$ pip install rlim
Create and use a RateLimiter
instance:
@RateLimiter(Rate(2), Limit(50, 40))
def f():
...
@RateLimiter(Rate(2), Limit(50, 40))
async def f():
...
Apply a RateLimiter
instance to a function decorated with placeholder
:
@placeholder
def f():
...
rl_set(f, RateLimiter(Rate(2), Limit(50, 40)))
@placeholder
async def f():
...
rl_set(f, RateLimiter(Rate(2), Limit(50, 40)))
Use an instance as a context manager:
rl = RateLimiter(Rate(2), Limit(50, 40))
def f():
with rl:
...
async def f():
async with rl:
...
Notice that in the above, Rate
and Limit
are two distinct types. Rate
is used to define a constant calling speed - for example, Rate(2)
would equate to 1
call every 0.5
seconds. Limit
is used to define a maximum number of calls (at any speed) within a certain period of time (sliding window) - for example, Limit(50, 40)
would mean the user could make calls at any speed, so long as they don't surpass 50 calls within the last 40 seconds. Together, this means the user can many calls at a max speed of 0.5s/call
, and must stay below (or equal to) 50
calls in the past 40
seconds.
Bundles allow you to bundle together numerous rate limiters, with methods for applying them to the methods of a given class or class instance. When a bundle is applied to a class instance, the RateLimiter
instances (or copies of them, if desired) within the bundle will be applied to each of the class's methods upon class instantiation.
Creating a Bundle
:
bdl = Bundle(
fn1=RateLimiter(...),
fn2=RateLimiter(...),
...
)
Applying a Bundle
instance to a class:
@bdl
class Example:
def fn1(self) -> None:
return
def fn2(self) -> None:
return
Now, when you create an instance of Example
, fn1
and fn2
will have their corresponding RateLimiter
instances applied to them.
RateLimiter.copy(**overrides) -> RateLimiter
Create a copy of the
RateLimiter
instance with optional overrides (that will be passed intoRateLimiter.__init__
).
RateLimiter.apply(func: Callable[_P, _R_co]) -> Callable[_P, _R_co]
RateLimiter.apply(func: Callable[_P, Awaitable[_R_co]]) -> Callable[_P, Awaitable[_R_co]]
Manually wrap the given function to use the
RateLimiter
instance for rate limiting.RateLimiter.__call__
(the function that makes it possible to decorate another function with aRateLimiter
instance) is simply an alias ofRateLimiter.apply
.
Bundle.apply(
inst: object,
ignore: bool = MISSING,
copy: bool = MISSING,
**overrides
) -> None
Apply the
RateLimiter
instances in thisBundle
to the given class instance.
ignore
(defaultFalse
) will make it soRateLimiterError
will not be raised if a function in the decorated class does not have a correspondingRateLimiter
.copy
(defaultTrue
) will make it so copies of theRateLimiter
instances are applied, instead of the same instances.**overrides
are keyword overrides that will be passed intoRateLimiter.copy
(only ifcopy
isTrue
).
Bundle.decorate(
ignore: bool = MISSING,
copy: bool = MISSING,
**overrides
) -> Callable[[Type[T]], Callable[_P, T]]
This function allows the user to have more control over how the
RateLimiter
instances are applied to the decorated class's methods. It returns a decorator.ignore
,copy
, and**overrides
will be passed intoBundle.apply
.
Bundle.bake(
ignore: bool = MISSING,
copy: bool = MISSING,
**overrides
) -> None
Bake arguments for calls to
Bundle.apply
andBundle.decorate
into theBundle
instance. Any arguments provided toBundle.apply
orBundle.decorate
have precedence over baked arguments.