end
+    #
+    # To use the specialized BZip output compressor, and the regular compressor for the input collections:
+    #   class MyJob < RocketJob::Job
+    #     include RocketJob::Batch
+    #     self.compress = {output: :bzip2, input: true}
+    #   end
+    def self.factory(type, category, job)
+      raise(ArgumentError, "Unknown type: #{type.inspect}") unless %i[input output].include?(type)
+
+      collection_name = "rocket_job.#{type}s.#{job.id}"
+      collection_name << ".#{category}" unless category == :main
+
+      args               = {collection_name: collection_name, slice_size: job.slice_size}
+      klass              = slice_class(type, job)
+      args[:slice_class] = klass if klass
+
+      if type == :input
+        RocketJob::Sliced::Input.new(args)
+      else
+        RocketJob::Sliced::Output.new(args)
+      end
+    end
+
+    private
+
+    # Parses the encrypt and compress options to determine which slice serializer to use.
+    # `encrypt` takes priority over any `compress` option.
+    def self.slice_class(type, job)
+      encrypt  = extract_value(type, job.encrypt)
+      compress = extract_value(type, job.compress)
+
+      if encrypt
+        case encrypt
+        when true
+          EncryptedSlice
+        else
+          raise(ArgumentError, "Unknown job `encrypt` value: #{compress}") unless compress.is_a?(Slices)
+          # Returns the supplied class to use for encryption.
+          encrypt
+        end
+      elsif compress
+        case compress
+        when true
+          CompressedSlice
+        when :bzip2
+          BZip2OutputSlice
+        else
+          raise(ArgumentError, "Unknown job `compress` value: #{compress}") unless compress.is_a?(Slices)
+          # Returns the supplied class to use for compression.
+          compress
+        end
+      end
+    end
+
+    def self.extract_value(type, value)
+      value.is_a?(Hash) ? value[type] : value
+    end
+  end
+end

data/lib/rocket_job/sliced/bzip2_output_slice.rb

@@ -0,0 +1,43 @@
+module RocketJob
+  module Sliced
+    # This is a specialized output serializer that renders each output slice as a single BZip2 compressed stream.
+    # BZip2 allows multiple output streams to be written into a single BZip2 file.
+    #
+    # Notes:
+    # * The `bzip2` linux command line utility supports multiple embedded BZip2 stream,
+    #   but some other custom implementations may not. They may only read the first slice and stop.
+    # * It is only designed for use on output collections.
+    #
+    # To download the output when using this slice:
+    #
+    #   # Download the binary BZip2 streams into a single file
+    #   IOStreams.path(output_file_name).stream(:none).writer do |io|
+    #     job.download { |slice| io << slice[:binary] }
+    #   end
+    class BZip2OutputSlice < ::RocketJob::Sliced::Slice
+      # This is a specialized binary slice for creating binary data from each slice
+      # that must be downloaded as-is into output files.
+      def self.binary?
+        true
+      end
+
+      private
+
+      def parse_records
+        records = attributes.delete("records")
+
+        # Convert BSON::Binary to a string
+        @records = [{binary: records.data}]
+      end
+
+      def serialize_records
+        return [] if @records.nil? || @records.empty?
+
+        lines = records.to_a.join("\n")
+        s = StringIO.new
+        IOStreams::Bzip2::Writer.stream(s) { |io| io.write(lines) }
+        BSON::Binary.new(s.string)
+      end
+    end
+  end
+end

data/lib/rocket_job/sliced/input.rb

@@ -5,7 +5,7 @@ module RocketJob
         # Create indexes before uploading
         create_indexes
         Writer::Input.collect(self, on_first: on_first, &block)
-      rescue StandardError => e
+      rescue Exception => e
         drop
         raise(e)
       end
@@ -73,7 +73,7 @@ module RocketJob
           count    += 1
         end
         count
-      rescue StandardError => e
+      rescue Exception => e
         drop
         raise(e)
       end
@@ -91,7 +91,7 @@ module RocketJob
           count  += 1
         end
         count
-      rescue StandardError => e
+      rescue Exception => e
         drop
         raise(e)
       end

data/lib/rocket_job/sliced/slice.rb

@@ -94,6 +94,12 @@ module RocketJob
         end
       end
 
+      # Returns whether this is a specialized binary slice for creating binary data from each slice
+      # that is then just downloaded as-is into output files.
+      def self.binary?
+        false
+      end
+
       # `records` array has special handling so that it can be modified in place instead of having
       # to replace the entire array every time. For example, when appending lines with `<<`.
       def records

