Skip to main content

Kafka

Kafka is a distributed messaging system that is used to publish and subscribe to streams of records. This demo shows how to use KafkaChatMessageHistory to store and retrieve chat messages from a Kafka cluster.

A running Kafka cluster is required to run the demo. You can follow this instruction to create a Kafka cluster locally.

from langchain_community.chat_message_histories import KafkaChatMessageHistory

chat_session_id = "chat-message-history-kafka"
bootstrap_servers = "localhost:64797"  # host:port. `localhost:Plaintext Ports` if setup Kafka cluster locally
history = KafkaChatMessageHistory(
    chat_session_id,
    bootstrap_servers,
)

Optional parameters to construct KafkaChatMessageHistory:

  • ttl_ms: Time to live in milliseconds for the chat messages.
  • partition: Number of partition of the topic to store the chat messages.
  • replication_factor: Replication factor of the topic to store the chat messages.

KafkaChatMessageHistory internally uses Kafka consumer to read chat messages, and it has the ability to mark the consumed position persistently. It has following methods to retrieve chat messages:

  • messages: continue consuming chat messages from last one.
  • messages_from_beginning: reset the consumer to the beginning of the history and consume messages. Optional parameters:
    1. max_message_count: maximum number of messages to read.
    2. max_time_sec: maximum time in seconds to read messages.
  • messages_from_latest: reset the consumer to the end of the chat history and try consuming messages. Optional parameters same as above.
  • messages_from_last_consumed: return messages continuing from the last consumed message, similar to messages, but with optional parameters.

max_message_count and max_time_sec are used to avoid blocking indefinitely when retrieving messages. As a result, messages and other methods to retrieve messages may not return all messages in the chat history. You will need to specify max_message_count and max_time_sec to retrieve all chat history in a single batch.

Add messages and retrieve.

history.add_user_message("hi!")
history.add_ai_message("whats up?")

history.messages
[HumanMessage(content='hi!'), AIMessage(content='whats up?')]

Calling messages again returns an empty list because the consumer is at the end of the chat history.

history.messages
[]

Add new messages and continue consuming.

history.add_user_message("hi again!")
history.add_ai_message("whats up again?")
history.messages
[HumanMessage(content='hi again!'), AIMessage(content='whats up again?')]

To reset the consumer and read from beginning:

history.messages_from_beginning()
[HumanMessage(content='hi again!'),
 AIMessage(content='whats up again?'),
 HumanMessage(content='hi!'),
 AIMessage(content='whats up?')]

Set the consumer to the end of the chat history, add a couple of new messages, and consume:

history.messages_from_latest()
history.add_user_message("HI!")
history.add_ai_message("WHATS UP?")
history.messages
[HumanMessage(content='HI!'), AIMessage(content='WHATS UP?')]

Was this page helpful?