magic_foundation
Minimalistic library that simplifies the adoption of async/await (asyncio) programming style in a multithreaded application.
Define 2 services and run in the same runloop (single thread)
import logging
from magic_foundation import Service, ServiceStatus, ServiceContext, Main
class TestService(Service):
def __init__(self, name:str):
self.name = name
async def initialize(self, ctx:ServiceContext):
log.info(f"[{self.name}] initialize")
async def run(self, ctx:ServiceContext):
index = 0
while self.status is ServiceStatus.running:
log.info(f"[{self.name}] run {index}")
index += 1
await asyncio.sleep(1.0)
async def terminate(self, ctx:ServiceContext):
log.info(f"[{self.name}] terminate")
if __name__ == "__main__":
log = logging.getLogger(__name__)
logging.basicConfig(level=logging.INFO)
logging.getLogger("magic_foundation").setLevel(logging.DEBUG)
main = Main.instance()
main.service_pools = {
'main': [
TestService(name="Service_1"),
TestService(name="Service_2")
],
}
main.run()If you want running the 2 services in different threads just define the service_pools as following:
...
main = Main.instance()
main.service_pools = {
'th1': [
TestService(name="Service_1")
],
'th2': [
TestService(name="Service_2")
],
}
main.run()Communication
The communication between Services mimics the publish/subscribe paradigm
Subscribe
async def coro(data):
log.info(f"[{self.name}] coro data:{data}")
await ctx.subscribe(queue_name="q://my_queue", handler=coro)Pubblish
data = {
"a": "A",
"b": 1234
}
await ctx.publish(queue_name="q://my_queue", data=data) It is possible to subscribe to the same queue from different services and each service will receive the message
WebSocket Service
As an additionan component the library provides a built-in websocket service.
from magic_foundation import Container, Service, ServiceStatus, ServiceContext, Main
from magic_foundation.websocket_service import WebSocketService
...
class TestService(Service):
...
async def run(self, ctx:ServiceContext):
log.info(f"[{self.name}] run")
async def handler(data):
log.info(f"[{self.name}] handler inbound data:{data}")
req = json.loads(data)
res = json.dumps({"status": "OK", "timestamp": req["timestamp"]})
await ctx.publish(queue_name="ws://outbound/client", data=res)
await ctx.subscribe(queue_name="ws://inbound/client", handler=handler)
...
async def terminate(self, ctx:ServiceContext):
log.info(f"[{self.name}] terminate")
...
if __name__ == "__main__":
main = Main.instance()
main.service_pools = {
"main" : [
WebSocketService(host="localhost", port=8765),
TestService(name="Consumer"),
]
}
main.run()The prefixes ws://inbound/ and ws://outbound/ refer to the WS endpoint (WebSocketService) and must be considered reserved.
The complete example is contained in the examples folder.
To trying it run in a shell the command: python examples/websocket.py and in a browser open the file websocket.html.
Limitations
Right now only the requested path is use as discriminant, so messages from different clients with the same path are routed to the same handler and vice versa one outbound message is sent to all clients connected to the same path.
Simple Logging Service
The included logging service is a simpple way to dump json maggase to a local file.
from magic_foundation import Container, Service, ServiceStatus, ServiceContext, Main
from magic_foundation.logging_service import LoggingService
...
file_path = "logging_out.log"
class TestService(Service):
...
async def run(self, ctx:ServiceContext):
log.info(f"[{self.name}] run")
...
async def terminate(self, ctx:ServiceContext):
log.info(f"[{self.name}] terminate")
index = 0;
while self.status == ServiceStatus.running:
index += 1
data = {"cmd": "log", "index": index}
await ctx.publish(queue_name=f"log://{file_path}", data=data)
await asyncio.sleep(1.0)
...
if __name__ == "__main__":
main = Main.instance()
main.service_pools = {
"backgound" : [
LoggingService(file_path=file_path, flush_interval_sec=4.0)
],
"main" : [
TestService(name="Producer")
]
}
main.run()The prefixes log:// refers to the simple logging service (LoggingService) and must be considered reserved.
The complete example is contained in the examples folder.
To trying it run in a shell the command: python examples/dump_to_file.py. In the root of the project a file called logging_out.log will be created and every 4 seconds the logging service will flush the collected messages.
