File size: 20,290 Bytes
fc5bba6
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
7812de0
fc5bba6
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
# coding: utf-8

# Author: Du Mingzhe (mingzhe@nus.edu.sg)
# Date: 2025-09-09

import os
import uuid
import random
import logging
import numpy as np
import matplotlib.pyplot as plt

from enum import Enum
from pathlib import Path
from openai import OpenAI
from scipy.stats import norm
from pydantic import BaseModel
from dotenv import load_dotenv
from abc import ABC, abstractmethod
from typing import List, Tuple, Dict
from trueskill import Rating, quality, rate
from openai_cost_calculator import estimate_cost_typed


# Logging Configuration
logging.getLogger("httpx").setLevel(logging.WARNING)
logger = logging.getLogger()
logger.setLevel(logging.INFO)
formatter = logging.Formatter('%(asctime)s %(levelname)s:%(name)s:%(message)s')
file_handler = logging.FileHandler('dixit.log')
file_handler.setFormatter(formatter)
stream_handler = logging.StreamHandler()
stream_handler.setFormatter(formatter)
if not logger.handlers:
    logger.addHandler(file_handler)
    logger.addHandler(stream_handler)

load_dotenv()

Role = Enum('Role', ['STORYTELLER', 'GUESSER', 'UNKNOWN'])

class StorytellResponse(BaseModel):
    clue: str
    explaination: str

class GuessResponse(BaseModel):
    card_id: int
    explaination: str

class VoteResponse(BaseModel):
    card_id: int
    explaination: str

StorytellPrompt = """
You are a master storyteller in the board game Dixit, known for your clever and evocative clues. 

Your challenge is to create a hard and ambiguous clue for the card below that strikes a perfect balance, as your score depends on it.

You can get 3 points if someone but not all of them guesses your card. Otherwise, you get 0 points.

# The Goal:
At least one player must guess your card, but not all of them. If everyone or no one guesses correctly, you score zero points.

# Your Task:
- Provide a clue for the card. This can be a word, a phrase, a quote, or a concise fragment of a story.
- Briefly explain the strategy behind your clue: Why do you think it's not too obvious, yet not too obscure?

# Response Format
class StorytellResponse(BaseModel):
    clue: str
    explaination: str
"""

GuessPrompt = """
You are a master player in the creative game of Dixit. The storyteller has provided an ambiguous clue. 

Your objective is to analyze this clue and select one card from your hand to serve as a decoy. 

Your goal is to mislead other players into voting for your card. You will earn 1 point for each vote your card receives.

Storyteller's Clue: {story}

# Response Format
class GuessResponse(BaseModel):
    card_id: int # the index of the image in the list of provided images (starts from 0)
    explaination: str
"""

VotePrompt = """
You are a master player in the creative game of Dixit. You are in the voting phase of the game.

The storyteller has provided an ambiguous clue, and a set of cards are on the table—one is the storyteller's, and the rest are decoys from other guessers. 

Your goal is to analyze the clue and the art on each card to deduce which one belongs to the storyteller.

Storyteller's Clue: {story}

# Response Format
class VoteResponse(BaseModel):
    card_id: int # the index of the image in the list of provided images (starts from 0)
    explaination: str
"""

class Card:
    def __init__(self, card_id: int, image_path: str):
        self.card_id = card_id
        self.guessed_by = None
        self.voted_by = list()
        self.image_path = image_path

class Player(ABC):
    def __init__(self, player_id: str):
        self.player_id = player_id
        self.role = Role.UNKNOWN
        self.selected_card_id = None
        self.guessed_card_id = None
        self.voted_card_id = None
        self.cost = 0
        logger.debug(f"[Player] Player {self.player_id} initialized with role {self.role}")
    
    @abstractmethod
    def storytell_inference(self, card: Card) -> Tuple[str, str, float]:
        raise NotImplementedError
    
    @abstractmethod
    def guess_inference(self, story: str, cards: List[Card]) -> Tuple[Card, str, float]:
        raise NotImplementedError
    
    @abstractmethod
    def vote_inference(self, story: str, cards: List[Card]) -> Tuple[Card, str, float]:
        raise NotImplementedError
    
    def storytell(self, card: Card) -> str:
        assert self.role == Role.STORYTELLER
        story, explaination, cost = self.storytell_inference(card)
        self.cost += cost
        self.selected_card_id = card.card_id
        logger.info(f"[Storytelling] [{self.player_id}] [{card.card_id}] -> '{story}' (cost: {cost:.6f} USD)")
        return story
    
    def guess(self, story: str, cards: List[Card]) -> Card:
        assert self.role == Role.GUESSER
        card, explaination, cost = self.guess_inference(story, cards)
        self.cost += cost
        self.guessed_card_id = card.card_id
        card.guessed_by = self.player_id
        logger.info(f"[Guessing] [{self.player_id}] [{', '.join([card.card_id for card in cards])}] -> [{card.card_id}] (cost: {cost:.6f} USD)")
        return card
    
