memoizing object speed demo

speedtest.py

class memo():
    '''adapted from https://github.com/timm/sbse14/wiki/basepy'''
    def __init__(self, **kwargs):
        self.__dict__.update(kwargs)

def test_memo():
    m = memo(stats=memo(mean=3.5, median=3),
         author=memo(name='Jim Witschey', twitter='mambocab', github='mambocab'))
    m.author.age = 25

def test_dict():
    d = { 'stats': { 'mean': 3.5, 'median': 3 },
        'author': { 'name': 'Jim Witschey',
                    'twitter': 'mambocab',
                    'github': 'mambocab' } }
    d['author']['age'] = 25

speedtest.sh

echo "class: " ; python -mtimeit -s'import speedtest' 'speedtest.test_memo()'
echo "dict:  " ; python -mtimeit -s'import speedtest' 'speedtest.test_dict()'

speedtest_output.txt

❯ ./speedtest.sh
class:
100000 loops, best of 3: 3.08 usec per loop
dict:
1000000 loops, best of 3: 0.57 usec per loop

❯ ./speedtest.sh
class:
100000 loops, best of 3: 2.98 usec per loop
dict:
1000000 loops, best of 3: 0.562 usec per loop

❯ ./speedtest.sh
class:
100000 loops, best of 3: 3.13 usec per loop
dict:
1000000 loops, best of 3: 0.606 usec per loop

❯ ./speedtest.sh
class:
100000 loops, best of 3: 3.2 usec per loop
dict:
1000000 loops, best of 3: 0.554 usec per loop

So, looks like the class-creation memoization approach is ~5x slower than dictionaries.

I should be more precise: it's slower at memo-creation time. I can't speak to access.