Create UrHeuristic agent

agents/ur_heuristic.py

@@ -0,0 +1,121 @@
+"""
+    This module contains a heuristic in the Royal Game of Ur.
+"""
+
+from __future__ import annotations
+
+import random
+
+from .agent import Agent, Payload
+
+BAD_OMEN = {
+    -1 : 4 / 16,
+    -2 : 6 / 16,
+    -3 : 4 / 16,
+    -4 : 1 / 16,
+}
+
+class UrHeuristic(Agent):
+    """
+        Describe here what your agent does and how it behaves! This will help
+        others understand how your agent works.
+    """
+
+    def __init__(self) -> None:
+        """
+            Create a custom instance of your agent. This function allows you
+            to define which games your agent can play.
+
+            You may add parameters to the function if your agent requires more
+            information to be able to operate.
+        """
+
+        super().__init__(
+            name="Ur heuristic agent",
+            author="Bram",
+            version="1.0.0",
+            profile={
+                "me.noordstar.peanuts.is_ai": False,
+            },
+        )
+
+        # Indicate that you're willing to play tic-tac-toe
+        # Remove this if you don't want your bot to participate there.
+        self.add_royal_game_of_ur(on_move=self.play_ur, profile={})
+    
+    def enemy_at(self, payload : Payload, i : int) -> bool:
+        """
+            Determine whether an enemy is at a given position.
+        """
+        return payload.get(f"{i}_enemy", payload.get(str(i), "")) == "ENEMY"
+    
+    def play_ur(self, payload : Payload) -> Payload:
+        """
+            Use a heuristic. The score of each state is the total number of
+            spaces that your pieces have moved forward minus the total number
+            of spaces that your opponent's pieces have moved forward.
+        """
+        moves = [ int(n) for n in payload.get("valid_moves", [0]) ]
+        star_fields = [ int(n) for n in payload.get("star_fields", []) ]
+        safe_fields = [ int(n) for n in payload.get("safe_fields", []) ]
+        roll = int(payload.get("roll", 0))
+
+        scores : dict[int, float] = { i : 0 for i in moves }
+
+        for m in moves:
+            dest = m + roll
+
+            # The total sum of steps taken will be the same regardless of which
+            # piece we move. Hence we ignore the roll itself.
+            #
+            # There's a few other aspects we need to keep in mind, however.
+
+            # If we can hit a star, we'll be able to roll again.
+            # The expected value is that we'll be able to take 2 extra steps.
+            if dest in star_fields:
+                scores[m] += 2
+            
+            # If we can capture an enemy's piece, they'll be put back to start.
+            # That is a massive gain of points.
+            if self.enemy_at(payload=payload, i=dest) and dest not in safe_fields and dest not in star_fields:
+                scores[m] += dest
+
+            # Every spot has a certain danger to it.
+            # Improve the score based on the danger of the spot we're leaving.
+            scores[m] += self.risk_score(payload=payload, i=m)
+
+            # Similarly, punish the score based on the danger we're getting
+            # ourselves into.
+            scores[m] -= self.risk_score(payload=payload, i=dest)
+
+            # Near the end, you can get stuck if it takes too long to finish
+            # Therefore, slightly before the finish, you're rewarded with
+            # finishing precisely.
+            off = m - 15
+            if off in BAD_OMEN:
+                scores[m] = 1 / BAD_OMEN[off]
+        
+        highscore = max(scores.values())
+        keys = [ pos for pos, score in scores.items() if score == highscore ]
+
+        return { "move": random.choice(keys) }
+
+    def risk_score(self, payload : Payload, i : int) -> float:
+        """
+            What is the risk score of a given field?
+
+            The risk score is the expected number of steps you'll need to move
+            back as a result of your opponent's next turn.
+        """
+
+        if i in list(payload.get("safe_fields", [])) or i in list(payload.get("star_fields", [])):
+            return 0.0
+
+        score : float = 0.0
+
+        for offset, risk in BAD_OMEN.items():
+            if self.enemy_at(payload=payload, i=i+offset):
+                score += risk
+
+        return i * score
+