Skip to content

Latest commit

 

History

History
160 lines (109 loc) · 5.02 KB

README.rst

File metadata and controls

160 lines (109 loc) · 5.02 KB

mqttwrapper

A little glue package to make it simple to quickly put together scripts that bridge MQTT and other libraries. See examples below.

Installation

Install from PyPI:

$ pip install mqttwrapper

By default paho-mqtt will be used as the MQTT library, but you can use hbmqtt if you wish, though support for this is deprecated and will be removed in the future. To install, use:

$ pip install mqttwrapper[hbmqtt]

Usage

mqttwrapper.run_script handles the setup/maintenance of the MQTT connection and subscriptions to topics, and calls a callback function when messages are received. It can run in blocking mode where it will take care of the MQTT client handling or in non blocking mode where a new thread takes care this leaving your main thread free for other things.

The simplest script looks something like this:

from mqttwrapper import run_script

def callback(topic, payload):
    print("Received payload {} on topic {}".format(payload, topic))

def main():
    run_script(callback, broker="mqtt://127.0.0.1", topics=["my/awesome/topic1", "another/awesome/topic2"])

or in non blocking mode:

def main():
    run_script(callback, broker="mqtt://127.0.0.1", topics=["my/awesome/topic1", "another/awesome/topic2"], blocking=False)
    while True:
        echo "Alive and kicking"
        time.sleep(1)

Any extra keyword arguments passed to run_script are passed back to the callback function when it's called:

from mqttwrapper import run_script

def callback(topic, payload, context, foo):
    assert context == "hello"
    assert foo == "bar"
    print("Received payload {} on topic {}".format(payload, topic))

def main():
    run_script(callback, broker="mqtt://127.0.0.1", topics=["my/awesome/topic1", "another/awesome/topic2"], context="hello", foo="bar")

If you don't need any context passed to your callback, just the topic and payload of messages, you can run in 'importless' mode by creating a callback module which has a callback function defined, then execute the mqttwrapper.run module:

$ cat callback.py
def callback(topic, payload):
    print(topic, payload)

$ python -m mqttwrapper.run

Retained messages

You can ignore MQTT retained messages by passing ignore_retained=True when calling run_script. NB this is currently not supported by the hbmqtt backend.

Publishing messages from the callback

Sometimes your callback function might want to publish its own MQTT messages, perhaps in reply to or an acknowledgement of a received message. This is possible by returning an iterable of (topic, payload) tuples from the callback, e.g.:

def callback(topic, payload):
  print("Received payload {} on topic {}".format(payload, topic))
  return [
    ("{}/ack".format(topic), payload)
  ]

mqttwrapper will take care of publishing these messages to the broker.

These tuples can optionally take the form (topic, payload, True) if you want the published message to be marked as retained.

Configuration

broker and topics can be omitted from the run_script call and environment variables used instead:

  • MQTT_BROKER: MQTT broker to connect to, e.g. mqtt://127.0.0.1/
  • MQTT_TOPICS: Comma-separated list of topics to subscribe to, e.g. my/topic1,my/topic2

asyncio/hbmqtt

NB This feature is deprecated and will be removed in the future.

The asyncio-powered hbmqtt MQTT library can be used instead, if you like:

from mqttwrapper.hbmqtt_backend import run_script

async def callback(topic, payload):
    print("Received payload {} on topic {}".format(payload, topic))

Note that your callback must be an awaitable in this case.

Your callback may require context arguments which themselves are async objects or awaitables which poses a challenge: how to set these up outside of an asyncio event loop before calling run_script? In this case, you can pass a context_callback awaitable as a kwarg to run_script. This is run at the start of the MQTT loop, and should return a dict which will be merged into the kwargs which are passed to your callback. For example:

from mqttwrapper.hbmqtt_backend import run_script

async def setup_db():
  return {
    "query_db": query_db
  }

async def query_db(value):
  # pretend this is some slow DB query, for example.
  await asyncio.sleep(3)
  return value * 2

async def callback(topic, payload, query_db):
    db_result = await query_db(int(payload))
    print("Received payload {} on topic {}, db result: {}".format(payload, topic, db_result))

def main():
    run_script(callback, context_callback=setup_db)

NB hbmqtt's reconnection handling does not resubscribe to topics upon reconnection, and mqttwrapper does not yet work around this.

Examples