STEP policy synthesis fails for n_max = 200, alpha = 0.01

[HarukiNishimura-TRI]

The following code fails to generate a step policy:

from sequentialized_barnard_tests import MirroredStepTest, Hypothesis

MirroredStepTest(Hypothesis.P0LessThanP1, n_max=200, alpha=0.01)

The error message is as follows:

STEP Policy Synthesis:  59%|█████▉    | 118/200 [00:05<00:03, 20.97it/s]
Traceback (most recent call last):
  File "~/sequentialized_barnard_tests/sequentialized_barnard_tests/scripts/synthesize_general_step_policy.py", line 547, in <module>
    ) = run_step_policy_synthesis(
  File "~/sequentialized_barnard_tests/sequentialized_barnard_tests/scripts/synthesize_general_step_policy.py", line 343, in run_step_policy_synthesis
    key_weights_rectified[second_feature_counter] = key_weights[
TypeError: 'NoneType' object is not subscriptable

[HarukiNishimura-TRI]

@dasnyder5 Perturbing alpha slightly from 0.01 to 0.010001 did indeed resolve the bug.

python sequentialized_barnard_tests/scripts/synthesize_general_step_policy.py --n_max 200 --alpha 0.010001 --n_points 129 --lambda_value 2.1 --major_axis_length 1.4 --log_p_norm 0.0 --use_p_norm False
STEP Policy Synthesis: 100%|████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 200/200 [00:19<00:00, 10.13it/s]