Huge News!Announcing our $40M Series B led by Abstract Ventures.Learn More
Socket
Sign inDemoInstall
Socket

airflow-provider-kafka

Package Overview
Dependencies
Maintainers
1
Alerts
File Explorer

Advanced tools

Socket logo

Install Socket

Detect and block malicious and high-risk dependencies

Install

airflow-provider-kafka

Apache Airflow Kafka provider containing Deferrable Operators & Sensors.

  • 0.2.2
  • PyPI
  • Socket score

Maintainers
1

Kafka Airflow Provider

GitHub release (latest by date)PyPIPyPI - Downloads

An airflow provider to:

  • interact with kafka clusters
  • read from topics
  • write to topics
  • wait for specific messages to arrive to a topic

This package currently contains

3 hooks (airflow_provider_kafka.hooks) :

  • admin_client.KafkaAdminClientHook - a hook to work against the actual kafka admin client
  • consumer.KafkaConsumerHook - a hook that creates a consumer and provides it for interaction
  • producer.KafkaProducerHook - a hook that creates a producer and provides it for interaction

4 operators (airflow_provider_kafka.operators) :

  • await_message.AwaitKafkaMessageOperator - a deferable operator (sensor) that awaits to encounter a message in the log before triggering down stream tasks.
  • consume_from_topic.ConsumeFromTopicOperator - an operator that reads from a topic and applies a function to each message fetched.
  • produce_to_topic.ProduceToTopicOperator - an operator that uses a iterable to produce messages as key/value pairs to a kafka topic.
  • event_triggers_function.EventTriggersFunctionOperator - an operator that listens for messages on the topic and then triggers a downstream function before going back to listening.

1 trigger airflow_provider_kafka.triggers :

  • await_message.AwaitMessageTrigger

Quick start

pip install airflow-provider-kafka

Example usages :

  • basic read/write/sense on a topic
  • event listener pattern

FAQs

Why confluent kafka and not (other library) ? A few reasons: the confluent-kafka library is guaranteed to be 1:1 functional with librdkafka, is faster, and is maintained by a company with a commercial stake in ensuring the continued quality and upkeep of it as a product.

Why not release this into airflow directly ? I could probably make the PR and get it through, but the airflow code base is getting huge and I don't want to burden the maintainers with code that they don't own for maintainence. Also there's been multiple attempts to get a Kafka provider in before and this is just faster.

Why is most of the configuration handled in a dict ? Because that's how confluent-kafka does it. I'd rather maintain interfaces that people already using kafka are comfortable with as a starting point - I'm happy to add more options/ interfaces in later but would prefer to be thoughtful about it to ensure that there difference between these operators and the actual client interface are minimal.

Local Development

Unit Tests

Unit tests are located at tests/unit, a kafka server isn't required to run these tests. execute with pytest

Setup on M1 Mac

Installing on M1 chip means a brew install of the librdkafka library before you can pip install confluent-kafka

brew install librdkafka
export C_INCLUDE_PATH=/opt/homebrew/Cellar/librdkafka/1.8.2/include
export LIBRARY_PATH=/opt/homebrew/Cellar/librdkafka/1.8.2/lib
pip install confluent-kafka

FAQs


Did you know?

Socket

Socket for GitHub automatically highlights issues in each pull request and monitors the health of all your open source dependencies. Discover the contents of your packages and block harmful activity before you install or update your dependencies.

Install

Related posts

SocketSocket SOC 2 Logo

Product

  • Package Alerts
  • Integrations
  • Docs
  • Pricing
  • FAQ
  • Roadmap
  • Changelog

Packages

npm

Stay in touch

Get open source security insights delivered straight into your inbox.


  • Terms
  • Privacy
  • Security

Made with ⚡️ by Socket Inc