    def vote(self, story: str, cards: List[Card]) -> Card:
        assert self.role == Role.GUESSER
        card, explaination, cost = self.vote_inference(story, cards)
        self.cost += cost
        self.voted_card_id = card.card_id
        card.voted_by.append(self.player_id)
        logger.info(f"[Voting] [{self.player_id}] [{', '.join([card.card_id for card in cards])}] -> [{card.card_id}] (cost: {cost:.6f} USD)")
        return card
    
class HumanPlayer(Player):
    def __init__(self, player_id: str):
        super().__init__(player_id)
    
    def storytell_inference(self, card: Card) -> Tuple[str, str, float]:
        # TODO: implement human storytell inference
        return "This is a test story", "This is a test explaination", 0.0
    
    def guess_inference(self, story: str, cards: List[Card]) -> Tuple[Card, str, float]:
        # TODO: implement human guess inference
        return random.choice(cards), "This is a test explaination", 0.0
    
    def vote_inference(self, story: str, cards: List[Card]) -> Tuple[Card, str, float]:
        # TODO: implement human vote inference
        return random.choice(cards), "This is a test explaination", 0.0
    
class OpenAIPlayer(Player):
    def __init__(self, player_id: str, model_name: str, debug_mode: bool = False):
        super().__init__(player_id)
        self.model_name = model_name
        self.debug_mode = debug_mode
        self.client = OpenAI(api_key=os.getenv("OPENAI_API_KEY"))
    
    def storytell_inference(self, card: Card) -> Tuple[str, str, float]:
        if self.debug_mode:
            return "test clue", "explaination", 0.0
    
        card_file = self.client.files.create(
            file=open(card.image_path.resolve(), "rb"),
            purpose="user_data"
        )   
        response = self.client.responses.parse(
            model=self.model_name,
            input=[
                {
                    "role": "user",
                    "content": [
                        {
                            "type": "input_text",
                            "text": StorytellPrompt,
                        },
                        {
                            "type": "input_image",
                            "file_id": card_file.id,
                        }
                    ]
                }
            ],
            text_format=StorytellResponse,
        )
        clue = response.output_parsed.clue
        explaination = response.output_parsed.explaination    
        cost = estimate_cost_typed(response)
        return clue, explaination, cost.total_cost
    
    def guess_inference(self, story: str, cards: List[Card]) -> Tuple[Card, str, float]:
        if self.debug_mode:
            return cards[0], "explaination", 0.0
    
        card_content_list = list()
        for card in cards:
            card_file = self.client.files.create(
                file=open(card.image_path.resolve(), "rb"), 
                purpose="user_data"
            )  
            card_content_list.append({
                "type": "input_image",
                "file_id": card_file.id,
            })
            

        response = self.client.responses.parse(
            model=self.model_name,
            input=[
                {
                    "role": "user",
                    "content": [
                        {
                            "type": "input_text",
                            "text": GuessPrompt.format(story=story)
                        },
                    ] + card_content_list
                }
            ],
            text_format=GuessResponse,
        )
        card_index = response.output_parsed.card_id
        explaination = response.output_parsed.explaination
        cost = estimate_cost_typed(response)

        return cards[card_index%len(cards)], explaination, cost.total_cost
    
    def vote_inference(self, story: str, cards: List[Card]) -> Tuple[Card, str, float]:
        if self.debug_mode:
            return cards[0], "explaination", 0.0
    
        card_content_list = list()
        for card in cards:
            card_file = self.client.files.create(
                file=open(card.image_path.resolve(), "rb"), 
                purpose="user_data"
            )  
            card_content_list.append({
                "type": "input_image",
                "file_id": card_file.id,
            })

        response = self.client.responses.parse(
            model=self.model_name,
            input=[
                {
                    "role": "user",
                    "content": [
                        {
                            "type": "input_text",
                            "text": VotePrompt.format(story=story)
                        },
                    ] + card_content_list
                }
            ],
            text_format=VoteResponse,
        )
        card_index = response.output_parsed.card_id
        explaination = response.output_parsed.explaination
        cost = estimate_cost_typed(response)
        
        return cards[card_index%len(cards)], explaination, cost.total_cost
    
class Leaderboard():
    def __init__(self, player_list: List[Player]):
        self.player_list = player_list
        self.player_rating_dict = {player.player_id: [Rating()] for player in player_list}

    @staticmethod
    def get_ranking(score_dict: Dict[str, int]) -> Dict[str, int]:
        # 1. Get all unique scores, and sort them in descending order
        unique_scores = sorted(set(score_dict.values()), reverse=True)

