🚀 The scheduler that simply works. 🚀
Documentation: https://asyncz.dymmond.com 📚
Source Code: https://github.com/dymmond/asyncz
Asyncz is a scheduler for any ASGI application that needs to have those complicated scheduled operations with the best of what pydantic can offer.
Nowadays using async frameworks with python is somewhat common and becoming even more mainstream. A lot of applications usually need a complex stack of technologies to fullfil their needs and directly or indirectly, a scheduler.
There are great frameworks out there that do the task extremely well, the best example is APScheduler, which is where Asyncz came from.
To be even more honest, Asyncz is a revamp of APScheduler. Without the APScheduler there is no Asyncz, so much that even the APScheduler tests are used within asyncz. That is how great APScheduler is!
So what was the reason why recreating another similar version of APScheduler? Well, it is not entirely the same thing. Asyncz was designed to work only with ASGI and AsyncIO as well as integrating pydantic and bring the modern python into the table.
APScheduler is widely used by millions of python developers and Asyncz does not aim to replace it, instead is a more focused and optimised solution for async and ASGI frameworks out there.
See the vendors for more details.
We all struggle with the logging and the configurations and with that in mind Asyncz comes with natice support for loguru.
This will make the logging a lot easier to understand and clear to read.
What does this mean? Well, Asyncz does not need to run inside any specific framework, actually you can use it completely indepent from any framework as well as inside ASGI frameworks such as Esmerald, FastAPI, Starlette, Starlite, Quart... You can pick one and go for it.
Asyncz comes with special support to Esmerald for the simple reason that the author is the same but it can be added more support. If you are interested in adding support to your favourite frameworks then see the contributing section.
Like APScheduler, Asyncz also brings four kinds of components:
- Python 3.7+
Asyncz wouldn't be possible without two giants:
$ pip install asyncz
How do you know if you are choosing the right scheduler, triggers, stores and executors?
Well, Asyncz is intentionally designed for specific systems and already helps you out with some of those questions.
- Schedulers - Natively only supports the AsyncIOScheduler.
- Triggers - Here it will depend of the periocidity of our tasks. Example:
- CronTrigger - UNIX like cron and gives you the same feeling as scheduling a task on a native UNIX like based system.
- DateTrigger - When you need to run a task once on a specific point of time.
- IntervalTrigger - When you need to run tasks in specific intervals of time.
- OrTrigger/AndTrigger - If you would like to combine more than one trigger (cron, interval and date) together.
- Stores - Natively only supports redis, mongo and memory.
- Executors - Natively only supports AsyncIOExecutor, ThreadPoolExecutor and ProcessPoolExecutor.
Sometimes having a lot of options makes the decision making very hard and Asyncz is intentionally designed and driven to simplify and for specific use cases but is not limited to those. In every section you have the option of uilding your own stores, executors, triggers and schedulers.
Due its simplificy, Asyncz provides some ways of configuring the scheduler for you.
First way:
from asyncz.schedulers.asyncio import AsyncIOScheduler
scheduler = AsyncIOScheduler()
Second way:
from asyncz.schedulers import AsyncIOScheduler
scheduler = AsyncIOScheduler()
Initialize the rest of the application after the scheduler
initialisation.
More details can be found with more thorough explanations.
This is in simple terms and in a nutshell how to start with Asyncz quickly. For more information, details and examples how to leverage Asyncz simply navigate through the documentation and have fun 😁🎉.
Asyncz currently supports ASGI, the Esmerald framework and brings some batteries that are currently used by Esmerald and leveraging Asyncz.
from asyncz.schedulers import AsyncIOScheduler
...
# handle_lifespan is optional, set to True if you don't want to pass it down because the underlying app doesn't support it
# this is true for django
application = AsyncIOScheduler().asgi(application, handle_lifespan=False)
# or more simple (please do not use both together)
application = AsyncIOScheduler().asgi()(application)
Using with lilya:
from asyncz.schedulers import AsyncIOScheduler
# Lilya middleware doesn't pass lifespan events
app = AsyncIOScheduler().asgi(Lilya(
routes=[...],
))
Or manually:
from asyncz.schedulers import AsyncIOScheduler
scheduler = AsyncIOScheduler()
app = Lilya(
routes=[...],
on_startup=[scheduler.start],
on_shutdown=[scheduler.shutdown],
)
Use as sync contextmanager
from asyncz.schedulers import AsyncIOScheduler
with AsyncIOScheduler() as scheduler:
...
Use as async contextmanager
from asyncz.schedulers import AsyncIOScheduler
async with AsyncIOScheduler() as scheduler:
...
For using with lifespan of starlette:
from asyncz.schedulers import AsyncIOScheduler
async lifespan(app):
with AsyncIOScheduler() as scheduler:
yield
# or yield a state
app = Starlette(
lifespan=lifespan,
)
You should use store encryption for security reasons.
All standard stores except MemoryStore support the environment variable ASYNCZ_STORE_ENCRYPTION_KEY
.
If set and non-empty the hash of the value is used for AESGCM encrypting the elements before sending them
to the store.
This way store entries are encrypted and authentificated so there is no security hole.
This is highly recommended! Because if someone can inject store entries he can execute code.
Currently there are no sponsors of Asyncz but you can financially help and support the author though GitHub sponsors and become a Special one or a Legend.