karafka 1.3.7 → 1.4.3

data/lib/karafka/contracts/consumer_group.rb

@@ -32,6 +32,7 @@ module Karafka
         required(:offset_retention_time).maybe(:integer)
         required(:heartbeat_interval).filled { (int? | float?) & gteq?(0) }
         required(:fetcher_max_queue_size).filled(:int?, gt?: 0)
+        required(:assignment_strategy).value(:any)
         required(:connect_timeout).filled { (int? | float?) & gt?(0) }
         required(:reconnect_timeout).filled { (int? | float?) & gteq?(0) }
         required(:socket_timeout).filled { (int? | float?) & gt?(0) }
@@ -70,13 +71,13 @@ module Karafka
 
       # Uri rule to check if uri is in a Karafka acceptable format
       rule(:seed_brokers) do
-        if value&.is_a?(Array) && !value.all?(&method(:kafka_uri?))
+        if value.is_a?(Array) && !value.all?(&method(:kafka_uri?))
           key.failure(:invalid_broker_schema)
         end
       end
 
       rule(:topics) do
-        if value&.is_a?(Array)
+        if value.is_a?(Array)
           names = value.map { |topic| topic[:name] }
 
           key.failure(:topics_names_not_unique) if names.size != names.uniq.size
@@ -84,7 +85,7 @@ module Karafka
       end
 
       rule(:topics) do
-        if value&.is_a?(Array)
+        if value.is_a?(Array)
           value.each_with_index do |topic, index|
             TOPIC_CONTRACT.call(topic).errors.each do |error|
               key([:topics, index, error.path[0]]).failure(error.text)
@@ -93,6 +94,10 @@ module Karafka
         end
       end
 
+      rule(:assignment_strategy) do
+        key.failure(:does_not_respond_to_call) unless value.respond_to?(:call)
+      end
+
       rule(:ssl_client_cert, :ssl_client_cert_key) do
         if values[:ssl_client_cert] && !values[:ssl_client_cert_key]
           key(:ssl_client_cert_key).failure(:ssl_client_cert_with_ssl_client_cert_key)

data/lib/karafka/contracts.rb

@@ -5,6 +5,6 @@ module Karafka
   module Contracts
     # Regexp for validating format of groups and topics
     # @note It is not nested inside of the contracts, as it is used by couple of them
-    TOPIC_REGEXP = /\A(\w|\-|\.)+\z/.freeze
+    TOPIC_REGEXP = /\A(\w|-|\.)+\z/.freeze
   end
 end

data/lib/karafka/helpers/class_matcher.rb

@@ -8,7 +8,7 @@ module Karafka
     class ClassMatcher
       # Regexp used to remove any non classy like characters that might be in the consumer
       # class name (if defined dynamically, etc)
-      CONSTANT_REGEXP = %r{[?!=+\-\*/\^\|&\[\]<>%~\#\:\s\(\)]}.freeze
+      CONSTANT_REGEXP = %r{[?!=+\-*/\^|&\[\]<>%~\#:\s()]}.freeze
 
       private_constant :CONSTANT_REGEXP
 

data/lib/karafka/instrumentation/logger.rb

@@ -29,17 +29,17 @@ module Karafka
 
       # @return [Karafka::Helpers::MultiDelegator] multi delegator instance
       #   to which we will be writing logs
-      # We use this approach to log stuff to file and to the STDOUT at the same time
+      # We use this approach to log stuff to file and to the $stdout at the same time
       def target
         Karafka::Helpers::MultiDelegator
           .delegate(:write, :close)
-          .to(STDOUT, file)
+          .to($stdout, file)
       end
 
