org.apache.kafka.streams.processor

Class TopologyBuilder

java.lang.Object

org.apache.kafka.streams.processor.TopologyBuilder

Direct Known Subclasses:

KStreamBuilder

public class TopologyBuilder
extends Object

A component that is used to build a ProcessorTopology. A topology contains an acyclic graph of sources, processors, and sinks. A source is a node in the graph that consumes one or more Kafka topics and forwards them to its child nodes. A processor is a node in the graph that receives input records from upstream nodes, processes that records, and optionally forwarding new records to one or all of its children. Finally, a sink is a node in the graph that receives records from upstream nodes and writes them to a Kafka topic. This builder allows you to construct an acyclic graph of these nodes, and the builder is then passed into a new KafkaStreams instance that will then begin consuming, processing, and producing records.

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	TopologyBuilder.TopicsInfo

Constructor Summary

Constructors

Constructor and Description
TopologyBuilder() Create a new builder.

Method Summary

Methods

Modifier and Type	Method and Description
TopologyBuilder	addInternalTopic(String topicName) Adds an internal topic
TopologyBuilder	addProcessor(String name, ProcessorSupplier supplier, String... parentNames) Add a new processor node that receives and processes records output by one or more parent source or processor node.
<K,V> TopologyBuilder	addSink(String name, String topic, Serializer<K> keySerializer, Serializer<V> valSerializer, StreamPartitioner<K,V> partitioner, String... parentNames) Add a new sink that forwards records from upstream parent processor and/or source nodes to the named Kafka topic.
TopologyBuilder	addSink(String name, String topic, Serializer keySerializer, Serializer valSerializer, String... parentNames) Add a new sink that forwards records from upstream parent processor and/or source nodes to the named Kafka topic.
TopologyBuilder	addSink(String name, String topic, StreamPartitioner partitioner, String... parentNames) Add a new sink that forwards records from upstream parent processor and/or source nodes to the named Kafka topic, using the supplied partitioner.
TopologyBuilder	addSink(String name, String topic, String... parentNames) Add a new sink that forwards records from upstream parent processor and/or source nodes to the named Kafka topic.
TopologyBuilder	addSource(String name, Deserializer keyDeserializer, Deserializer valDeserializer, Pattern topicPattern) Add a new source that consumes from topics matching the given pattern and forwards the records to child processor and/or sink nodes.
TopologyBuilder	addSource(String name, Deserializer keyDeserializer, Deserializer valDeserializer, String... topics) Add a new source that consumes the named topics and forwards the records to child processor and/or sink nodes.
TopologyBuilder	addSource(String name, Pattern topicPattern) Add a new source that consumes from topics matching the given pattern and forward the records to child processor and/or sink nodes.
TopologyBuilder	addSource(String name, String... topics) Add a new source that consumes the named topics and forward the records to child processor and/or sink nodes.
TopologyBuilder	addStateStore(StateStoreSupplier supplier, String... processorNames) Adds a state store
ProcessorTopology	build(Integer topicGroupId) Build the topology for the specified topic group.
TopologyBuilder	connectProcessorAndStateStores(String processorName, String... stateStoreNames) Connects the processor and the state stores
TopologyBuilder	connectProcessors(String... processorNames) Connects a list of processors.
protected TopologyBuilder	connectSourceStoreAndTopic(String sourceStoreName, String topic)
Collection<Set<String>>	copartitionGroups() Returns the copartition groups.
TopologyBuilder	copartitionSources(Collection<String> sourceNodes) Asserts that the streams of the specified source nodes must be copartitioned.
Map<Integer,Set<String>>	nodeGroups() Returns the map of node groups keyed by the topic group id.
TopologyBuilder	setApplicationId(String applicationId) Set the applicationId to be used for auto-generated internal topics.
Pattern	sourceTopicPattern()
Set<String>	sourceTopics() Get the names of topics that are to be consumed by the source nodes created by this builder.
Map<String,Set<String>>	stateStoreNameToSourceTopics()
Map<Integer,TopologyBuilder.TopicsInfo>	topicGroups() Returns the map of topic groups keyed by the group id.
void	updateSubscriptions(SubscriptionUpdates subscriptionUpdates, String threadId)

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

TopologyBuilder

public TopologyBuilder()

Create a new builder.

Method Detail

setApplicationId

public final TopologyBuilder setApplicationId(String applicationId)

Set the applicationId to be used for auto-generated internal topics. This is required before calling sourceTopics(), topicGroups(), copartitionSources(java.util.Collection<java.lang.String>), stateStoreNameToSourceTopics and build(Integer).

Parameters:

applicationId - the streams applicationId. Should be the same as set by StreamsConfig.APPLICATION_ID_CONFIG

addSource

public final TopologyBuilder addSource(String name,
                        String... topics)

Add a new source that consumes the named topics and forward the records to child processor and/or sink nodes. The source will use the default key deserializer and default value deserializer specified in the stream configuration.

