deep learning and go
如下所示。如果我执行python.exe bot_v_bot.py,程序将运行。
如果我从eclipse/pydev运行bot_v_bot.py,那么它就可以工作了。
.ipnb文件与bot_v_bot.py位于同一文件夹中。
如果我说:
from bot_v_bot import main
main()
输入.ipnb文件中的一个单元格并运行它,它显示:
ModuleNotFoundError Traceback (most recent call last)
<ipython-input-4-248b35949c67> in <module>()
----> 1 from bot_v_bot import main
2 main()
ModuleNotFoundError: No module named 'bot_v_bot'
编辑:下面的代码有效。eclipse在python路径上有src。
import sys
sys.path.append('src')
from bot_v_bot import main
main()
from __future__ import print_function
# tag::bot_vs_bot[]
from dlgo import agent
from dlgo import goboard_slow
from dlgo import gotypes
from dlgo.utils import print_board, print_move
import time
def main():
board_size = 9
game = goboard_slow.GameState.new_game(board_size)
bots = {
gotypes.Player.black: agent.naive.RandomBot(),
gotypes.Player.white: agent.naive.RandomBot(),
}
while not game.is_over():
time.sleep(0.3) # <1>
print(chr(27) + "[2J") # <2>
print_board(game.board)
bot_move = bots[game.next_player].select_move(game)
print_move(game.next_player, bot_move)
game = game.apply_move(bot_move)
if __name__ == '__main__':
main()
# <1> We set a sleep timer to 0.3 seconds so that bot moves aren't printed too fast to observe
# <2> Before each move we clear the screen. This way the board is always printed to the same position on the command line.
# end::bot_vs_bot[]