fix trasure_on_right choose_action bug

[chucklqsun]

.any() is checking whether any value is true,
.all() is checking whether all values are true.
state_actions.all() == 0 will be true if s[n]'s values are (0, 1) and select action randomly. It is a bug.

  A  B
0  0  0
1  0  1
2  1  2
==below is any()==
('0:', False)
('1:', True)
('2:', True)
==below is all()==
('0:', False)
('1:', False)
('2:', True)

[MorvanZhou]

Thanks mate, this is indeed a bug. But to be more intuitive, I think we can change not state_actions.any() to (state_actions == 0).all(). I will approve it after you applying this change.

[chucklqsun]

@MorvanZhou Good solution. I have modified and committed the revision.

[MorvanZhou]

Closed with thanks