Skip to content

bluesky/bluesky-kafka

Repository files navigation

bluesky-kafka

Kafka integration for bluesky.

  • Free software: 3-clause BSD license

Features

  • BlueskyConsumer
  • MongoConsumer
  • Publisher
  • RemoteDispatcher

Test

Install docker and docker-compose.

Start a Kafka server:

$ cd bluesky-kafka/scripts
$ sudo docker-compose -f bitnami-kafka-docker-compose.yml up

Run tests:

$ cd bluesky-kafka
$ pytest

Optionally increase logging output to the console by specifying a logging level:

$ pytest --log-cli-level=DEBUG

Run a Mongo Consumer Group

Create a conda environment:

$ conda create -n consumers python=3.8
$ conda activate consumers

Install packages:

$ pip install bluesky-kafka supervisor

Setup environment variables: mongo_uri reference: https://docs.mongodb.com/manual/reference/connection-string/ bootstrap_servers: comma-separated list of brokers.

$ export BLUESKY_MONGO_URI="mongodb://username:password@machine1:port1,machine2:port2,machine3:port3
$ export KAFKA_BOOTSTRAP_SERVERS="machine1:9092, machine2:9092, machine3:9092"

Update the bluesky_kafka/supervisor/supervisord.conf file with the correct path for your installation.

Start the consumer processes:

$ supervisord -c bluesky_kafka/supervisor/supervisord.conf

Monitor the consumer processes:

$ supervisorctl -c bluesky_kafka/supervisor/supervisorctl.conf

Remote Best Effort Callback

To run bluesky.callbacks.best_effort.BestEffortCallback consuming from a Kafka topic use:

import matplotlib.pyplot as plt

from bluesky.callbacks.best_effort import BestEffortCallback
from bluesky_kafka import RemoteDispatcher

bec = BestEffortCallback

kafka_config = {...}

kafka_consumer = RemoteDispatcher(**kafka_config)
kafka_consumer.subscribe(bec)
kafka_consumer.start(work_during_wait=lambda: plt.pause(0.05))