tokenizer.py

from __future__ import annotations

import json
import os
import re
from typing import Dict, List, Optional, Tuple

from transformers import PreTrainedTokenizer


class ChessTokenizer(PreTrainedTokenizer):

    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]"

    # Structure token
    MOVE_TOKEN = "[MOVE]"

    _MOVE_RE = re.compile(
        r'^(?P<color>[WB])(?P<piece>[PNBRQK])(?P<from>[a-h][1-8])(?P<to>[a-h][1-8])(?P<rest>.*)$'
    )
    _PROMO_RE = re.compile(r'=?([QRBNqrbn])')

    def __init__(
        self,
        vocab_file: Optional[str] = None,
        vocab: Optional[Dict[str, int]] = None,
        **kwargs,
    ):
        # 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

        # Remove any duplicate special-token entries passed through kwargs
        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._create_default_vocab()

        self._ids_to_tokens = {v: k for k, v in self._vocab.items()}

        super().__init__(
            pad_token=self._pad_token,
            bos_token=self._bos_token,
            eos_token=self._eos_token,
            unk_token=self._unk_token,
            **kwargs,
        )

    def _create_default_vocab(self) -> Dict[str, int]:
        special = [self.PAD_TOKEN, self.BOS_TOKEN, self.EOS_TOKEN, self.UNK_TOKEN, self.MOVE_TOKEN]
        return {t: i for i, t in enumerate(special)}

    @classmethod
    def build_structured_vocab(cls) -> "ChessTokenizer":
        special = [cls.PAD_TOKEN, cls.BOS_TOKEN, cls.EOS_TOKEN, cls.UNK_TOKEN, cls.MOVE_TOKEN]

        files = "abcdefgh"
        ranks = "12345678"
        squares = [f"{f}{r}" for f in files for r in ranks]  # 64

        promo = [f"promo_{p}" for p in ("q", "r", "b", "n")]

        tokens = special + squares + promo
        vocab = {t: i for i, t in enumerate(tokens)}
        return cls(vocab=vocab)

    @classmethod
    def build_vocab_from_dataset(
        cls,
        dataset_name: str = "dlouapre/lichess_2025-01_1M",
        split: str = "train",
        column: str = "text",
        min_frequency: int = 500,
        max_samples: Optional[int] = 100000,
    ) -> "ChessTokenizer":
        return cls.build_structured_vocab()

    @property
    def vocab_size(self) -> int:
        return len(self._vocab)

    def get_vocab(self) -> Dict[str, int]:
        return dict(self._vocab)

    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:
        drop = {self.PAD_TOKEN, self.BOS_TOKEN, self.EOS_TOKEN, self.UNK_TOKEN}
        return " ".join(t for t in tokens if t not in drop)

    def _decompose_one_move(self, move_tok: str) -> List[str]:
        m = self._MOVE_RE.match(move_tok)
        if not m:
            return [self.UNK_TOKEN]

        from_sq = m.group("from")
        to_sq = m.group("to")
        rest = m.group("rest") or ""

        out = [self.MOVE_TOKEN, from_sq, to_sq]

        # Promotion detection (best-effort)
        pm = self._PROMO_RE.search(rest)
        if pm:
            p = pm.group(1).lower()
            if p in ("q", "r", "b", "n"):
                out.append(f"promo_{p}")

        return out

    def _tokenize(self, text: str) -> List[str]:
        text = text.strip()
        if not text:
            return []

        special = {self.PAD_TOKEN, self.BOS_TOKEN, self.EOS_TOKEN, self.UNK_TOKEN, self.MOVE_TOKEN}

        if " " not in text:
            if text in special:
                return [text]
            if text in self._vocab:
                return [text]
            return self._decompose_one_move(text)

        out: List[str] = []
        for part in text.split():
            if part in special:
                out.append(part)
            elif part in self._vocab:
                out.append(part)
            else:
                out.extend(self._decompose_one_move(part))
        return out

    def save_vocabulary(
        self,
        save_directory: str,
        filename_prefix: Optional[str] = None,
    ) -> Tuple[str]:
        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,)


def count_vocab_from_dataset(
    dataset_name: str = "dlouapre/lichess_2025-01_1M",
    split: str = "train",
    column: str = "text",
    max_samples: Optional[int] = 10000,
) -> Dict[str, int]:
    from collections import Counter
    from datasets import load_dataset

    dataset = load_dataset(dataset_name, split=split)
    if max_samples is not None:
        dataset = dataset.select(range(min(max_samples, len(dataset))))

    token_counts = Counter()
    for example in dataset:
        moves = example[column].strip().split()
        token_counts.update(moves)

    return dict(token_counts)