Welcome to Software Development on Codidact!
Will you help us build our independent community of developers helping developers? We're small and trying to grow. We welcome questions about all aspects of software development, from design to code to QA and more. Got questions? Got answers? Got code you'd like someone to review? Please join us.
How to add a percent_jitter strategy to backoff Python package and pass extra args in the decorator?
I'm trying to add a percent-based jitter function to the backoff package and use that modified package in my project instead of the PyPI release.
Current jitter algorithm I want to implement (from Stack Overflow):
https://stackoverflow.com/q/46939285
Proposed function:
def percent_jitter(value: float, percentage: float = 0.2) -> float:
max_jitter = value * percentage
return value + random.uniform(-max_jitter, max_jitter)
Questions:
- What is the best way to modify the backoff package source to add this percent_jitter function? The existing jitter functions are called with only the backoff value. I would need to modify the library so a jitter function can accept an extra parameter (like percentage) and let users pass that parameter when applying the @backoff.on_exception with jitter=percent_jitter(percentage={input}).
- How can I install and use my modified version in my project instead of the PyPI package?
- Alternatively, if modifying the library is too complex, how can I use this custom jitter function with the original backoff package?
1 answer
One option is to write a (parameterised) decorator that transforms your jitter function with a given percentage value:
from typing import Callable
def percent_jitter_dec(percentage: float = 0.2) -> Callable[[float], float]:
def _jitter(value: float) -> float:
max_jitter = value * percentage
return value + random.uniform(-max_jitter, max_jitter)
return _jitter
Now, you should be able to use jitter=percent_jitter_dec(percentage=0.5).
In this simple case, you actually do not need to write the wrapper yourself. The functools package (normally included with any standard Python installation) provides the partial function that implements exactly this kind of transformations. To use it, you would do something as follows:
from functools import partial
@backoff.on_exception(jitter=partial(percent_jitter, percentage=0.5))
def func(): ...
Note that we are using the plain percent_jitter function you provided in your question, here.
0 comment threads
1 comment thread