asynchronous

trame.app.asynchronous.create_task(coroutine, loop=None)

Create a task from a coroutine while also attaching a done callback so any exception or error could be caught and reported.

Parameters:

  • coroutine – A coroutine to execute as an independent task

  • loop – Optionally provide the loop on which the task should be scheduled on. By default we will use the current running loop.

Returns:

The decorated task

Return type:

asyncio.Task

trame.app.asynchronous.decorate_task(task)

Decorate a task by attaching a done callback so any exception or error could be caught and reported.

Parameters:

task (asyncio.Task) – A coroutine to execute as an independent task

Returns:

The same task object

Return type:

asyncio.Task

trame.app.asynchronous.create_state_queue_monitor_task(server, queue, delay=1)

Create and schedule a task to watch over the provided queue to update a server state. This is especially useful when using a multiprocess executor and you want to report progress into your current server.

Parameters:

  • server (trame_server.core.Server) – A coroutine to execute as an independent task

  • queue (multiprocessing.Queue) – A queue instance meant to exchange state from the parallel process to the given server

  • delay (float) – Time to sleep in seconds before processing the queue once emptied

Returns:

The monitoring task

Return type:

asyncio.Task

class trame.app.asynchronous.StateQueue(queue, auto_flush=True)

Bases: object

Class use to decorate a multiprocessing.Queue inside your external process to simulate your server state object.

Parameters:

  • queue (multiprocessing.Queue) – A queue instance meant to exchange state from the parallel process to the given server

  • auto_flush (Boolean) – Should you manage the state update phase or just propagate as soon as you update a property

property queue

Provide access to the decorated queue

update(_dict)

Update the distributed state from a set of key/value pair

Parameters:

_dict (dict) – A dict containing one or many key/value pair

flush()

Explicitly push any local change to the queue.

exit()

Release the monitoring task as we are done with our work

trame.app.asynchronous.task(func)

Function decorator to make its async execution within a task