Skip to content

Instantly share code, notes, and snippets.

@dotchetter
Last active March 17, 2020 17:22
Show Gist options
  • Star 0 You must be signed in to star a gist
  • Fork 1 You must be signed in to fork a gist
  • Save dotchetter/4713c40f3e48189c14879a8f77e549b8 to your computer and use it in GitHub Desktop.
Save dotchetter/4713c40f3e48189c14879a8f77e549b8 to your computer and use it in GitHub Desktop.
Call functions through the PollCache object for only receiving output from the function if new output is detected.
class PollCache:
"""
This object is designed to act as a cushion between a
function or other callable, and its caller, and only
give returns when new data from the function is identified.
This way you can pass functions to an instance of this class
and loop indefinitely, where only new results will be returned.
The PollCache object will treat every function and its
constellation of arguments and keyword arguments as unique,
meaning that you can use the same function or other callable
with different parameters in a loop, and it will not be
overwritten in the PollCache just because the function is
the same, but will be treated as its own cache.
This class uses the __call__ method as its main interface.
Call the instance of this class as you would a function.
The syntax is simply:
>> cache = PollCache()
>> cache(function_name, parameter1 = val, parameter2 = val2, ...)
:silent_first_call:
Boolean. Set it to True if the desired behavior is to
build up a cache, and only start returning values if
any of the cached method returns deviate.
Default is that all initial calls with a new function
or a past function with new arguments, will return
values. Suppress this with this parameter.
---- Example with default behavior--------------------------------------
>> cache = PollCache()
>> cache(function, a = 1, b = 2) # Will produce a return value
>> cache(function, a = 1, b = 2) # Will NOT produce a return value (identical output)
>> cache(function, a = 10, b = 20) # Will produce a return value (new output)
----- Example with silent first call -----------------------------------
>> cache = PollCache(silent_first_call = True)
>> cache(function, a = 1, b = 2) # Will NOT produce a return value
>> cache(function, a = 1, b = 2) # Will NOT produce a return value (identical output)
>> cache(function, a = 10, b = 20) # Will produce a return value (new output)
"""
def __init__(self, silent_first_call = False):
self.cached_polls = dict()
self.silent_first_call = silent_first_call
def __call__(self, func: 'function', *args, **kwargs):
try:
new_result = func(*args, **kwargs)
except:
raise
if not func in self.cached_polls.keys():
self.cached_polls[func] = [
{'args': args,
'kwargs': kwargs,
'result': new_result,
'calls': 1}]
return new_result if not self.silent_first_call else None
for call in self.cached_polls[func]:
if call['args'] == args and call['kwargs'] == kwargs and call['result'] != new_result:
call['result'] = new_result
call['calls'] += 1
return new_result
if not [i for i in self.cached_polls[func] if i['args'] == args and i['kwargs'] == kwargs]:
self.cached_polls[func].append({
'args': args,
'kwargs': kwargs,
'result': new_result,
'calls': 1})
return new_result if not self.silent_first_call else None
if __name__ == '__main__':
"""
Testing the pollcache object can be done here.
This test has a simple function, print_numbers.
It'll just print out whatever you give it really.
There's an optional parameter: add_time, which will
artificially make the method return different valules
every minute. This will simulate a function that might
fetch data from the internet through an API or alike,
and the goal is to only hear from it if it returns
new data.
"""
from datetime import datetime
from pprint import pprint
from time import sleep
def print_numbers(*args, add_time = False):
if not add_time:
return args
return f'{args}, {datetime.now().minute}'
pollcache = PollCache(silent_first_call = True)
while True:
r = pollcache(print_numbers, 1,2,3, add_time = True)
if r: print(r)
sleep(1)
pprint(pollcache.cached_polls)
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment