fixed

main.py

@@ -6,6 +6,7 @@ import numpy as np
 import random
 import argparse
 
+
 ozai_url = 'http://localhost:8000/api/'
 
 
@@ -108,6 +109,8 @@ def init_game(names=["a", "b"]):
     if response.status_code == 200:
         game_id = response.json()
         print("Game ID:", game_id)
+        #delay
+        sleep(7)
         return game_id
     else:
         return None
@@ -127,7 +130,7 @@ def create_game(names=["a", "b"]):
         join_game(game_id, name)
     return game_id
 
-def submit_action(game_id, player_name, source=0, destination=0, color="start", policy="strict"):
+def submit_action(game_id, player_name, source=0, destination=0, color="start", policy="random"):
     # Submit an action to the game. The action is a dictionary with the keys being the action type and the values being the arguments.
     # Example: {"player": 0, "market": true, "factory": 0, "color": "blue", "patternLine": 0}
     if source != "market":
@@ -143,11 +146,20 @@ def submit_action(game_id, player_name, source=0, destination=0, color="start",
         'source': source,
         'destination': destination
     }
+
+    print(f"action {action}")
     response = requests.put(ozai_url + 'game/' + game_id, json=action)
-    if response.status_code == 200:
-        return True
-    else:
-        return False
+    print(f"response {response}")
+    if response.status_code != 200:
+        action = {
+            'player': str(player_name),
+            'policy': "loose",
+            'color': str(color).lower(),
+            'source': source,
+            'destination': destination
+        }
+        response = requests.put(ozai_url + 'game/' + game_id, json=action)
+    return response
 
 def get_gamestate(game_id) -> GameState:
     response = requests.get(ozai_url + 'game/' + game_id)
@@ -247,6 +259,125 @@ def strategy_2(move,GameState,player):
             return True
     return False
 
+
+def strategy_advanced_old(move, GameState, player):
+    if move[1] == 'floor':
+        # Avoid overfilling the floor
+        if sum(player.floor.values()) + move[3] > 7:
+            return False
+        return True
+
+    remaining_space = move[1] + 1 - player.pattern_lines[move[1]]['number']
+    
+    # Prioritize moves that fill the pattern line exactly, enabling a wall tile placement.
+    if move[3] == remaining_space:
+        return True
+    
+    # Deprioritize moves that will overfill a pattern line
+    if move[3] > remaining_space:
+        return False
+
+    # Ensure that placing the tile does not block you from placing other colors on the same row
+    row_color_in_wall = any(player.wall[move[1]][i] == move[2] for i in range(5))
+    if row_color_in_wall:
+        return False
+
+    # Prioritize moves that fill pattern lines near completion (1 tile left)
+    if player.pattern_lines[move[1]]['number'] == remaining_space - 1:
+        return True
+
+    return True
+
+
+def strategy_advanced(move, GameState, player):
+    # Destructure move for clarity
+    factory, pattern_line, tile_color, tile_count = move
+    
+    # Floor management: Avoid overfilling the floor line
+    if pattern_line == 'floor':
+        if sum(player.floor.values()) + tile_count > 7:
+            return False
+        return True
+
+    # Calculate the remaining space in the selected pattern line
+    remaining_space = pattern_line + 1 - player.pattern_lines[pattern_line]['number']
+    
+    # If the move exactly fills the pattern line, prioritize it
+    if tile_count == remaining_space:
+        return True
+    
+    # Deprioritize moves that overfill a pattern line (tiles that would spill over)
+    if tile_count > remaining_space:
+        return False
+
+    # Check if placing the tile blocks placing other colors in the same row
+    # This prevents placing a tile in a pattern line where the corresponding wall row already has that color
+    if any(player.wall[pattern_line][i] == tile_color for i in range(5)):
+        return False
+
+    # Prioritize filling pattern lines that are close to completion (one tile away)
+    if remaining_space == 1:
+        return True
+    
+    # If none of the conditions above trigger, accept the move
+    return True
+
+
+
+
+def strategy_lookahead(move, GameState, player):
+    def simulate_future_state(GameState, move, player):
+        """
+        Simulate the GameState after making the current move and return the hypothetical player's state.
+        """
+        future_GameState = GameState  # Assume deep copy or similar for actual implementation
+        # Apply the move
+        if move[1] == 'floor':
+            future_floor_count = sum(player.floor.values()) + move[3]
+            if future_floor_count > 7:
+                return None  # Overfilling the floor is a bad move
+        else:
+            player.pattern_lines[move[1]]['number'] += move[3]
+            if player.pattern_lines[move[1]]['number'] == move[1] + 1:
+                # Tile will be placed on the wall in the next turn
+                player.wall[move[1]][player.wall[move[1]].index(False)] = move[2]
+                player.pattern_lines[move[1]]['number'] = 0  # Reset the pattern line
+        
+        # Simulate scoring or any other immediate effect
+        # (Depending on how your game state works, update this appropriately)
+
+        return player
+
+    def evaluate_future_state(future_player):
+        """
+        Evaluate the future state for the player and return a score.
+        """
+        # Simple heuristic: prioritize a mix of empty spaces filled, tiles on the wall, and penalty on the floor
+        score = sum(sum(row) for row in future_player.wall) - sum(future_player.floor.values())
+        return score
+
+    # Evaluate the immediate impact
+    immediate_impact = strategy_advanced(move, GameState, player)
+
+    if not immediate_impact:
+        return False
+
+    # Simulate the game state after the current move
+    future_player_state = simulate_future_state(GameState, move, player)
+
+    if future_player_state is None:
+        return False  # Avoid moves that lead to a bad future state
+
+    # Evaluate the future state
+    future_score = evaluate_future_state(future_player_state)
+
+    # Set a threshold or comparison logic to decide whether the move is good enough
+    if future_score > evaluate_future_state(player):  # Basic comparison with current state
+        return True
+
+    return False
+
+
 def strategy_random(move, GameState,player):
     return True #do not filter any moves
 
