Sometimes algorithms run better on certain inputs, leading to 'zig-zag' graphs. A simple example of this would be the in operator on a list. Take 0 in [0, 1, ..., n] as opposed to n in [0, 1, ..., n].
This can also be seen in this answer.
We should:
- Add a new timer to call
build_functions with an additional _ parameter.
- Add a decorator/class to wrap
args_conv to add:
- An unbound
functools.lru_cache so that each function has equal disadvantages/advantages to their speed.
- It should remove the argument-
_ - which will be used to allow the cache to output different random values.
- Add a new class that interfaces with
random and numpy to set a default seed so the output is reproducible. Allowing further analysis of the data.
It seems like 2 and 3 could use the same class.
Sometimes algorithms run better on certain inputs, leading to 'zig-zag' graphs. A simple example of this would be the in operator on a list. Take
0 in [0, 1, ..., n]as opposed ton in [0, 1, ..., n].This can also be seen in this answer.
We should:
build_functionswith an additional_parameter.args_convto add:functools.lru_cacheso that each function has equal disadvantages/advantages to their speed._- which will be used to allow the cache to output different random values.randomandnumpyto set a default seed so the output is reproducible. Allowing further analysis of the data.It seems like 2 and 3 could use the same class.