Convert between token IDs and names

token.py

import re

TOKEN_BASE = 36
TOKEN_ID_MAX = 4873763662273663091  # ZZZZZZZZZZZZ
TOKEN_MAX = "ZZZZZZZZZZZZ"


def token_char_encode(char: str) -> int:
    if ord(char) >= ord("A") and ord(char) <= ord("Z"):
        return 10 + ord(char) - ord("A")
    if ord(char) >= ord("0") and ord(char) <= ord("9"):
        return ord(char) - ord("0")
    raise AssertionError("unknown character {}".format(char))


def token_char_decode(id: int) -> str:
    assert TOKEN_BASE > id >= 0, "invalid char"
    if id < 10:
        return chr(ord("0") + id)
    return chr(ord("A") + id - 10)


def token_id_encode(name: str) -> int:
    """
    encode native token name to uint64
    """
    assert len(name) < 13, "name too long"
    assert re.match(r"^[0-9A-Z]+$", name), "name can only contain 0-9, A-Z"
    id = token_char_encode(name[-1])
    base = TOKEN_BASE
    for c in name[-2::-1]:
        id += base * (token_char_encode(c) + 1)
        base *= TOKEN_BASE
    return id


def token_id_decode(id: int) -> str:
    """
    decode native token name from uint64
    """
    assert 0 <= id <= TOKEN_ID_MAX, "id too big or negative"
    name = token_char_decode(id % TOKEN_BASE)
    id = id // TOKEN_BASE - 1
    while id >= 0:
        name += token_char_decode(id % TOKEN_BASE)
        id = id // TOKEN_BASE - 1
    return name[::-1]


if __name__ == '__main__':
    print(token_id_encode("QFIRE"))  # 46121774
    print(token_id_decode(46121774))  # QFIRE