        # 2. Create a mapping from score to rank (rank starts from 1)
        score_to_rank = {score: rank + 1 for rank, score in enumerate(unique_scores)}

        # 3. Assign the correct rank to each player using the mapping
        player_ranking = {player: score_to_rank[score] for player, score in score_dict.items()}

        return player_ranking
    
    def get_rating_trace(self, player_id: str) -> List[Rating]:
        return self.player_rating_dict[player_id]
    
    def get_latest_rating(self) -> Dict[str, Rating]:
        return {player.player_id: self.player_rating_dict[player.player_id][-1] for player in self.player_list}
    
    def rating_update(self, score_dict: Dict[str, int]):
        player_ranking_dict = self.get_ranking(score_dict)
        player_rating_groups = [[self.player_rating_dict[player.player_id][-1]] for player in self.player_list]
        player_ranks = [player_ranking_dict[player.player_id] for player in self.player_list]
        new_player_rating_list = rate(rating_groups=player_rating_groups, ranks=player_ranks)
        for player, new_rating in zip(self.player_list, new_player_rating_list):
            self.player_rating_dict[player.player_id].append(new_rating[0])

    def visualize_rating(self, game_id: str) -> None:
        fig, ax = plt.subplots(figsize=(24, 8))
        for player_id, rating_trace in self.player_rating_dict.items():
            rating = rating_trace[-1]
            mu = rating.mu
            sigma = rating.sigma
            dist = norm(loc=mu, scale=sigma)
            x = np.linspace(mu - 4*sigma, mu + 4*sigma, 200)
            pdf = dist.pdf(x)
            ax.plot(x, pdf, label=f'{player_id} (μ={mu:.2f}, σ={sigma:.2f})')

        ax.set_title("Player Skill Distributions (TrueSkill Ratings)", fontsize=16)
        ax.set_xlabel("Skill Level (μ)", fontsize=12)
        ax.set_ylabel("Probability Density", fontsize=12)
        ax.legend()
        ax.grid(True, linestyle='--', alpha=0.6)

        plt.tight_layout()
        plt.savefig(f"player_skill_distributions_{game_id}.png")
        
class Game:
    def __init__(self, player_list: List[Player], card_dir: Path):
        self.game_id = str(uuid.uuid4())
        self.game_cost = 0
        self.card_dir = card_dir
        self.player_list = player_list
        self.num_players = len(player_list)
        self.storyteller = None
        self.guesser_list = list()
        self.table_cards = dict()
        self.deck = self.initialize_deck(card_dir)
        self.score_dict = {player.player_id: 0 for player in self.player_list}
        logger.info(f"[Game] Game {self.game_id} initialized (num_players: {self.num_players})")

    def initialize_deck(self, card_dir: Path):
        temp_deck = list()
        for card_file in card_dir.glob("*.png"):
            card = Card(card_file.stem, card_file)
            temp_deck.append(card)
        random.shuffle(temp_deck)
        logger.info(f"[Game] Game {self.game_id} deck initialized ({len(temp_deck)} cards)")
        return temp_deck
    
    def pick_cards(self, num_cards: int):
        return random.sample(self.deck, num_cards)
    
    def get_card_by_id(self, card_id: int) -> Card:
        for card in self.deck:
            if card.card_id == card_id:
                return card
        raise ValueError(f"[Game] Card {card_id} not found")
    
    def role_assignment(self) -> None:
        random.shuffle(self.player_list)
        self.storyteller = self.player_list[0]
        self.guesser_list = self.player_list[1:]

        self.storyteller.role = Role.STORYTELLER
        for guesser in self.guesser_list:
            guesser.role = Role.GUESSER
        logger.info(f"[Game] Game {self.game_id} role assigned (storyteller: {self.storyteller.player_id}, guesser: {', '.join([guesser.player_id for guesser in self.guesser_list])})")
    
    def dealing(self) -> Tuple[Card, Dict[str, List[Card]]]:
        picked_cards = self.pick_cards(1 + len(self.guesser_list) * 6)
        storyteller_card = picked_cards.pop(0)
        storyteller_card.guessed_by = self.storyteller.player_id
        guesser_card_dict = {guesser.player_id: picked_cards[index*6:(index+1)*6] for index, guesser in enumerate(self.guesser_list)}
        return storyteller_card, guesser_card_dict

    def calculate_score(self):
        correct_voted_list = [guesser for guesser in self.guesser_list if guesser.voted_card_id == self.storyteller.selected_card_id]

        if self.num_players == 5:
            # Storyteller
            self.score_dict[self.storyteller.player_id] += 3 if len(correct_voted_list) in [1, 2, 3] else 0