       # Makes sure the log directory exists as long as we can write to it
       def ensure_dir_exists
         FileUtils.mkdir_p(File.dirname(log_path))
-      rescue Errno::EACCES
+      rescue Errno::EACCES, Errno::EROFS
         nil
       end
 

data/lib/karafka/instrumentation/stdout_listener.rb

@@ -43,7 +43,7 @@ module Karafka
         # so it returns a topic as a string, not a routing topic
         debug(
           <<~MSG.chomp.tr("\n", ' ')
-            Params deserialization for #{event[:caller].topic} topic
+            Params deserialization for #{event[:caller].metadata.topic} topic
             successful in #{event[:time]} ms
           MSG
         )
@@ -52,7 +52,9 @@ module Karafka
       # Logs unsuccessful deserialization attempts of incoming data
       # @param event [Dry::Events::Event] event details including payload
       def on_params_params_deserialize_error(event)
-        error "Params deserialization error for #{event[:caller].topic} topic: #{event[:error]}"
+        topic = event[:caller].metadata.topic
+        error = event[:error]
+        error "Params deserialization error for #{topic} topic: #{error}"
       end
 
       # Logs errors that occurred in a listener fetch loop

data/lib/karafka/params/batch_metadata.rb

@@ -0,0 +1,26 @@
+# frozen_string_literal: true
+
+module Karafka
+  module Params
+    # Simple batch metadata object that stores all non-message information received from Kafka
+    # cluster while fetching the data
+    # @note This metadata object refers to per batch metadata, not `#params.metadata`
+    BatchMetadata = Struct.new(
+      :batch_size,
+      :first_offset,
+      :highwater_mark_offset,
+      :unknown_last_offset,
+      :last_offset,
+      :offset_lag,
+      :deserializer,
+      :partition,
+      :topic,
+      keyword_init: true
+    ) do
+      # @return [Boolean] is the last offset known or unknown
+      def unknown_last_offset?
+        unknown_last_offset
+      end
+    end
+  end
+end

data/lib/karafka/params/builders/batch_metadata.rb

@@ -0,0 +1,30 @@
+# frozen_string_literal: true
+
+module Karafka
+  module Params
+    module Builders
+      # Builder for creating batch metadata object based on the batch informations
+      module BatchMetadata
+        class << self
+          # Creates metadata based on the kafka batch data
+          # @param kafka_batch [Kafka::FetchedBatch] kafka batch details
+          # @param topic [Karafka::Routing::Topic] topic for which we've fetched the batch
+          # @return [Karafka::Params::BatchMetadata] batch metadata object
+          def from_kafka_batch(kafka_batch, topic)
+            Karafka::Params::BatchMetadata.new(
+              batch_size: kafka_batch.messages.count,
+              first_offset: kafka_batch.first_offset,
+              highwater_mark_offset: kafka_batch.highwater_mark_offset,
+              unknown_last_offset: kafka_batch.unknown_last_offset?,
+              last_offset: kafka_batch.last_offset,
+              offset_lag: kafka_batch.offset_lag,
+              deserializer: topic.deserializer,
+              partition: kafka_batch.partition,
+              topic: topic.name
+            ).freeze
+          end
+        end
+      end
+    end
+  end
+end

data/lib/karafka/params/builders/params.rb

@@ -12,22 +12,24 @@ module Karafka
         class << self
           # @param kafka_message [Kafka::FetchedMessage] message fetched from Kafka
           # @param topic [Karafka::Routing::Topic] topic for which this message was fetched
-          # @return [Karafka::Params::Params] params object
+          # @return [Karafka::Params::Params] params object with payload and message metadata
           def from_kafka_message(kafka_message, topic)
-            Karafka::Params::Params
-              .new
-              .merge!(
-                'create_time' => kafka_message.create_time,
-                'headers' => kafka_message.headers || {},
-                'is_control_record' => kafka_message.is_control_record,
-                'key' => kafka_message.key,
-                'offset' => kafka_message.offset,
-                'deserializer' => topic.deserializer,
-                'partition' => kafka_message.partition,
-                'receive_time' => Time.now,
-                'topic' => kafka_message.topic,
-                'payload' => kafka_message.value
-              )
+            metadata = Karafka::Params::Metadata.new(
+              create_time: kafka_message.create_time,
+              headers: kafka_message.headers || {},
+              is_control_record: kafka_message.is_control_record,
+              key: kafka_message.key,
+              offset: kafka_message.offset,
+              deserializer: topic.deserializer,
+              partition: kafka_message.partition,
+              receive_time: Time.now,
+              topic: topic.name
+            ).freeze
+
+            Karafka::Params::Params.new(
+              kafka_message.value,
+              metadata
+            )
           end
         end
       end

