Source code for pythainlp.util.encoding

# -*- coding: utf-8 -*-
# SPDX-FileCopyrightText: 2016-2026 PyThaiNLP Project
# SPDX-FileType: SOURCE
# SPDX-License-Identifier: Apache-2.0
from __future__ import annotations




def tis620_to_utf8(text: str) -> str:
    """Convert TIS-620 to UTF-8

    :param str text: TIS-620 encoded text
    :return: UTF-8 encoded text
    :rtype: str

    :Example:

        >>> from pythainlp.util import tis620_to_utf8
        >>> tis620_to_utf8("¡ÃзÃǧÍصÊÒË¡ÃÃÁ")
        'กระทรวงอุตสาหกรรม'
    """
    return text.encode("cp1252", "ignore").decode("tis-620")





def to_idna(text: str) -> str:
    """Encode text with IDNA, as used in Internationalized Domain Name (IDN).

    :param str text: Thai text
    :return: IDNA-encoded text
    :rtype: str

    :Example:

        >>> from pythainlp.util import to_idna
        >>> to_idna("คนละครึ่ง.com")
        'xn--42caj4e6bk1f5b1j.com'
    """
    return text.encode("idna").decode("utf-8")