tonybaloney / rich-bench Goto Github PK

Reading the documentation made me wonder why the __benchmarks__ entries are tuples with two functions a and b. Apparently this assumes that a benchmark always has to be performed against some kind of reference?

I'm wondering if the framework can be easily adapted to use cases where one doesn't have such a reference function. Say, I just to benchmark sort_three without the kind of arbitrary sort_seven reference.

Is this already supported somehow?

I am using Python 3.11 on Arch Linux and found the example work but my other function does not work. Here is my source code:

import logging

logging.basicConfig(filename='app.log', filemode='w', format='%(name)s - %(levelname)s - %(message)s')


def _log():
    for _ in range(10_000):
        logging.warning(_)

def _print():
    for _ in range(10_000):
        print(_)

__benchmarks__ = [
    (_log, _print, "Log >< Print"),
    (_print, _log, "Print >< Log")
]

Output:

• fit it to the user's screen (abbreviate benchmark titles)
• increase number of decimal places to accommodate tiny numbers (currently "**kwargs..." and "slicing with memoryview...")

I'd be happy to submit a PR for these.