python-depends

A FastAPI like dependecy injector

Install

# stable
pip3 install python-depends

# latest
pip3 install git+https://github.com/Dimfred/depends.git

Examples

from depends import Depends, inject

async def d1():
    # do some stuff, which takes long
    return "some stuff"

async def d2():
    # do some other stuff, which also takes long
    return "some other stuff"

# inject the dependency into a function
@inject
async def main(d1_=Depends(d1), d2_=Depends(d2)):
    print(d1_)  # some stuff
    print(d2_)  # some other stuff

Nested dependencies

from depends import Depends, inject

async def d1():
    # do some stuff, which takes long
    return "some stuff"

async def d2(d1_=Depends(d1)):
    # do some other stuff, which also takes long
    # you can work with d2_ here
    return "some other stuff"

# d1 was called only once and is cached during the whole call
@inject
async def main(d1_=Depends(d1), d2_=Depends(d2)):
    print(d1_)  # some stuff
    print(d2_)  # some other stuff

You can also use parameters in your injected function which will be forwarded to your dependencies. The detection is done by name, no type checking is applied here.

from depends import Depends, inject

async def d1(a):
    return a


# d1 was called only once and is cached during the whole call
@inject
async def main(a, d1_=Depends(d1)):
    return a, d1_

assert (await main(1)) == (1, 1)

Another cool thing is that you can use context managed objects inside an injected function. Like for example a database session.

from depends import Depends, inject

async def get_db():
    async with Session() as db:
        yield db

@inject
async def main(db=Depends(get_db)):
    # do stuff with your async db connection
    # after the exit the connection will be teared down

TODO

• support sync dependencies (only async rn)
• replace the caching mechanism with maybe the correct dependency tree