data/lib/karafka/params/builders/params_batch.rb

@@ -12,11 +12,11 @@ module Karafka
           # @param topic [Karafka::Routing::Topic] topic for which we're received messages
           # @return [Karafka::Params::ParamsBatch<Karafka::Params::Params>] batch with params
           def from_kafka_messages(kafka_messages, topic)
-            params_array = kafka_messages.map! do |message|
+            params_array = kafka_messages.map do |message|
               Karafka::Params::Builders::Params.from_kafka_message(message, topic)
             end
 
-            Karafka::Params::ParamsBatch.new(params_array)
+            Karafka::Params::ParamsBatch.new(params_array).freeze
           end
         end
       end

data/lib/karafka/params/metadata.rb

@@ -2,34 +2,19 @@
 
 module Karafka
   module Params
-    # Simple metadata object that stores all non-message information received from Kafka cluster
-    # while fetching the data
-    class Metadata < Hash
-      # Attributes that should be accessible as methods as well (not only hash)
-      METHOD_ATTRIBUTES = %w[
-        batch_size
-        first_offset
-        highwater_mark_offset
-        last_offset
-        offset_lag
-        deserializer
-        partition
-        topic
-      ].freeze
-
-      private_constant :METHOD_ATTRIBUTES
-
-      METHOD_ATTRIBUTES.each do |attr|
-        # Defines a method call accessor to a particular hash field.
-        define_method(attr) do
-          self[attr]
-        end
-      end
-
-      # @return [Boolean] is the last offset known or unknown
-      def unknown_last_offset?
-        self['unknown_last_offset']
-      end
-    end
+    # Single message / params metadata details that can be accessed without the need for the
+    # payload deserialization
+    Metadata = Struct.new(
+      :create_time,
+      :headers,
+      :is_control_record,
+      :key,
+      :offset,
+      :deserializer,
+      :partition,
+      :receive_time,
+      :topic,
+      keyword_init: true
+    )
   end
 end

data/lib/karafka/params/params.rb

@@ -6,68 +6,44 @@ module Karafka
     # It provides lazy loading not only until the first usage, but also allows us to skip
     # using deserializer until we execute our logic. That way we can operate with
     # heavy-deserialization data without slowing down the whole application.
-    class Params < Hash
-      # Params attributes that should be available via a method call invocation for Kafka
-      # client compatibility.
-      # Kafka passes internally Kafka::FetchedMessage object and the ruby-kafka consumer
-      # uses those fields via method calls, so in order to be able to pass there our params
-      # objects, have to have same api.
-      METHOD_ATTRIBUTES = %w[
-        create_time
-        headers
-        is_control_record
-        key
-        offset
-        deserializer
-        deserialized
-        partition
-        receive_time
-        topic
-        payload
-      ].freeze
+    class Params
+      extend Forwardable
 
-      private_constant :METHOD_ATTRIBUTES
+      attr_reader :raw_payload, :metadata
 
-      METHOD_ATTRIBUTES.each do |attr|
-        # Defines a method call accessor to a particular hash field.
-        # @note Won't work for complex key names that contain spaces, etc
-        # @param key [Symbol] name of a field that we want to retrieve with a method call
-        # @example
-        #   key_attr_reader :example
-        #   params.example #=> 'my example payload'
-        define_method(attr) do
-          self[attr]
-        end
-      end
+      def_delegators :metadata, *Metadata.members
 
