backtest.optimize throws error

[Kayne88]

Expected Behavior

backtest.optimize(fast_period=range(2,10,2),medium_period=range(6,20,2),slow_period=range(10,30,2))
is executing without error.

Actual Behavior

It returns this traceback:

Traceback (most recent call last):
File "c:/Users/User/Documents/Trading/Back testing/run_backtest.py", line 19, in
optimal_parameters = backtest.optimize(fast_period=range(2,10,2),medium_period=range(6,20,2),slow_period=range(10,30,2))
File "C:\Users\User\Anaconda3\lib\site-packages\backtesting\backtesting.py", line 1487, in optimize
output = _optimize_grid()
File "C:\Users\User\Anaconda3\lib\site-packages\backtesting\backtesting.py", line 1324, in _optimize_grid
rand = np.random.RandomState(random_state).random
AttributeError: 'mtrand.RandomState' object has no attribute 'random'

Analysis

If I look at numpy RandomState doc: https://numpy.org/doc/1.16/reference/generated/numpy.random.RandomState.html
There is indeed no method handle called random, only rand.

Additional info

• Backtesting version: 0.3.2

[kernc]

Lovely, thanks. Seems like np.random.default_rng().random random.random() is the way to go. 👍

[kernc]

Apparently this only happens on some platforms. For me:

>>> np.__version__
'1.21.2'
>>> np.random.RandomState().random()
0.9464831509552952

But the method is not documented. Therefore, thanks!