def _step(self, solutions, depth):
if depth == 0:
return solutions
else:
next_solutions = []
for score, moves, board, row, column in solutions:
for delta_r, delta_c in self._swaps(board, row, column):
swapped_board = Board.copy_board(board).swap(row, column, row + delta_r, column + delta_c)
if self._remember(swapped_board):
# only add move to solutions if it is a board layout that has not been seen before
next_solutions.append((score + self._score(swapped_board), (moves + ((delta_r, delta_c),)), swapped_board, row + delta_r, column + delta_c))
if next_solutions:
# prune solutions down before recursing to next depth
return self._step(self._prune(next_solutions), depth - 1)
else:
return solutions
def _step(self, score, moves, board, row, column, depth):
if depth == 0:
return (score, moves, board)
else:
# get possible swaps from current row, column position
swaps = self._swaps(board, row, column)
solutions = []
for delta_r, delta_c in swaps:
swapped_board = Board.copy_board(board).swap(
row, column, row + delta_r, column + delta_c)
if self._remember(swapped_board):
# only add move to solutions if it is a board layout that has not been seen before
solutions.append((self._score(swapped_board),
swapped_board, (delta_r, delta_c)))
if solutions:
# calculate score on all possible swaps and order them from highest to lowest. recurse into the highest one
best = max(solutions, key=lambda x: x[0])
return self._step(score + best[0], (moves + (best[2], )),
best[1], row + best[2][0],
column + best[2][1], depth - 1)
else:
return (score, moves, board)