@@ -259,21 +390,25 @@ def do_move(game_id, player_name, filter_strategy=strategy_random):
         for move in moves:
             if filter_strategy(move,GameState,player):
                 filtered_moves.append(move)
+        
+        if len(filtered_moves) <= 1:
+            raise Exception("No valid moves")
         # Submit a random move, of the filtered ones.
         move = random.choice(filtered_moves)
-        submit_action(game_id, player_name, move[0], move[1], move[2])
         return move
     except:
         #if filtered all moves, just submit a random move from the moves
-        result = random.choice(moves)
-        return result
+        move = random.choice(moves)
+        return move
 
 def play_game(gameid, players, strategy):
     print(f"Playing game {gameid} with players {players} and strategy {strategy.__name__}")
     while not game_over(gameid):
         for player in players:
             mov = do_move(gameid, player, filter_strategy=strategy)
-            if game_over(game_id):
+            submit_action(gameid, player, mov[0], mov[1], mov[2])
+            print(f"Player {player} did move {mov}, current score: {get_score(gameid, player)}, current round {get_gamestate(gameid).rounds}")
+            if game_over(gameid):
                 #get the score of the players
                 score = []
                 for player in players:
@@ -283,9 +418,9 @@ def play_game(gameid, players, strategy):
 
 if __name__ == "__main__":
     parser = argparse.ArgumentParser(description='Play a game with specified strategies.')
-    parser.add_argument('--game_id', type=int, default=0, help='The id of the game to play.')
+    parser.add_argument('--game_id', type=str, default="", help='The id of the game to play.')
     parser.add_argument('--players', nargs='+', default=["a", "b"], help='The names of the players.')
-    parser.add_argument('--strategy', type=str, default="", help='The strategy to use. Can be "1", "2", or "" for random.')
+    parser.add_argument('--strategy', type=str, default="", help='The strategy to use. Can be "1", "2", "advanced", "lookahead", or "" for random.')
     parser.add_argument('--ozai_url', type=str, default='http://localhost:8000/api/', help='The url to the ozai server.')
     args = parser.parse_args()
     
@@ -297,11 +432,19 @@ if __name__ == "__main__":
     #use global ozai url and update it
     ozai_url = args.ozai_url
 
-    if game_id == 0:
+    if game_id == "":
         game_id = create_game(names=players)
+    
     if strategy == "1":
         play_game(game_id, players, strategy_1)
     elif strategy == "2":
         play_game(game_id, players, strategy_2)
+    elif strategy == "advanced":
+        play_game(game_id, players, strategy_advanced)
+    elif strategy == "lookahead":
+        play_game(game_id, players, strategy_lookahead)
+    elif strategy == "ai":
+        import ml
+        play_game(game_id, players, ml.strategy_ai)
     else:
         play_game(game_id, players, strategy_random)