asyncio wrapper for grpc.io. Currently, only client-side interfaces are supported.
aiogrpc has exactly the same interface as grpc.io (https://grpc.io/grpc/python/grpc.html#module-grpc), except:
call_soon_threadsafe()
to run in the same thread with the event loop__call__
for unary methods, code()
, details()
, ...) become coroutinesCompile your proto file with protoc to generate stub files
Use the stub files with aiogrpc.Channel, instead of grpc.Channel:
from aiogrpc import insecure_channel
import asyncio
from mystub import MyStub
channel = insecure_channel('ipv4:///127.0.0.1:8080')
mystub = MyStub(channel)
async def test_call():
return await mystub.mymethod(...)
async def test_call_stream():
async for v in mystub.my_stream_method(...):
...
Same as the original grpc.io, balancing (fail over) between multiple servers are supported:
Use multiple IPv4 or IPv6 addresses for balancing - use target like: "ipv4:///1.2.3.4:9999,1.2.3.5:9999,1.2.3.6:9999"
or "ipv6:///[1::2]:9999,[1::3]:9999,[1::4]:9999"
Specify ONE DNS name which resolves to multiple addresses. GRPC will balance between the resolved addresses, and try
to update the resolving result in interval. Example: "dns:///myserver:9999"
Notice that you cannot mix up IPv4, IPv6 and DNS addresses; only addresses with the same type can be balanced. Also, when using DNS target, only one DNS name can be specified.
The default load balancing strategy is "pick_first"
(fail-over). Set "grpc.lb_policy_name"
option to "round_robin"
for
round-robin load balancing: channel = insecure_channel('ipv4:///1.2.3.4:9999,1.2.3.5:9999,1.2.3.6:9999', [('grpc.lb_policy_name', 'round_robin')])
All code are published under Apache 2.0 License
Some code and docstrings are modified from grpc.io (https://grpc.io/grpc/python/grpc.html#module-grpc)
此处可能存在不合适展示的内容,页面不予展示。您可通过相关编辑功能自查并修改。
如您确认内容无涉及 不当用语 / 纯广告导流 / 暴力 / 低俗色情 / 侵权 / 盗版 / 虚假 / 无价值内容或违法国家有关法律法规的内容,可点击提交进行申诉,我们将尽快为您处理。