ex.
- lst = [A,B,C], i=0 => ans = [A,B,C]
- lst = [A,B,C], i=1 => ans = [B,C,A]
- lst = [A,B,C], i=2 => ans = [C,A,B]
sorted(lst, key=lambda e: i-lst.index(e))みたいにして，iに近い程前に来るみたいにしたかったけど，
iより前にある値の評価が負になって上手く並び替えられなかった．
=> [lst[(i+j)%len(lst)] for j in range(len(lst))] で実装することにした．
以下実装をきれいにできないものか...
p1.action_histories = [A,B]
p2.action_histories = [C]
p3.action_histories = [E,F]
playerの順番 : p3 => p1 => p2
求めるもの : all_action_histories = [E,A,C,F,B]
def __order_histories(self, start_pos, players):
ordered_players = [players[(start_pos+i)%len(players)] for i in range(len(players))]
all_player_histories = [p.action_histories for p in ordered_players]
max_len = max([len(h) for h in all_player_histories])
unified_histories = [self.__unify_length(max_len, l) for l in all_player_histories]
ordered_histories = reduce(lambda acc, zp: acc + list(zp), zip(*unified_histories), [])
return [history for history in ordered_histories if not history is None]
def __unify_length(self, max_len, lst):
for _ in range(max_len-len(lst)):
lst.append(None)
return lst