I did some modification as the original answer have some issues that doesnt resolve according to the use of utils.py after that it pass all the public testcase but fail on the private and state that Make sure that 'evaluate' is called with the correct input, especially for white. But i think i have manage to do that but still facing this issues, why?

def minimax$($

board: Board,

depth: int,

max$\_$depth: int,

is$\_$black: bool

$)->$ tuple$[$Score$,$ Move$]$:

$"""$

Finds the best move for the input board state.

Note that you are black.

Parameters

$----------$

board: $2$D list of lists. Contains characters $"$B$","$W$",$ and $"\_",$

representing black pawn, white pawn, and empty cell, respectively.

depth: int, the depth to search for the best move. When this is equal

to $`$max$\_$depth$`,$ you should get the evaluation of the position using

the provided heuristic function.

max$\_$depth: int, the maximum depth for cutoff.

is$\_$black: bool. True when finding the best move for black, False

otherwise.

Returns

$-------$

A tuple $($evalutation$,(($src$\_$row, src$\_$col$),($dst$\_$row, dst$\_$col$)))$:

evaluation: the best score that black can achieve after this move.

src$\_$row, src$\_$col: position of the pawn to move.

dst$\_$row, dst$\_$col: position to move the pawn to$.$

$"""$

if depth $==$ max$\_$depth or utils.is$\_$game$\_$over$($board$)$:

return evaluate$($board$),$ None

# Determine the best move and its evaluation

if is$\_$black:

best$\_$evaluation $=$ float$(\text{'}-$inf'$)$

best$\_$move $=$ None

for action in generate$\_$valid$\_$moves$($board$)$:

new$\_$board $=$ utils.state$\_$change$($board$,$ action$[0],$ action$[1],$ in$\_$place$=$False$)$

opponent$\_$evaluation, $\_=$ minimax$($new$\_$board, depth $+1,$ max$\_$depth, False$)$

if opponent$\_$evaluation $>$ best$\_$evaluation:

best$\_$evaluation $=$ opponent$\_$evaluation

best$\_$move $=($action$[0],$ action$[1])$

return best$\_$evaluation, best$\_$move

else:

best$\_$evaluation $=$ float$(\text{'}$inf$\text{'})$

best$\_$move $=$ None

for action in generate$\_$valid$\_$moves$($utils$.$invert$\_$board$($board$,$ in$\_$place$=$False$))$:

new$\_$board $=$ utils.state$\_$change$($utils$.$invert$\_$board$($board$,$ in$\_$place$=$False$),$ action$[0],$ action$[1],$ in$\_$place$=$False$)$

opponent$\_$evaluation, $\_=$ minimax$($utils$.$invert$\_$board$($new$\_$board, in$\_$place$=$False$),$ depth $+1,$ max$\_$depth, True$)$

if opponent$\_$evaluation $<$ best$\_$evaluation:

best$\_$evaluation $=$ opponent$\_$evaluation

best$\_$move $=($action$[0],$ action$[1])$

return best$\_$evaluation, best$\_$move