Django-nameko

+ Add a new feature in heartbeat_check thread to cleanup replies that remain for longer than timeout duration, it would take from 10-20 loop cycle to clear these remaining message in replies that no handler pick up. This will reduce memory consumption of long running process that use async rpc call
+ Speed up the graceful shutdown progress when heartbeat is enabled.

+ Add new feature to dispatch event right from django-nameko
+ Support RBMQ Heart Beat feature
+ Fix bugs

+Rename global variable pool to nameko_global_pools to avoid conflict
+Add a behaviour of resetting context_data to original pool.context_data once RpcClusterProxy exit() out of block
+Minor changes to reading version from VERSION file in setup.py

**Notes:**
'POOL_SIZE' is equivalent to NAMEKO_POOL_SIZE
'POOL_CONTEXT_DATA' is equivalent to NAMEKO_CONTEXT_DATA
'POOL_TIMEOUT' is equivalent to NAMEKO_TIMEOUT
in the previous release.
Old configuration with a single pool still works as normal. No code changes required
This version setting config is compatible with the config of version 0.1,0.2,0.3 but not 0.4

NAMEKO_CONFIG={
'default': {
'AMQP_URL': 'amqp://',
'POOL_SIZE': 4,
'POOL_CONTEXT_DATA': {"common": "multi"},
'POOL_TIMEOUT': None
},
'pool1': {
'AMQP_URL': 'amqp://pool2',
'POOL_CONTEXT_DATA': {"name": "pool1", "data": 123},
},
'pool2': {
'AMQP_URL': 'amqp://pool3',
'POOL_CONTEXT_DATA': {"name": "pool2", "data": 321},
'POOL_TIMEOUT': 60
},
'pool3': {
'POOL_SIZE': 8,
'POOL_TIMEOUT': 60
}
}