# Urwid unicode character processing tables
#    Copyright (C) 2004-2011  Ian Ward
#
#    This library is free software; you can redistribute it and/or
#    modify it under the terms of the GNU Lesser General Public
#    License as published by the Free Software Foundation; either
#    version 2.1 of the License, or (at your option) any later version.
#
#    This library is distributed in the hope that it will be useful,
#    but WITHOUT ANY WARRANTY; without even the implied warranty of
#    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
#    Lesser General Public License for more details.
#
#    You should have received a copy of the GNU Lesser General Public
#    License along with this library; if not, write to the Free Software
#    Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
#
# Urwid web site: https://urwid.org/


from __future__ import annotations

import re
import typing
import warnings

import wcwidth

if typing.TYPE_CHECKING:
    from typing_extensions import Literal

SAFE_ASCII_RE = re.compile(r"^[ -~]*$")
SAFE_ASCII_BYTES_RE = re.compile(rb"^[ -~]*$")

_byte_encoding: Literal["utf8", "narrow", "wide"] = "narrow"


def get_char_width(char: str) -> Literal[0, 1, 2]:
    """
    Return the screen column width for a single character.

    .. deprecated:: 3.0.4
    """
    warnings.warn(
        "get_char_width is deprecated in favor of wcwidth.width",
        DeprecationWarning,
        stacklevel=2,
    )
    if (width := wcwidth.wcwidth(char)) >= 0:
        return width

    return 0


def get_width(o: int) -> Literal[0, 1, 2]:
    """
    Return the screen column width for unicode ordinal o.

    .. deprecated:: 3.0.4
    """
    warnings.warn(
        "get_width is deprecated in favor of wcwidth.width",
        DeprecationWarning,
        stacklevel=2,
    )
    if (width := wcwidth.wcwidth(chr(o))) >= 0:
        return width

    return 0


def _decode_grapheme_at(text: bytes, start: int, end: int) -> tuple[str, int]:
    """
    Decode bytes starting at `start` to get the first grapheme cluster.

    :param text: UTF-8 encoded bytes
    :param start: starting byte position
    :param end: ending byte position
    :returns: (grapheme_string, next_byte_position)

    Assumes caller provides valid UTF-8 byte boundaries.
    """
    decoded = text[start:end].decode("utf-8")
    grapheme = next(wcwidth.iter_graphemes(decoded), "")
    grapheme_bytes = grapheme.encode("utf-8")
    return grapheme, start + len(grapheme_bytes)


def decode_one(text: bytes | str, pos: int) -> tuple[int, int]:
    """
    Return (ordinal at pos, next position) for UTF-8 encoded text.
    """
    lt = len(text) - pos

    b2 = 0  # Fallback, not changing anything
    b3 = 0  # Fallback, not changing anything
    b4 = 0  # Fallback, not changing anything

    try:
        if isinstance(text, str):
            b1 = ord(text[pos])
            if lt > 1:
                b2 = ord(text[pos + 1])
            if lt > 2:
                b3 = ord(text[pos + 2])
            if lt > 3:
                b4 = ord(text[pos + 3])
        else:
            b1 = text[pos]
            if lt > 1:
                b2 = text[pos + 1]
            if lt > 2:
                b3 = text[pos + 2]
            if lt > 3:
                b4 = text[pos + 3]
    except Exception as e:
        raise ValueError(f"{e}: text={text!r}, pos={pos!r}, lt={lt!r}").with_traceback(e.__traceback__) from e

