flask_nameko

A wrapper for using nameko services with Flask

Installation

Install it via PyPI:

pip install flask_nameko

Usage

To start using flask_nameko, you need to create and configure a new FlaskPooledClusterRpcProxy singleton, which you'll use to communicate with your Nameko cluster.

# __init__.py
from flask import Flask
from flask_nameko import FlaskPooledClusterRpcProxy

rpc = FlaskPooledClusterRpcProxy()

def create_app():
    app = Flask(__name__)
    app.config.update(dict(
        NAMEKO_AMQP_URI='amqp://localhost'
    ))

    rpc.init_app(app)

app = create_app()

Then, you can use the FlaskPooledClusterRpcProxy singleton just as you would normally use a ClusterRpcProxy, by accessing individual services by name and calling methods on them:

# routes.py

from . import (
    app,
    rpc
)

@app.route('/'):
def index():
    result = rpc.service.do_something('test')
    return result

API

Configuration

FlaskPooledClusterRpcProxy accepts all nameko configuration values, prefixed with the NAMEKO_ prefix. In addition, it exposes additional configuration options:

NAMEKO_INITIAL_CONNECTIONS (int, default=2) - the number of initial connections to the Nameko cluster to create
NAMEKO_MAX_CONNECTIONS (int, default=8) - the max number of connections to the Nameko cluster to create before raises an error
NAMEKO_CONNECT_ON_METHOD_CALL (bool, default=True) - whether connections to services should be loaded when the service is accessed (False) or when a method is called on a service (True)
NAMEKO_RPC_TIMEOUT (int, default=None) - the default timeout before raising an error when trying to call a service RPC method
NAMEKO_POOL_RECYCLE (int, default=None) - if specified, connections that are older than this interval, specified in seconds, will be automatically recycled on checkout. This setting is useful for environments where connections are happening through a proxy like HAProxy which automatically closes connections after a specified interval.

Proxies

flask_nameko.FlaskPooledClusterRpcProxy(app=None, connect_on_method_call=True)

This class is used to create a pool of connections to a Nameko cluster. It provides the following options:

   * `connect_on_method_call` - if this is true, the connection to a service is created when a method is called on a service rather than when the service is accessed

init_app(app=None)

Configure the proxy for a given app.

Development

$ git clone git@github.com:jessepollak/flask_nameko.git flask_nameko
$ cd flask_nameko
$ make develop