Thrift Connector

Simple connection pool for thrift. thrift_connector can be used both for native thrift client and thriftpy client with minor difference in usage.

Examples can be found in examples as well as the following sections.

How it works

It maintains a connection pool, and it'll do a precheck(calling ping) to validate the connection to ensure the connection is always usable.

Options

service Defined thrift service. It should be the container class of apis for thriftpy (thriftpy2.load('...').XXXService), and api module for thrift (XXX_sdk.xxx.xxxService) host Server host port Server port [timeout] Socket timeout, in seconds. [name] Connection pool name, for identity. [raise_empty] Whether to raise exception if pool is empty while trying to obtain a connection. [max_conn] Number of connections to manage in pool. [connection_class] Connection class implementation. Builtin classes are: ThriftClient for native thrift, ThriftPyClient and ThriftPyCyClient for thriftpy, the latter one utilizes Cython for better performance. [keepalive] Seconds each connection is able to stay alive. If oen connection has lived longer than this period, it will be closed.

Usage

Examples for thriftpy is:

.. code:: python

import thriftpy2
import thrift_connector.connection_pool as connection_pool

service = thriftpy2.load("pingpong_app/pingpong.thrift")
pool = connection_pool.ClientPool(
    service.PingService,
    'localhost',
    8880,
    connection_class=connection_pool.ThriftPyCyClient
    )

print "Sending Ping..."
print "Receive:", pool.ping()
print "Winning the match..."
print "Receive:", pool.win()

Examples for thrift is:

.. code:: python

# -*- coding: utf-8 -*-

from pingpong_app.pingpong_sdk.pingpong import PingService
import connection_pool

pool = connection_pool.ClientPool(
    PingService,
    'localhost',
    8880,
    connection_class=connection_pool.ThriftClient
    )

print "Sending Ping..."
print "Receive:", pool.ping()
print "Winning the match..."
print "Receive:", pool.win()

Examples for Tornado clients can be found in examples.

Test

.. code:: python

py.test tests