A client library for a publish-subscribe service.
- It exposes a simplified API to clients that suits ployst conventions.
- Serves as an anticorruption layer to isolate us from the chosen provider.
Current implementation supports Google Cloud PubSub.
pubsub automatically handles the creation of topics and subscriptions,
as well as making sure subscription names do not clash, based on the
app_name provided when instantiating PubSub.
Messages may be anything that is JSON-serializable (can be fed to
json.dumps). pubsub will take care of encoding and decoding messages in a
format compatible with the service used.
WARNING: under
Delivery contract, Google PubSub documentation states:For the most part Pub/Sub delivers each message once, and in the order in which it was published. However, once-only and in-order delivery are not guaranteed: it may happen that a message is delivered more than once, and out of order. Therefore, your subscriber should be idempotent when processing messages, and, if necessary, able to handle messages received out of order. If ordering is important, we recommend that the publisher of the topic to which you subscribe include some kind of sequence information in the message.
from pubsub.client import PubSub
project = 'ployst-proto'
app_name = 'integration-tests'
pubsub = PubSub(project, app_name)
topic = 'tests.onetwothree'
payload = 'I will be sent'
# publish
pubsub.publish(topic, [payload])
# pull and acknowledge messages
pulled_messages = self.pubsub.pull(topic)
payload = pulled_messages[0]['payload']
pubsub.acknowledge(topic, pulled_messages)
# or acknowledge individual messages:
for message in pulled_messages:
do_stuff(message)
message.ack()Unacknowledged messages will not be deleted, and may be retrieved at a later point (although not necessarily in the original sequence).
For more examples, look at the included tests.