data/lib/rocket_job/sliced/slices.rb

@@ -42,6 +42,12 @@ module RocketJob
         slice
       end
 
+      # Returns whether this collection contains specialized binary slices for creating binary data from each slice
+      # that is then just downloaded as-is into output files.
+      def binary?
+        slice_class.binary?
+      end
+
       # Returns output slices in the order of their id
       # which is usually the order in which they were written.
       def each

data/lib/rocket_job/version.rb

@@ -1,3 +1,3 @@
 module RocketJob
-  VERSION = "5.3.3".freeze
+  VERSION = "5.4.0.beta1".freeze
 end

data/lib/rocketjob.rb

@@ -29,6 +29,7 @@ module RocketJob
   autoload :Worker,             "rocket_job/worker"
   autoload :Performance,        "rocket_job/performance"
   autoload :Server,             "rocket_job/server"
+  autoload :Sliced,             "rocket_job/sliced"
   autoload :Subscriber,         "rocket_job/subscriber"
   autoload :Supervisor,         "rocket_job/supervisor"
   autoload :ThrottleDefinition, "rocket_job/throttle_definition"
@@ -48,10 +49,6 @@ module RocketJob
       autoload :Transaction,         "rocket_job/plugins/job/transaction"
       autoload :Worker,              "rocket_job/plugins/job/worker"
     end
-    module Rufus
-      autoload :CronLine,       "rocket_job/plugins/rufus/cron_line"
-      autoload :ZoTime,         "rocket_job/plugins/rufus/zo_time"
-    end
     autoload :Cron,             "rocket_job/plugins/cron"
     autoload :Document,         "rocket_job/plugins/document"
     autoload :ProcessingWindow, "rocket_job/plugins/processing_window"
@@ -80,21 +77,6 @@ module RocketJob
     end
   end
 
-  module Sliced
-    autoload :CompressedSlice,  "rocket_job/sliced/compressed_slice"
-    autoload :EncryptedSlice,   "rocket_job/sliced/encrypted_slice"
-    autoload :Input,            "rocket_job/sliced/input"
-    autoload :Output,           "rocket_job/sliced/output"
-    autoload :Slice,            "rocket_job/sliced/slice"
-    autoload :Slices,           "rocket_job/sliced/slices"
-    autoload :Store,            "rocket_job/sliced/store"
-
-    module Writer
-      autoload :Input,          "rocket_job/sliced/writer/input"
-      autoload :Output,         "rocket_job/sliced/writer/output"
-    end
-  end
-
   module Subscribers
     autoload :Logger,           "rocket_job/subscribers/logger"
     autoload :Server,           "rocket_job/subscribers/server"

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: rocketjob
 version: !ruby/object:Gem::Version
-  version: 5.3.3
+  version: 5.4.0.beta1
 platform: ruby
 authors:
 - Reid Morrison
-autorequire: 
+autorequire:
 bindir: bin
 cert_chain: []
-date: 2020-07-16 00:00:00.000000000 Z
+date: 2020-10-21 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: aasm
@@ -108,8 +108,8 @@ dependencies:
     - - "~>"
       - !ruby/object:Gem::Version
         version: '1.3'
-description: 
-email: 
+description:
+email:
 executables:
 - rocketjob
 - rocketjob_perf
@@ -186,6 +186,8 @@ files:
 - lib/rocket_job/server.rb
 - lib/rocket_job/server/model.rb
 - lib/rocket_job/server/state_machine.rb
+- lib/rocket_job/sliced.rb
+- lib/rocket_job/sliced/bzip2_output_slice.rb
 - lib/rocket_job/sliced/compressed_slice.rb
 - lib/rocket_job/sliced/encrypted_slice.rb
 - lib/rocket_job/sliced/input.rb
@@ -210,7 +212,7 @@ homepage: http://rocketjob.io
 licenses:
 - Apache-2.0
 metadata: {}
-post_install_message: 
+post_install_message:
 rdoc_options: []
 require_paths:
 - lib
@@ -221,12 +223,12 @@ required_ruby_version: !ruby/object:Gem::Requirement
       version: '2.3'
 required_rubygems_version: !ruby/object:Gem::Requirement
   requirements:
-  - - ">="
+  - - ">"
     - !ruby/object:Gem::Version
-      version: '0'
+      version: 1.3.1
 requirements: []
 rubygems_version: 3.0.8
-signing_key: 
+signing_key:
 specification_version: 4
 summary: Ruby's missing batch processing system.
 test_files: []