self.request.close()
+class Thing:
+
+ def __init__(self, position):
+ self.position = position
+ self.task = 'wait'
+ self.todo = 0
+
+ def task_wait(self):
+ pass
+
+ def task_moveup(self):
+ self.position[0] -= 1
+
+ def task_movedown(self):
+ self.position[0] += 1
+
+ def decide_task(self):
+ self.set_task('wait')
+
+ def set_task(self, task):
+ self.task = task
+ self.todo = 1
+
+ def proceed(self, is_AI=True):
+ """Further the thing in its tasks.
+
+ Decrements self.todo; if it thus falls to <= 0, enacts the method whose
+ name is 'task_' + self.task and sets self.task = None. If is_AI, calls
+ self.decide_task to decide a new self.task.
+ """
+ self.todo -= 1
+ if self.todo <= 0:
+ task= getattr(self, 'task_' + self.task)
+ task()
+ self.task = None
+ if is_AI and self.task is None:
+ self.decide_task()
+
+
class World:
- turn = 0
- map_size = (5, 5)
- map_ = 'xxxxx\n'+\
- 'x...x\n'+\
- 'x.X.x\n'+\
- 'x...x\n'+\
- 'xxxxx'
- player_pos = [3, 3]
+
+ def __init__(self):
+ self.turn = 0
+ self.map_size = (5, 5)
+ self.map_ = 'xxxxx\n'+\
+ 'x...x\n'+\
+ 'x.X.x\n'+\
+ 'x...x\n'+\
+ 'xxxxx'
+ self.things = [Thing(position=[3, 3]), Thing([1, 1])]
+ self.player_i = 0
+ self.player = self.things[self.player_i]
def fib(n):
def __init__(self, queues_out):
from multiprocessing import Pool
self.queues_out = queues_out
- self.pool = Pool()
self.world = World()
+ # self.pool and self.pool_result are currently only needed by the FIB
+ # command and the demo of a parallelized game loop in cmd_inc_p.
+ self.pool = Pool()
self.pool_result = None
def send_to(self, connection_id, msg):
"""Transform tuple (y,x) into string 'Y:'+str(y)+',X:'+str(x)."""
return 'Y:' + str(tuple_[0]) + ',X:' + str(tuple_[1])
+ def proceed_to_next_player_turn(self, connection_id):
+ """Run game world turns until player can decide their next step.
+
+ Sends a 'TURN_FINISHED' message, then iterates through all non-player
+ things, on each step furthering them in their tasks (and letting them
+ decide new ones if they finish). The iteration order is: first all
+ things that come after the player in the world things list, then (after
+ incrementing the world turn) all that come before the player; then the
+ player's .proceed() is run, and if it does not finish his task, the
+ loop starts at the beginning. Once the player's task is finished, the
+ loop breaks, and client-relevant game data is sent.
+ """
+ self.send_all('TURN_FINISHED ' + str(self.world.turn))
+ while True:
+ for thing in self.world.things[self.world.player_i+1:]:
+ thing.proceed()
+ self.world.turn += 1
+ for thing in self.world.things[:self.world.player_i]:
+ thing.proceed()
+ self.world.player.proceed(is_AI=False)
+ if self.world.player.task is None:
+ break
+ self.send_all('NEW_TURN ' + str(self.world.turn))
+ self.send_all('MAP_SIZE ' + self.stringify_yx(self.world.map_size))
+ self.send_all('TERRAIN\n' + self.world.map_)
+ self.send_all('POSITION ' + self.stringify_yx(self.world.player.position))
+
def cmd_fib(self, tokens, connection_id):
"""Reply with n-th Fibonacci numbers, n taken from tokens[1:].
reply = ' '.join([str(r) for r in results])
self.send_to(connection_id, reply)
- def cmd_inc(self, connection_id):
+ def cmd_inc_p(self, connection_id):
"""Increment world.turn, send game turn data to everyone.
To simulate game processing waiting times, a one second delay between
TURN_FINISHED and NEW_TURN occurs; after NEW_TURN, some expensive
calculations are started as pool processes that need to be finished
until a further INC finishes the turn.
+
+ This is just a demo structure for how the game loop could work when
+ parallelized. One might imagine a two-step game turn, with a non-action
+ step determining actor tasks (the AI determinations would take the
+ place of the fib calculations here), and an action step wherein these
+ tasks are performed (where now sleep(1) is).
"""
from time import sleep
if self.pool_result is not None:
self.send_all('NEW_TURN ' + str(self.world.turn))
self.send_all('MAP_SIZE ' + self.stringify_yx(self.world.map_size))
self.send_all('TERRAIN\n' + self.world.map_)
- self.send_all('POSITION ' + self.stringify_yx(self.world.player_pos))
+ self.send_all('POSITION ' + self.stringify_yx(self.world.player.position))
self.pool_result = self.pool.map_async(fib, (35, 35))
def cmd_get_turn(self, connection_id):
"""Send world.turn to caller."""
self.send_to(connection_id, str(self.world.turn))
- def cmd_move(self, direction):
- """Move player 'UP' or 'DOWN' depending on direction string."""
+ def cmd_move(self, direction, connection_id):
+ """Set player task to 'moveup' or 'movedown', finish player turn."""
if not direction in {'UP', 'DOWN'}:
raise ArgumentError('MOVE ARGUMENT MUST BE "UP" or "DOWN"')
if direction == 'UP':
- self.world.player_pos[0] -= 1
+ self.world.player.set_task('moveup')
else:
- self.world.player_pos[0] += 1
- self.send_all('POSITION ' + self.stringify_yx(self.world.player_pos))
+ self.world.player.set_task('movedown')
+ self.proceed_to_next_player_turn(connection_id)
+
+ def cmd_wait(self, connection_id):
+ """Set player task to 'wait', finish player turn."""
+ self.world.player.set_task('wait')
+ self.proceed_to_next_player_turn(connection_id)
def cmd_echo(self, tokens, input_, connection_id):
"""Send message in input_ beyond tokens[0] to caller."""
try:
if len(tokens) == 0:
self.send_to(connection_id, 'EMPTY COMMAND')
- elif len(tokens) == 1 and tokens[0] == 'INC':
- self.cmd_inc(connection_id)
+ elif len(tokens) == 1 and tokens[0] == 'INC_P':
+ self.cmd_inc_p(connection_id)
elif len(tokens) == 1 and tokens[0] == 'GET_TURN':
self.cmd_get_turn(connection_id)
+ elif len(tokens) == 1 and tokens[0] == 'WAIT':
+ self.cmd_wait(connection_id)
elif len(tokens) == 2 and tokens[0] == 'MOVE':
- self.cmd_move(tokens[1])
+ self.cmd_move(tokens[1], connection_id)
elif len(tokens) >= 1 and tokens[0] == 'ECHO':
self.cmd_echo(tokens, input_, connection_id)
elif len(tokens) >= 1 and tokens[0] == 'ALL':
projects / plomrogue2-experiments / commitdiff