| Python RabbitMQ Publisher and Receiver |
|---|
A rabbitmq publisher and receiver that implements a safe and reliable rabbitmq publisher and listener. |
| Changelog |
- Changed the logging of message reeived and acknowledged to debug mode instead of info mode.
* Adding extra module for RabbitMQ Publisher. This update has backward incompatibility for Receiver class
- Reconnecting bug fix. Applies to all the branch.
- Logging of message body changed to DEBUG mode instead of INFO
- Wrote a basic RabbitMQ Receiver
This package rmq-pub-sub contains two modules - rmqproducer and rmqreceiver. The module rmqreceiver has a Receiver class which can be imported to implement a rabbitmq consumer. This class contains a lot of functions for tasks like: to make connection to rabbitmq server, to change configuration of exchange and queue binding, to start listing the messages, to safely stop the connection. SelectConnection is being used by Receiver class for its asynchronous design. Similarly the module rmqproducer has a Publisher class which can be imported to implement a rabbitmq publisher.
- To install the latest version of the package, user can use the command:
pip install git+https://github.com/loconsolutions/python-rabbitmq-pubsub.git
- To install a specific version x.x use the following command:
pip install git+https://github.com/loconsolutions/python-rabbitmq-pubsub.git@version_x.x
- for example to install version 0.1 command to be used will be:
pip install git+https://github.com/loconsolutions/python-rabbitmq-pubsub.git@version_0.1
- To uninstall the package use the command:
pip uninstall rmq-pub-sub
For documentation one can refer to the code in file rmqreceiver/rabbitmq_receiver.py and rmqproducer/rabbitmq_producer.py
To use RabbitMQ listener import the Receiver class and based on what behaviour you want from the RabbitMQ listner, pass the parameter values during initalizing the class. Here is the list of parameters (including optional params) to be passed on initializing Receiver class:
- param method consumer_callback
The method to callback when consuming (messages) with the signature consumer_callback(channel, method, properties, body), where
channel: pika.Channel, method: pika.spec.Basic.Deliver, properties: pika.spec.BasicProperties, body: str, unicode, or bytes (python 3.x)
- param str amqp_url
The AMQP url to connect with
- param str exchange
Name of exchange
- param str exchange_type
The exchange type to use. If no vaue is given for exchange type, it will assume that the exchange already exists and will use the existing exchange.
- param str queue
Name of the queue. Its default value is ''. When the queue name is empty string i.e. '', server chooses a random queue name for us.
- param list binding_keys
The list of binding keys to be used. It's a list of strings. It's default value is [None]
- param bool queue_exclusive
Only allow access by the current connection. This is is the exclusive flag used in queue_declare() function of pika channel. If the flag is true, consumer queue is deleted on disconnection. It's default value is False. If server is going to choose a name for queue we set this variable True irrespective of what value user has given for queue_exclusive
- param bool queue_durable
Survive reboots of the broker. This is the durable flag used in queue_declare() function of pika channel. If this flag is True, messages in queue are saved on disk in case RabbitMQ quits or crashes. It's default value is True
- param bool no_ack
Tell the broker to not expect a response (acknowledgement). It's default value is False
- param bool safe_stop
If this option is True, system will try to gracefully stop the connection if the process is killed (with SIGTERM signal). Its default value is True
Use run() function to start the RabbitMQ listener. It will then keep on consuming the messages. Use stop() function to stop the listner whenever you want. Logging of all the events is already added in the class.
Similarly to use RabbitMQ publisher import the Publisher class and based on what behaviour you want from the RabbitMQ publisher, pass the parameter values during initalizing the class. Here is the list of parameters (including optional params) to be passed on initializing Producer class:
- param str amqp_url
The AMQP url to connect with
- param str exchange
Name of exchange
- param str exchange_type
The exchange type to use. It's default value is topic
- param bool exchange_durable
Survive a reboot of RabbitMQ. This is the durable flag used in exchange_declare() function of pika channel. It's default value is True
- param bool exchange_auto_delete
Remove when no more queues are bound to it. This is the auto_delete flag used in exchange_declare() function of pika channel. It's default value is False
- param bool exchange_internal
Can only be published to by other exchanges. This is the internal flag used in exchange_declare() function of pika channel. It's default value is False
- param bool delivery_confirmation
If the confirmation of published message is required. It's default value is True.
- param method nack_callback
The method to callback when publishing of a message fails. Signature of the method: nack_callback(failed_message) where failed_message is the message which failed
- param bool safe_stop
If this option is True, system will try to gracefully stop the connection if the process is killed (with SIGTERM signal). Its default value is True
- param reconnect_time
The number of seconds after which connection will automatically restart if it accidently stops. Its default value is 5 seconds.
Simply initialize the class, start publishing the message using publish_message() method and stop() when done publishing. Inside the code we are maintaining a connection pool. Users are strongly recommended to use stop() method after they are done with the publishing of messages so that connection can be sent back to the pool and reused by some other user saving the cost of creating a new connection
Here is the sample code to use the rabbitmq receiver.
from rmq import Receiver
def consumer_callback(unused_channel, basic_deliver, properties, body):
#do something.
print "The message received is: %s" % body
def main():
try:
url = 'amqp://guest:guest@127.0.0.1:5672/%2F'
exchange = 'something.something'
exchange_type = 'topic'
binding_key = 'something.something.*'
queue_name='my_queue'
my_receiver = Receiver(consumer_callback, url, exchange,
binding_keys=[binding_key], queue=queue_name,
queue_durable=True, queue_exclusive=False)
# Since we haven't passed the exchange_type, it will connect to
# existing exchange instead of initializing a new one on its own
my_receiver.run()
# Since safe_stop option is True (by default), when a kill
# process signal is raised my_receiver.stop() function will be
# automatically called before the process ends
except KeyboardInterrupt:
my_receiver.stop()
if __name__ == '__main__':
main()A sample code to use the rabbitmq publisher
import time
import logging
from rmq import Publisher
logging.basicConfig(level=logging.INFO)
my_publisher = Publisher(
'amqp://guest:guest@localhost:5672/%2F?connection_attempts=3&heartbeat_interval=3600', 'my.exchange.name')
# Make sure exchange doesn't already exist with different properties
for count in range(1, 6):
my_publisher.publish_message(
"message number {num}".format(num=count), 'my.routing.key')
time.sleep(1)
my_publisher.stop()
# Users are strongly recommended to use stop() method after they are done
# with the publishing of messages so that connection can be sent back to
# the connection pool and reused by some other user saving the cost of
# creating a new connection