File size: 5,938 Bytes
bae78d7 | 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 | """
Custom Chess Tokenizer for the Chess Challenge.
This tokenizer uses sub-token decomposition to achieve a minimal vocabulary
by breaking moves into atomic components (squares + modifiers).
Example:
WPe2e4 → ["e2", "e4"]
BNg8f6(+) → ["g8", "f6", "+"]
This approach trades sequence length (3x longer) for vocabulary size (77 vs 1800+).
"""
from __future__ import annotations
import json
import os
from typing import Dict, List, Optional
from transformers import PreTrainedTokenizer
class ChessTokenizer(PreTrainedTokenizer):
"""
Sub-token chess tokenizer with minimal fixed vocabulary.
Decomposes each move into:
- Source square (e.g., e2)
- Destination square (e.g., e4)
- Optional modifiers (x, +, +*, Q/R/B/N, O/o)
Vocabulary composition:
- 64 squares (a1-h8)
- 9 modifiers (x, +, +*, Q, R, B, N, O, o)
- 4 special tokens ([PAD], [BOS], [EOS], [UNK])
Total: 77 tokens
"""
model_input_names = ["input_ids", "attention_mask"]
vocab_files_names = {"vocab_file": "vocab.json"}
# Special tokens
PAD_TOKEN = "[PAD]"
BOS_TOKEN = "[BOS]"
EOS_TOKEN = "[EOS]"
UNK_TOKEN = "[UNK]"
def __init__(
self,
vocab_file: Optional[str] = None,
vocab: Optional[Dict[str, int]] = None,
**kwargs,
):
"""
Initialize the chess tokenizer.
"""
# Initialize special tokens
self._pad_token = self.PAD_TOKEN
self._bos_token = self.BOS_TOKEN
self._eos_token = self.EOS_TOKEN
self._unk_token = self.UNK_TOKEN
kwargs.pop("pad_token", None)
kwargs.pop("bos_token", None)
kwargs.pop("eos_token", None)
kwargs.pop("unk_token", None)
# Load or create vocabulary
if vocab is not None:
self._vocab = vocab
elif vocab_file is not None and os.path.exists(vocab_file):
with open(vocab_file, "r", encoding="utf-8") as f:
self._vocab = json.load(f)
else:
self._vocab = self.build_minimal_vocab().get_vocab()
# Reverse mapping
self._ids_to_tokens = {v: k for k, v in self._vocab.items()}
# Call parent init
super().__init__(
pad_token=self._pad_token,
bos_token=self._bos_token,
eos_token=self._eos_token,
unk_token=self._unk_token,
**kwargs,
)
@classmethod
def build_minimal_vocab(cls) -> "ChessTokenizer":
"""
Build tokenizer with minimal fixed vocabulary (77 tokens).
"""
files = "abcdefgh"
ranks = "12345678"
squares = [f + r for f in files for r in ranks]
modifiers = ["x", "+", "+*", "Q", "R", "B", "N", "O", "o"]
special_tokens = [cls.PAD_TOKEN, cls.BOS_TOKEN, cls.EOS_TOKEN, cls.UNK_TOKEN]
vocab_tokens = special_tokens + squares + modifiers
vocab = {tok: i for i, tok in enumerate(vocab_tokens)}
return cls(vocab=vocab)
@property
def vocab_size(self) -> int:
return len(self._vocab)
def get_vocab(self) -> Dict[str, int]:
return dict(self._vocab)
def _tokenize(self, text: str) -> List[str]:
"""
Tokenize moves into squares + modifiers.
Examples:
WPe2e4 -> ["e2", "e4"]
BNg8f6(+) -> ["g8", "f6", "+"]
WKe1g1(O) -> ["e1", "g1", "O"]
"""
tokens = []
for move in text.strip().split():
if len(move) < 4:
continue
core = move[2:] # Remove color + piece
# Squares
from_sq = core[0:2]
to_sq = core[2:4]
tokens.extend([from_sq, to_sq])
# Modifiers
suffix = core[4:]
if "x" in suffix:
tokens.append("x")
if "+*" in suffix:
tokens.append("+*")
elif "+" in suffix:
tokens.append("+")
for promo in ["Q", "R", "B", "N"]:
if f"({promo})" in suffix:
tokens.append(promo)
# Castling
if "O" in move or "o" in move:
tokens.append("O")
return tokens
def _convert_token_to_id(self, token: str) -> int:
return self._vocab.get(token, self._vocab.get(self.UNK_TOKEN, 0))
def _convert_id_to_token(self, index: int) -> str:
return self._ids_to_tokens.get(index, self.UNK_TOKEN)
def convert_tokens_to_string(self, tokens: List[str]) -> str:
"""Convert sub-tokens back to string representation."""
special = {self.PAD_TOKEN, self.BOS_TOKEN, self.EOS_TOKEN, self.UNK_TOKEN}
return " ".join(t for t in tokens if t not in special)
def save_vocabulary(self, save_directory: str, filename_prefix: Optional[str] = None) -> tuple:
if not os.path.isdir(save_directory):
os.makedirs(save_directory, exist_ok=True)
vocab_file = os.path.join(
save_directory,
(filename_prefix + "-" if filename_prefix else "") + "vocab.json",
)
with open(vocab_file, "w", encoding="utf-8") as f:
json.dump(self._vocab, f, ensure_ascii=False, indent=2)
return (vocab_file,)
# ===== Example usage =====
if __name__ == "__main__":
tokenizer = ChessTokenizer.build_minimal_vocab()
print(f"Vocabulary size: {tokenizer.vocab_size}")
test_games = [
"WPe2e4 BPe7e5",
"WNg1f3 BNb8c6",
"WBb5c6(x) BPd7d6",
"WPe7e8(Q) BKe8d7",
"WKe1g1(O) BKe8c8(o)",
]
for game in test_games:
print(f"\nOriginal: {game}")
tokens = tokenizer._tokenize(game)
print(f"Tokens: {tokens}")
ids = tokenizer.convert_tokens_to_ids(tokens)
print(f"IDs: {ids}")
decoded = tokenizer.convert_ids_to_tokens(ids)
print(f"Decoded: {decoded}")
|