Parameters:

name - the unique name of the source used to reference this node when adding processor children.

topics - the name of one or more Kafka topics that this source is to consume

Returns:

this builder instance so methods can be chained together; never null

addSource

public final TopologyBuilder addSource(String name,
                        Pattern topicPattern)

Add a new source that consumes from topics matching the given pattern and forward the records to child processor and/or sink nodes. The source will use the default key deserializer and default value deserializer specified in the stream configuration.

Parameters:

name - the unique name of the source used to reference this node when adding processor children.

topicPattern - regular expression pattern to match Kafka topics that this source is to consume

Returns:

this builder instance so methods can be chained together; never null

addSource

public final TopologyBuilder addSource(String name,
                        Deserializer keyDeserializer,
                        Deserializer valDeserializer,
                        String... topics)

Add a new source that consumes the named topics and forwards the records to child processor and/or sink nodes. The source will use the specified key and value deserializers.

Parameters:

name - the unique name of the source used to reference this node when adding processor children.

keyDeserializer - the key deserializer used when consuming records; may be null if the source should use the default key deserializer specified in the stream configuration

valDeserializer - the value deserializer used when consuming records; may be null if the source should use the default value deserializer specified in the stream configuration

topics - the name of one or more Kafka topics that this source is to consume

Returns:

this builder instance so methods can be chained together; never null

Throws:

TopologyBuilderException - if processor is already added or if topics have already been registered by another source

addSource

public final TopologyBuilder addSource(String name,
                        Deserializer keyDeserializer,
                        Deserializer valDeserializer,
                        Pattern topicPattern)

Add a new source that consumes from topics matching the given pattern and forwards the records to child processor and/or sink nodes. The source will use the specified key and value deserializers. The provided de-/serializers will be used for all matched topics, so care should be taken to specify patterns for topics that share the same key-value data format.

Parameters:

name - the unique name of the source used to reference this node when adding processor children.

keyDeserializer - the key deserializer used when consuming records; may be null if the source should use the default key deserializer specified in the stream configuration

valDeserializer - the value deserializer used when consuming records; may be null if the source should use the default value deserializer specified in the stream configuration

topicPattern - regular expression pattern to match Kafka topics that this source is to consume

Returns:

this builder instance so methods can be chained together; never null

Throws:

TopologyBuilderException - if processor is already added or if topics have already been registered by name

addSink

public final TopologyBuilder addSink(String name,
                      String topic,
                      String... parentNames)

Add a new sink that forwards records from upstream parent processor and/or source nodes to the named Kafka topic. The sink will use the default key serializer and default value serializer specified in the stream configuration.

Parameters:

name - the unique name of the sink

topic - the name of the Kafka topic to which this sink should write its records

parentNames - the name of one or more source or processor nodes whose output records this sink should consume and write to its topic

Returns:

this builder instance so methods can be chained together; never null

See Also:

addSink(String, String, StreamPartitioner, String...), addSink(String, String, Serializer, Serializer, String...), addSink(String, String, Serializer, Serializer, StreamPartitioner, String...)

addSink

public final TopologyBuilder addSink(String name,
                      String topic,
                      StreamPartitioner partitioner,
                      String... parentNames)

Add a new sink that forwards records from upstream parent processor and/or source nodes to the named Kafka topic, using the supplied partitioner. The sink will use the default key serializer and default value serializer specified in the stream configuration.

The sink will also use the specified StreamPartitioner to determine how records are distributed among the named Kafka topic's partitions. Such control is often useful with topologies that use state stores in its processors. In most other cases, however, a partitioner needs not be specified and Kafka will automatically distribute records among partitions using Kafka's default partitioning logic.

Parameters:

name - the unique name of the sink

topic - the name of the Kafka topic to which this sink should write its records

partitioner - the function that should be used to determine the partition for each record processed by the sink

parentNames - the name of one or more source or processor nodes whose output records this sink should consume and write to its topic

Returns:

this builder instance so methods can be chained together; never null

See Also:

addSink(String, String, String...), addSink(String, String, Serializer, Serializer, String...), addSink(String, String, Serializer, Serializer, StreamPartitioner, String...)

addSink

public final TopologyBuilder addSink(String name,
                      String topic,
                      Serializer keySerializer,
                      Serializer valSerializer,
                      String... parentNames)

Add a new sink that forwards records from upstream parent processor and/or source nodes to the named Kafka topic. The sink will use the specified key and value serializers.

Parameters:

name - the unique name of the sink

topic - the name of the Kafka topic to which this sink should write its records

keySerializer - the key serializer used when consuming records; may be null if the sink should use the default key serializer specified in the stream configuration

valSerializer - the value serializer used when consuming records; may be null if the sink should use the default value serializer specified in the stream configuration

parentNames - the name of one or more source or processor nodes whose output records this sink should consume and write to its topic

Returns:

this builder instance so methods can be chained together; never null

See Also:

addSink(String, String, String...), addSink(String, String, StreamPartitioner, String...), addSink(String, String, Serializer, Serializer, StreamPartitioner, String...)

addSink

public final <K,V> TopologyBuilder addSink(String name,
                            String topic,
                            Serializer<K> keySerializer,
                            Serializer<V> valSerializer,
                            StreamPartitioner<K,V> partitioner,
                            String... parentNames)

Add a new sink that forwards records from upstream parent processor and/or source nodes to the named Kafka topic. The sink will use the specified key and value serializers, and the supplied partitioner.

