adori/fastapi-cloud-tasks

GCP's Cloud Tasks + Cloud Scheduler + FastAPI = Partial replacement for celery.

fastapi-cloud-tasks's Issues

Add local instructions with cloud tasks emulator

https://github.com/aertje/cloud-tasks-emulator

Deployment

How would you deploy this?

Question: Is the project abandoned, and do you need help?

Hi @ofpiyush, are you still maintaining this library?

The library is very useful, and I'm not actually missing features, but it is a little difficult to use the library with the latest developments in Pydantic and FastAPI, and it would be nice to either get an official patch with #8 or a more permanent fix/update.

Thank you for this library, and if you need any help with maintenance, let me know.

.delay() shoud be async

The example in README shows how to queue a task with sync .delay() function, which is not a good idea in async handler.
We have remote IO there: usually low-latency, but no warranty. So .delay() blocks asyncio loop when called from async route.

README presents examples with async routes calling sync .delay() call example (e.g. "https://github.com/Adori/fastapi-cloud-tasks#local"). This might be misleading for a less experienced user.

Making .delay(...) async would allow to call

await make_dinner.delay(...)

or even

task = create_task(make_dinner.delay(...))
# make something else here in the meanwhile
# ...
# ensure the task was enqueued before the request has been finalized
await task  

Cloud Tasks library already provides async variant of the client for some time, and it works well:
https://cloud.google.com/python/docs/reference/cloudtasks/latest/google.cloud.tasks_v2.services.cloud_tasks.CloudTasksAsyncClient

Add a differences from Celery section