-      # Overrides `Hash#[]` to allow lazy deserialization of payload. This allows us to fetch
-      # metadata without actually triggering deserialization of the payload until it is needed
-      # @param key [String, Symbol] hash key
-      # @return [Object] content of a given params key
-      def [](key)
-        # Payload will be deserialized only when we request for it.
-        deserialize! if key == 'payload'
-        super
+      # @param raw_payload [Object] incoming payload before deserialization
+      # @param metadata [Karafka::Params::Metadata] message metadata object
+      def initialize(raw_payload, metadata)
+        @raw_payload = raw_payload
+        @metadata = metadata
+        @deserialized = false
+        @payload = nil
       end
 
-      # @return [Karafka::Params::Params] This method will trigger deserializer execution. If we
-      #   decide to retrieve data, deserializer will be executed to get data. Output of that will
-      #   be merged to the current object. This object will be also marked as already deserialized,
-      #   so we won't deserialize it again.
-      def deserialize!
-        return self if self['deserialized']
+      # @return [Object] lazy-deserialized data (deserialized upon first request)
+      def payload
+        return @payload if deserialized?
+
+        @payload = deserialize
+        # We mark deserialization as successful after deserialization, as in case of an error
+        # this won't be falsely set to true
+        @deserialized = true
+        @payload
+      end
 
-        self['deserialized'] = true
-        self['payload'] = deserialize
-        self
+      # @return [Boolean] did given params payload were deserialized already
+      def deserialized?
+        @deserialized
       end
 
       private
 
-      # @return [Object] deserialized data
+      # @return [Object] tries de-serializes data
       def deserialize
         Karafka.monitor.instrument('params.params.deserialize', caller: self) do
-          self['deserializer'].call(self)
+          metadata.deserializer.call(self)
         end
       rescue ::StandardError => e
         Karafka.monitor.instrument('params.params.deserialize.error', caller: self, error: e)

data/lib/karafka/params/params_batch.rb

@@ -15,47 +15,46 @@ module Karafka
         @params_array = params_array
       end
 
-      # @yieldparam [Karafka::Params::Params] each deserialized and loaded params instance
-      # @note Invocation of this method will cause loading and deserializing each param after
-      #   another. If you want to get access without deserializing, please access params_array
-      #   directly
+      # @yieldparam [Karafka::Params::Params] each params instance
+      # @note Invocation of this method will not cause loading and deserializing each param after
+      #   another.
       def each
-        @params_array.each { |param| yield(param.deserialize!) }
+        @params_array.each { |param| yield(param) }
       end
 
       # @return [Array<Karafka::Params::Params>] returns all the params in a loaded state, so they
       #   can be used for batch insert, etc. Without invoking all, up until first use, they won't
       #   be deserialized
       def deserialize!
-        each(&:itself)
+        each(&:payload)
       end
 
       # @return [Array<Object>] array with deserialized payloads. This method can be useful when
       #   we don't care about metadata and just want to extract all the data payloads from the
       #   batch
       def payloads
-        deserialize!.map(&:payload)
+        map(&:payload)
       end
 
-      # @return [Karafka::Params::Params] first element after the deserialization process
+      # @return [Karafka::Params::Params] first element
       def first
-        @params_array.first.deserialize!
+        @params_array.first
       end
 
-      # @return [Karafka::Params::Params] last element after the deserialization process
+      # @return [Karafka::Params::Params] last element
       def last
-        @params_array.last.deserialize!
-      end
-
-      # @return [Array<Karafka::Params::Params>] pure array with params (not deserialized)
-      def to_a
-        @params_array
+        @params_array.last
       end
 
       # @return [Integer] number of messages in the batch
       def size
         @params_array.size
       end
+
+      # @return [Array<Karafka::Params::Params>] pure array with params
+      def to_a
+        @params_array
+      end
     end
   end
 end

data/lib/karafka/routing/builder.rb

@@ -16,6 +16,7 @@ module Karafka
       private_constant :CONTRACT
 
       def initialize
+        super
         @draws = Concurrent::Array.new
       end
 

data/lib/karafka/routing/consumer_group.rb

@@ -8,9 +8,11 @@ module Karafka
     class ConsumerGroup
       extend Helpers::ConfigRetriever
 