            # Guesser (Find the Storyteller's card)
            if len(correct_voted_list) == 0:
                for guesser in self.guesser_list:
                    self.score_dict[guesser.player_id] += 2
            elif len(correct_voted_list) in [1, 2, 3]:
                for guesser in correct_voted_list:
                    self.score_dict[guesser.player_id] += 3
            elif len(correct_voted_list) == 4:
                for guesser in correct_voted_list:
                    self.score_dict[guesser.player_id] += 2

            # Guesser (Others vote for guesser's card)
            for guesser in self.guesser_list:
                guessed_card = self.get_card_by_id(guesser.voted_card_id)
                if guessed_card.guessed_by != self.storyteller.player_id:
                    self.score_dict[guessed_card.guessed_by] += 1
        elif self.num_players == 6:
            # Storyteller
            self.score_dict[self.storyteller.player_id] += 3 if len(correct_voted_list) in [1, 2, 3, 4] else 0

            # Guesser (Find the Storyteller's card)
            if len(correct_voted_list) == 0:
                for guesser in self.guesser_list:
                    self.score_dict[guesser.player_id] += 2
            elif len(correct_voted_list) in [1, 2, 3, 4]:
                for guesser in correct_voted_list:
                    self.score_dict[guesser.player_id] += 3
            elif len(correct_voted_list) == 5:
                for guesser in correct_voted_list:
                    self.score_dict[guesser.player_id] += 2

            # Guesser (Others vote for guesser's card)
            for guesser in self.guesser_list:
                guessed_card = self.get_card_by_id(guesser.voted_card_id)
                if guessed_card.guessed_by != self.storyteller.player_id:
                    self.score_dict[guessed_card.guessed_by] += 1
        else:
            raise ValueError(f"[Game] Unsupported number of players {self.num_players}")
    
    def loop(self):
        logger.info(f"[Game] Game {self.game_id} started")
        self.game_cost = 0
        self.table_cards = dict()
        self.score_dict = {player.player_id: 0 for player in self.player_list}

        # Step 1: Role assignment
        self.role_assignment()
        logger.info("[+] [Role Assignment]" + "="*50)

        # Step 2: Dealing
        storyteller_card, guesser_card_dict = self.dealing()
        logger.info("[+] [Dealing]" + "="*50)

        # Step 3: Storytelling
        story = self.storyteller.storytell(storyteller_card)
        logger.info(f"Game {self.game_id} storyteller card: {storyteller_card.card_id} -> '{story}'")
        logger.info("[+] [Storytelling]" + "="*50)
        
        # Step 4: Guessing
        for guesser in self.guesser_list:
            guessed_card = guesser.guess(story, guesser_card_dict[guesser.player_id])
            guessed_card.guessed_by = guesser.player_id
            self.table_cards[guessed_card.card_id] = guessed_card
        self.table_cards[storyteller_card.card_id] = storyteller_card
        logger.info(f"Game {self.game_id} guessed cards: {', '.join([card.card_id for card in self.table_cards.values()])}")
        logger.info("[+] [Guessing]" + "="*50)

        # Step 5: Voting
        for guesser in self.guesser_list:
            voted_card = guesser.vote(story, [self.table_cards[card_id] for card_id in self.table_cards])
            self.table_cards[voted_card.card_id] = voted_card
        logger.info(f"Game {self.game_id} voted cards: {', '.join([card.card_id for card in self.table_cards.values()])}")
        logger.info("[+] [Voting]" + "="*50)

        # Step 6: Calculate score
        self.calculate_score()
        logger.info(f"Game {self.game_id} scores calculated: {self.score_dict}")
        logger.info("[+] [Score Calculation]" + "="*50)

        # Step 7: Calculate cost
        self.game_cost = sum([player.cost for player in self.player_list])
        logger.info(f"Game {self.game_id} ended (cost: {self.game_cost:.6f} USD)")
        logger.info("[+] [End]" + "="*50)

        return self.score_dict

if __name__ == "__main__":
    player_list = [
        OpenAIPlayer(player_id="player_0", model_name="gpt-4.1"),
        OpenAIPlayer(player_id="player_1", model_name="gpt-4o-mini"),
        OpenAIPlayer(player_id="player_2", model_name="gpt-5"),
        OpenAIPlayer(player_id="player_3", model_name="gpt-5-mini"),
        OpenAIPlayer(player_id="player_4", model_name="gpt-5-nano"),
    ]
    card_dir = Path('asset/cards')
    leaderboard = Leaderboard(player_list)
    game = Game(player_list, card_dir)

    for i in range(20):
        print(f"==================== Game {i} started ====================")
        try:
            score_dict= game.loop()
            leaderboard.rating_update(score_dict)
            leaderboard.visualize_rating(f"round_{i}_{game.game_id}")
            logger.info(f"Game {game.game_id} leaderboard: {leaderboard.get_latest_rating()}")
        except Exception as e:
            print(e)
        print(f"====================  Game {i} ended  ====================")