hunter.py

from typing import List, Optional, Tuple

from axelrod._strategy_utils import detect_cycle
from axelrod.action import Action
from axelrod.player import Player

C, D = Action.C, Action.D


class DefectorHunter(Player):
    """A player who hunts for defectors.

    Names:

    - Defector Hunter: Original name by Karol Langner
    """

    name = "Defector Hunter"
    classifier = {
        "memory_depth": float("inf"),  # Long memory
        "stochastic": False,
        "long_run_time": False,
        "inspects_source": False,
        "manipulates_source": False,
        "manipulates_state": False,
    }

    def strategy(self, opponent: Player) -> Action:
        """Actual strategy definition that determines player's action."""
        if (
            len(self.history) >= 4
            and len(opponent.history) == opponent.defections
        ):
            return D
        return C


class CooperatorHunter(Player):
    """A player who hunts for cooperators.

    Names:

    - Cooperator Hunter: Original name by Karol Langner
    """

    name = "Cooperator Hunter"
    classifier = {
        "memory_depth": float("inf"),  # Long memory
        "stochastic": False,
        "long_run_time": False,
        "inspects_source": False,
        "manipulates_source": False,
        "manipulates_state": False,
    }

    def strategy(self, opponent: Player) -> Action:
        """Actual strategy definition that determines player's action."""
        if (
            len(self.history) >= 4
            and len(opponent.history) == opponent.cooperations
        ):
            return D
        return C


def is_alternator(history: List[Action]) -> bool:
    for i in range(len(history) - 1):
        if history[i] == history[i + 1]:
            return False
    return True


class AlternatorHunter(Player):
    """A player who hunts for alternators.

    Names:

    - Alternator Hunter: Original name by Karol Langner
    """

    name = "Alternator Hunter"
    classifier = {
        "memory_depth": float("inf"),  # Long memory
        "stochastic": False,
        "long_run_time": False,
        "inspects_source": False,
        "manipulates_source": False,
        "manipulates_state": False,
    }

    def __init__(self) -> None:
        super().__init__()
        self.is_alt = False

    def strategy(self, opponent: Player) -> Action:
        """Actual strategy definition that determines player's action."""
        if len(opponent.history) < 6:
            return C
        if len(self.history) == 6:
            self.is_alt = is_alternator(opponent.history)
        if self.is_alt:
            return D
        return C


class CycleHunter(Player):
    """Hunts strategies that play cyclically, like any of the Cyclers,
    Alternator, etc.

    Names:

    - Cycle Hunter: Original name by Marc Harper
    """

    name = "Cycle Hunter"
    classifier = {
        "memory_depth": float("inf"),  # Long memory
        "stochastic": False,
        "long_run_time": False,
        "inspects_source": False,
        "manipulates_source": False,
        "manipulates_state": False,
    }

    def __init__(self) -> None:
        super().__init__()
        self.cycle = None  # type: Optional[Tuple[Action]]

    def strategy(self, opponent: Player) -> Action:
        """Actual strategy definition that determines player's action."""
        if self.cycle:
            return D
        cycle = detect_cycle(opponent.history, min_size=3)
        if cycle:
            if len(set(cycle)) > 1:
                self.cycle = cycle
                return D
        return C


class EventualCycleHunter(CycleHunter):
    """Hunts strategies that eventually play cyclically.

    Names:

    - Eventual Cycle Hunter: Original name by Marc Harper
    """

    name = "Eventual Cycle Hunter"

    def strategy(self, opponent: Player) -> None:
        """Actual strategy definition that determines player's action."""
        if len(opponent.history) < 10:
            return C
        if len(opponent.history) == opponent.cooperations:
            return C
        if len(opponent.history) % 10 == 0:
            # recheck
            self.cycle = detect_cycle(opponent.history, offset=10, min_size=3)
        if self.cycle:
            return D
        else:
            return C


class MathConstantHunter(Player):
    """A player who hunts for mathematical constant players.

    Names:

    Math Constant Hunter: Original name by Karol Langner
    """

    name = "Math Constant Hunter"
    classifier = {
        "memory_depth": float("inf"),  # Long memory
        "stochastic": False,
        "long_run_time": False,
        "inspects_source": False,
        "manipulates_source": False,
        "manipulates_state": False,
    }

    def strategy(self, opponent: Player) -> Action:
        """
        Check whether the number of cooperations in the first and second halves
        of the history are close. The variance of the uniform distribution (1/4)
        is a reasonable delta but use something lower for certainty and avoiding
        false positives. This approach will also detect a lot of random players.
        """

        n = len(self.history)
        if n >= 8 and opponent.cooperations and opponent.defections:
            start1, end1 = 0, n // 2
            start2, end2 = n // 4, 3 * n // 4
            start3, end3 = n // 2, n
            count1 = opponent.history[start1:end1].count(C) + self.history[
                start1:end1
            ].count(C)
            count2 = opponent.history[start2:end2].count(C) + self.history[
                start2:end2
            ].count(C)
            count3 = opponent.history[start3:end3].count(C) + self.history[
                start3:end3
            ].count(C)
            ratio1 = 0.5 * count1 / (end1 - start1)
            ratio2 = 0.5 * count2 / (end2 - start2)
            ratio3 = 0.5 * count3 / (end3 - start3)
            if abs(ratio1 - ratio2) < 0.2 and abs(ratio1 - ratio3) < 0.2:
                return D
        return C


class RandomHunter(Player):
    """A player who hunts for random players.

    Names:

    - Random Hunter: Original name by Karol Langner
    """

    name = "Random Hunter"
    classifier = {
        "memory_depth": float("inf"),  # Long memory
        "stochastic": False,
        "long_run_time": False,
        "inspects_source": False,
        "manipulates_source": False,
        "manipulates_state": False,
    }

    def __init__(self) -> None:
        self.countCC = 0
        self.countDD = 0
        super().__init__()

    def strategy(self, opponent: Player) -> Action:
        """
        A random player is unpredictable, which means the conditional frequency
        of cooperation after cooperation, and defection after defections, should
        be close to 50%... although how close is debatable.
        """
        # Update counts
        if len(self.history) > 1:
            if self.history[-2] == C and opponent.history[-1] == C:
                self.countCC += 1
            if self.history[-2] == D and opponent.history[-1] == D:
                self.countDD += 1

        n = len(self.history)
        if n > 10:
            probabilities = []
            if self.cooperations > 5:
                probabilities.append(self.countCC / self.cooperations)
            if self.defections > 5:
                probabilities.append(self.countDD / self.defections)
            if probabilities and all(
                [abs(p - 0.5) < 0.25 for p in probabilities]
            ):
                return D
        return C