-      attr_reader :topics
-      attr_reader :id
-      attr_reader :name
+      attr_reader(
+        :topics,
+        :id,
+        :name
+      )
 
       # @param name [String, Symbol] raw name of this consumer group. Raw means, that it does not
       #   yet have an application client_id namespace, this will be added here by default.

data/lib/karafka/serialization/json/deserializer.rb

@@ -17,8 +17,8 @@ module Karafka
         #   }
         #   Deserializer.call(params) #=> { 'a' => 1 }
         def call(params)
-          ::MultiJson.load(params['payload'])
-        rescue ::MultiJson::ParseError => e
+          params.raw_payload.nil? ? nil : ::JSON.parse(params.raw_payload)
+        rescue ::JSON::ParserError => e
           raise ::Karafka::Errors::DeserializationError, e
         end
       end

data/lib/karafka/server.rb

@@ -58,7 +58,7 @@ module Karafka
       def stop_supervised
         Karafka::App.stop!
 
-        # Temporary patch until https://github.com/dry-rb/dry-configurable/issues/93 is fixed
+        # See https://github.com/dry-rb/dry-configurable/issues/93
         timeout = Thread.new { Karafka::App.config.shutdown_timeout }.join.value
 
         # We check from time to time (for the timeout period) if all the threads finished

data/lib/karafka/setup/config.rb

@@ -89,6 +89,9 @@ module Karafka
         #   are stored for further processing. Note, that each item in the queue represents a
         #   response from a single broker
         setting :fetcher_max_queue_size, 10
+        # option assignment_strategy [Object] a strategy determining the assignment of
+        #   partitions to the consumers.
+        setting :assignment_strategy, Karafka::AssignmentStrategies::RoundRobin.new
         # option max_bytes_per_partition [Integer] the maximum amount of data fetched
         #   from a single partition at a time.
         setting :max_bytes_per_partition, 1_048_576

data/lib/karafka/version.rb

@@ -3,5 +3,5 @@
 # Main module namespace
 module Karafka
   # Current Karafka version
-  VERSION = '1.3.7'
+  VERSION = '1.4.3'
 end

data/lib/karafka.rb

@@ -6,10 +6,10 @@
   waterdrop
   kafka
   envlogic
+  json
   thor
   forwardable
   fileutils
-  multi_json
   dry-configurable
   dry-validation
   dry/events/publisher