Parameters:

name - the unique name of the sink

topic - the name of the Kafka topic to which this sink should write its records

keySerializer - the key serializer used when consuming records; may be null if the sink should use the default key serializer specified in the stream configuration

valSerializer - the value serializer used when consuming records; may be null if the sink should use the default value serializer specified in the stream configuration

partitioner - the function that should be used to determine the partition for each record processed by the sink

parentNames - the name of one or more source or processor nodes whose output records this sink should consume and write to its topic

Returns:

this builder instance so methods can be chained together; never null

Throws:

TopologyBuilderException - if parent processor is not added yet, or if this processor's name is equal to the parent's name

See Also:

addSink(String, String, String...), addSink(String, String, StreamPartitioner, String...), addSink(String, String, Serializer, Serializer, String...)

addProcessor

public final TopologyBuilder addProcessor(String name,
                           ProcessorSupplier supplier,
                           String... parentNames)

Add a new processor node that receives and processes records output by one or more parent source or processor node. Any new record output by this processor will be forwarded to its child processor or sink nodes.

Parameters:

name - the unique name of the processor node

supplier - the supplier used to obtain this node's Processor instance

parentNames - the name of one or more source or processor nodes whose output records this processor should receive and process

Returns:

this builder instance so methods can be chained together; never null

Throws:

TopologyBuilderException - if parent processor is not added yet, or if this processor's name is equal to the parent's name

addStateStore

public final TopologyBuilder addStateStore(StateStoreSupplier supplier,
                            String... processorNames)

Adds a state store

Parameters:

supplier - the supplier used to obtain this state store StateStore instance

Returns:

this builder instance so methods can be chained together; never null

Throws:

TopologyBuilderException - if state store supplier is already added

connectProcessorAndStateStores

public final TopologyBuilder connectProcessorAndStateStores(String processorName,
                                             String... stateStoreNames)

Connects the processor and the state stores

Parameters:

processorName - the name of the processor

stateStoreNames - the names of state stores that the processor uses

Returns:

this builder instance so methods can be chained together; never null

connectSourceStoreAndTopic

protected final TopologyBuilder connectSourceStoreAndTopic(String sourceStoreName,
                                         String topic)

connectProcessors

public final TopologyBuilder connectProcessors(String... processorNames)

Connects a list of processors. NOTE this function would not needed by developers working with the processor APIs, but only used for the high-level DSL parsing functionalities.

Parameters:

processorNames - the name of the processors

Returns:

this builder instance so methods can be chained together; never null

Throws:

TopologyBuilderException - if less than two processors are specified, or if one of the processors is not added yet

addInternalTopic

public final TopologyBuilder addInternalTopic(String topicName)

Adds an internal topic

Parameters:

topicName - the name of the topic

Returns:

this builder instance so methods can be chained together; never null

copartitionSources

public final TopologyBuilder copartitionSources(Collection<String> sourceNodes)

Asserts that the streams of the specified source nodes must be copartitioned.

Parameters:

sourceNodes - a set of source node names

Returns:

this builder instance so methods can be chained together; never null

nodeGroups

public Map<Integer,Set<String>> nodeGroups()

Returns the map of node groups keyed by the topic group id.

Returns:

groups of node names

build

public ProcessorTopology build(Integer topicGroupId)

Build the topology for the specified topic group. This is called automatically when passing this builder into the KafkaStreams.KafkaStreams(TopologyBuilder, org.apache.kafka.streams.StreamsConfig) constructor.

See Also:

KafkaStreams.KafkaStreams(TopologyBuilder, org.apache.kafka.streams.StreamsConfig)

topicGroups

public Map<Integer,TopologyBuilder.TopicsInfo> topicGroups()

Returns the map of topic groups keyed by the group id. A topic group is a group of topics in the same task.

Returns:

groups of topic names

sourceTopics

public Set<String> sourceTopics()

Get the names of topics that are to be consumed by the source nodes created by this builder.

Returns:

the unmodifiable set of topic names used by source nodes, which changes as new sources are added; never null

stateStoreNameToSourceTopics

public Map<String,Set<String>> stateStoreNameToSourceTopics()

Returns:

a mapping from state store name to a Set of source Topics.

copartitionGroups

public Collection<Set<String>> copartitionGroups()

Returns the copartition groups. A copartition group is a group of source topics that are required to be copartitioned.

Returns:

groups of topic names

sourceTopicPattern

public Pattern sourceTopicPattern()

updateSubscriptions

public void updateSubscriptions(SubscriptionUpdates subscriptionUpdates,
                       String threadId)