pyfuncbuffer

A simple to use decorator to buffer function calls. Supports python versions 3.7 and up. Works for both regular and async functions.

Install

$ pip install pyfuncbuffer

Example usage

Let's say you have a scraper, and don't want sites to timeout you. You can use the @buffer() wrapper to make your function calls buffered!

from pyfuncbuffer import buffer

# We specify scrape_links to always buffer at least 0.5 seconds
# and by a random delay of 0 to 0.5
@buffer(seconds=0.5, random_delay=0.5)
def scrape_links(url) -> []: ...

links = scrape_links("https://example.org")

while True:
    link = links.pop(0)
    links.append(scrape_links(link))

The @buffer() wrapper works both for regular functions, and instance methods!

Parameters

• seconds: Required

Seconds to buffer. Can be an int or a float.

• random_delay: Optional

Seconds to define a random delay between 0 and random_delay. Can be an int or a float. Alternativelly if a tuple is passed, delay is chosen between random_delay[0] and random_delay[1].

• always_buffer: Optional

Whether or not to always buffer. If specified, buffer_on_same_arguments is ignored.

• buffer_on_same_arguments: Optional

Only buffer if the arguments on the buffered function are the same. False by default.

• share_buffer: Optional

Share buffer between processes. This is only useful when using multiprocessing, and still wanting to have function calls buffered even if called in seperate processes.

Testing

Testing is done using pytest and pytest-asyncio.

Run

$ python -m pytest tests/test_pyfuncbuffer.py

in the project root to run all the tests.