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: karafka
 version: !ruby/object:Gem::Version
-  version: 1.3.7
+  version: 1.4.3
 platform: ruby
 authors:
 - Maciej Mensfeld
@@ -36,7 +36,7 @@ cert_chain:
   2DND//YJUikn1zwbz1kT70XmHd97B4Eytpln7K+M1u2g1pHVEPW4owD/ammXNpUy
   nt70FcDD4yxJQ+0YNiHd0N8IcVBM1TMIVctMNQ==
   -----END CERTIFICATE-----
-date: 2020-08-11 00:00:00.000000000 Z
+date: 2021-03-24 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: dry-configurable
@@ -122,20 +122,6 @@ dependencies:
     - - "~>"
       - !ruby/object:Gem::Version
         version: '1.0'
-- !ruby/object:Gem::Dependency
-  name: multi_json
-  requirement: !ruby/object:Gem::Requirement
-    requirements:
-    - - ">="
-      - !ruby/object:Gem::Version
-        version: '1.12'
-  type: :runtime
-  prerelease: false
-  version_requirements: !ruby/object:Gem::Requirement
-    requirements:
-    - - ">="
-      - !ruby/object:Gem::Version
-        version: '1.12'
 - !ruby/object:Gem::Dependency
   name: rake
   requirement: !ruby/object:Gem::Requirement
@@ -184,14 +170,14 @@ dependencies:
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: 1.3.0
+        version: 1.4.0
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: 1.3.0
+        version: 1.4.0
 - !ruby/object:Gem::Dependency
   name: zeitwerk
   requirement: !ruby/object:Gem::Requirement
@@ -241,6 +227,7 @@ files:
 - karafka.gemspec
 - lib/karafka.rb
 - lib/karafka/app.rb
+- lib/karafka/assignment_strategies/round_robin.rb
 - lib/karafka/attributes_map.rb
 - lib/karafka/backends/inline.rb
 - lib/karafka/base_consumer.rb
@@ -251,6 +238,7 @@ files:
 - lib/karafka/cli/flow.rb
 - lib/karafka/cli/info.rb
 - lib/karafka/cli/install.rb
+- lib/karafka/cli/missingno.rb
 - lib/karafka/cli/server.rb
 - lib/karafka/code_reloader.rb
 - lib/karafka/connection/api_adapter.rb
@@ -259,9 +247,9 @@ files:
 - lib/karafka/connection/client.rb
 - lib/karafka/connection/listener.rb
 - lib/karafka/connection/message_delegator.rb
+- lib/karafka/consumers/batch_metadata.rb
 - lib/karafka/consumers/callbacks.rb
 - lib/karafka/consumers/includer.rb
-- lib/karafka/consumers/metadata.rb
 - lib/karafka/consumers/responders.rb
 - lib/karafka/consumers/single_params.rb
 - lib/karafka/contracts.rb
@@ -280,7 +268,8 @@ files:
 - lib/karafka/instrumentation/monitor.rb
 - lib/karafka/instrumentation/proctitle_listener.rb
 - lib/karafka/instrumentation/stdout_listener.rb
-- lib/karafka/params/builders/metadata.rb
+- lib/karafka/params/batch_metadata.rb
+- lib/karafka/params/builders/batch_metadata.rb
 - lib/karafka/params/builders/params.rb
 - lib/karafka/params/builders/params_batch.rb
 - lib/karafka/params/metadata.rb
@@ -331,7 +320,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.1.2
+rubygems_version: 3.2.3
 signing_key: 
 specification_version: 4
 summary: Ruby based framework for working with Apache Kafka

data/lib/karafka/consumers/metadata.rb

@@ -1,10 +0,0 @@
-# frozen_string_literal: true
-
-module Karafka
-  module Consumers
-    # Brings the metadata into consumers that support batch_fetching
-    module Metadata
-      attr_accessor :metadata
-    end
-  end
-end

data/lib/karafka/params/builders/metadata.rb

@@ -1,33 +0,0 @@
-# frozen_string_literal: true
-
-module Karafka
-  module Params
-    module Builders
-      # Builder for creating metadata object based on the message or batch informations
-      # @note We have 2 ways of creating metadata based on the way ruby-kafka operates
-      module Metadata
-        class << self
-          # Creates metadata based on the kafka batch data
-          # @param kafka_batch [Kafka::FetchedBatch] kafka batch details
-          # @param topic [Karafka::Routing::Topic] topic for which we've fetched the batch
-          # @return [Karafka::Params::Metadata] metadata object
-          def from_kafka_batch(kafka_batch, topic)
-            Karafka::Params::Metadata
-              .new
-              .merge!(
-                'batch_size' => kafka_batch.messages.count,
-                'first_offset' => kafka_batch.first_offset,
-                'highwater_mark_offset' => kafka_batch.highwater_mark_offset,
-                'last_offset' => kafka_batch.last_offset,
-                'offset_lag' => kafka_batch.offset_lag,
-                'deserializer' => topic.deserializer,
-                'partition' => kafka_batch.partition,
-                'topic' => kafka_batch.topic,
-                'unknown_last_offset' => kafka_batch.unknown_last_offset?
-              )
-          end
-        end
-      end
-    end
-  end
-end