A programmatic approach to function runtime estimation.
This package first came about as a way to serve my needs during an Algorithms class in college.
It started as a function that stored time elapsed results in a global dictionary, where the key was a tuple of the functions name and return value.
Now it's just a bit different.
-
PYPI:
pip install functimer
-
Manual:
poetry install --no-dev poetry build pip install dist/*.whl
How to install Poetry.
-
Python
@timed(unit=Unit.SECOND, number=1) def timed_sleep(seconds): sleep(seconds) runtime = timed_sleep(1) "1.00 s"
-
Command Line
$ python -m functimer "sum([1, 2, 3])" Average runtime of 10,000 executions: 0.15 µs $ python -m functimer "sum([1, 2, 3])" --return Average runtime of 10,000 executions: 0.15 µs sum([1, 2, 3]) -> 6 $ python -m functimer "(lambda x: x+x)(10)" --return Average runtime of 10,000 executions: 0.14 µs (lambda x: x+x)(10) -> 20 $ python -m functimer "functimer.util.get_unit('1.00 s')" --return Average runtime of 10,000 executions: 0.50 µs functimer.util.get_unit('1.00 s') -> Unit.SECOND $ python -m functimer "functimer.classes.Unit.from_str('s')" --return Average runtime of 10,000 executions: 0.25 µs functimer.classes.Unit.from_str('s') -> Unit.SECOND
Run tox in the root directory of the repo.
MIT