Topical

A simple, lightweight event pipeline in Python.

Installation

pip install ap-topical

Usage

import logging, asyncio
import ap.topical
logging.basicConfig(level=logging.INFO)
logger = logging.getLogger(__name__)

@ap.topical.event('start')
async def start_handler(payload):
    """Do stuff here"""

    await ap.topical.publish('end', payload)


@ap.topical.event('start')
async def start_auditor(payload):
    logger.info('auditing start event')

@ap.topical.event('end')
async def end_handler(payload):
    """Do stuff here"""


async main():
    await ap.topical.publish('start', {})

if __name__ == '__main__':
    asyncio.run(main())

Events

Events are strings (or other hashable objects) that are published via ap.topical.publish(). These are used as both a topic and a specific event.

Subscribing Callbacks to Events

You can subscribe a callback to a given event using ap.topical.subscribe as in the following example:

async def my_event_handler(payload):
    ...

ap.topical.subscribe('my-event', my_event_handler)

Additionally, you can use the handy @ap.topical.event() decorator to do this for you automatically:

@ap.topical.event('my-event')
async def my_event_handler(payload):
    ...

Publishing Events

Events can be published by providing an event and a payload to pass between event handlers.

await ap.topical.publish('my-event', {})

Multiple Handlers

A given event can have any number of handlers (limited by available compute resources). These will all trigger asynchronously.

@ap.topical.event('start')
async def start_handler(payload):
    ...

@ap.topical.event('start')
async def start_auditor(payload):
    ...

await ap.topical.publish('start', {})

This allows you to create robust graphs of event handlers.

Future Plans

• add unsubscribe functionality
• add wildcard event support
• add default event payload support
• add payload validation via msgspec