zimer

A lightweight utility package that ships two decorators that make everyday Python development a bit easier:

• zimer – measure the execution time of synchronous and asynchronous functions, optionally repeating them and printing the average runtime.
• with_retry – automatically retry a function when it raises an exception, with optional exponential back-off. Works with both sync and async callables.

Installation

pip install zimer

zimer – timing decorator

from zimer import zimer
import time, asyncio

@zimer                      # one run, prints elapsed time
def slow_sync():
    time.sleep(1)

@zimer(repeats=3)           # three runs, prints average
async def slow_async():
    await asyncio.sleep(0.5)

slow_sync()
asyncio.run(slow_async())

with_retry – retry decorator

from zimer import with_retry
import random, asyncio

# --- synchronous example -----------------------------------------
@with_retry(num_retries=3, backoff=1, backoff_exponent=2)  # 1s, 4s between tries
def flaky_sync():
    if random.random() < 0.7:
        raise ValueError("Still broken …")
    return "✓ finally worked"

# --- asynchronous example ----------------------------------------
@with_retry                # default: 5 retries, no back-off
async def flaky_async():
    if random.random() < 0.5:
        raise RuntimeError("Try again …")
    return "✓ async worked"

print(flaky_sync())
print(asyncio.run(flaky_async()))

Arguments:

• num_retries – total attempts (initial call + retries). Must be ≥ 1.
• backoff – initial pause between retries (seconds). Must be ≥ 0.
• backoff_exponent – growth factor applied to the retry count to compute the wait time: sleep = backoff * ((attempt + 1) ** backoff_exponent).

Testing

Clone the repo and run:

pip install -r requirements-dev.txt  # pytest, etc.
pytest -q

All unit tests live under tests/ and cover both decorators.

License

MIT © 2024 Zubin Aysola