    if not b1 & 0x80:
        return b1, pos + 1
    error = ord("?"), pos + 1

    if lt < 2:
        return error
    if b1 & 0xE0 == 0xC0:
        if b2 & 0xC0 != 0x80:
            return error
        if (o := ((b1 & 0x1F) << 6) | (b2 & 0x3F)) >= 0x80:
            return o, pos + 2
        return error
    if lt < 3:
        return error
    if b1 & 0xF0 == 0xE0:
        if b2 & 0xC0 != 0x80:
            return error
        if b3 & 0xC0 != 0x80:
            return error
        if (o := ((b1 & 0x0F) << 12) | ((b2 & 0x3F) << 6) | (b3 & 0x3F)) >= 0x800:
            return o, pos + 3
        return error
    if lt < 4:
        return error
    if b1 & 0xF8 == 0xF0:
        if b2 & 0xC0 != 0x80:
            return error
        if b3 & 0xC0 != 0x80:
            return error
        if b4 & 0xC0 != 0x80:
            return error
        if (o := ((b1 & 0x07) << 18) | ((b2 & 0x3F) << 12) | ((b3 & 0x3F) << 6) | (b4 & 0x3F)) >= 0x10000:
            return o, pos + 4
        return error
    return error


def decode_one_uni(text: str, i: int) -> tuple[int, int]:
    """
    decode_one implementation for unicode strings
    """
    return ord(text[i]), i + 1


def decode_one_right(text: bytes, pos: int) -> tuple[int, int] | None:
    """
    Return (ordinal at pos, next position) for UTF-8 encoded text.
    pos is assumed to be on the trailing byte of a utf-8 sequence.
    """
    if not isinstance(text, bytes):
        raise TypeError(text)
    error = ord("?"), pos - 1
    p = pos
    while p >= 0:
        if text[p] & 0xC0 != 0x80:
            o, _next_pos = decode_one(text, p)
            return o, p - 1
        p -= 1
        if p == p - 4:
            return error
    return None


def set_byte_encoding(enc: Literal["utf8", "narrow", "wide"]) -> None:
    if enc not in {"utf8", "narrow", "wide"}:
        raise ValueError(enc)
    global _byte_encoding  # noqa: PLW0603  # pylint: disable=global-statement
    _byte_encoding = enc


def get_byte_encoding() -> Literal["utf8", "narrow", "wide"]:
    return _byte_encoding


def calc_string_text_pos(text: str, start_offs: int, end_offs: int, pref_col: int) -> tuple[int, int]:
    """
    Calculate the closest position to the screen column pref_col in text
    where start_offs is the offset into text assumed to be screen column 0
    and end_offs is the end of the range to search.

    Iterates by grapheme clusters for emoji ZWJ sequences, flags,
    combining characters, and other multi-codepoint unicode sequences.

    :param text: string
    :param start_offs: starting text position
    :param end_offs: ending text position
    :param pref_col: target column
    :returns: (position, actual_col)
    """
    if start_offs > end_offs:
        raise ValueError((start_offs, end_offs))

    cols = 0
    pos = start_offs
    for grapheme in wcwidth.iter_graphemes(text[start_offs:end_offs]):
        grapheme_width = wcwidth.width(grapheme, control_codes="ignore")
        if grapheme_width + cols > pref_col:
            return pos, cols
        cols += grapheme_width
        pos += len(grapheme)

    return end_offs, cols


def calc_text_pos(text: str | bytes, start_offs: int, end_offs: int, pref_col: int) -> tuple[int, int]:
    """
    Calculate the closest position to the screen column pref_col in text
    where start_offs is the offset into text assumed to be screen column 0
    and end_offs is the end of the range to search.

    text may be unicode or a byte string in the target _byte_encoding

    Returns (position, actual_col).
    """
    if start_offs > end_offs:
        raise ValueError((start_offs, end_offs))

    if isinstance(text, str):
        return calc_string_text_pos(text, start_offs, end_offs, pref_col)

    if not isinstance(text, bytes):
        raise TypeError(text)

    if _byte_encoding == "utf8":
        decoded = text[start_offs:end_offs].decode("utf-8")
        str_pos, cols = calc_string_text_pos(decoded, 0, len(decoded), pref_col)
        byte_offset = len(decoded[:str_pos].encode("utf-8"))
        return start_offs + byte_offset, cols

    # "wide" and "narrow"
    i = start_offs + pref_col
    if i >= end_offs:
        return end_offs, end_offs - start_offs
    if _byte_encoding == "wide" and within_double_byte(text, start_offs, i) == 2:
        i -= 1
    return i, i - start_offs


def calc_width(text: str | bytes, start_offs: int, end_offs: int) -> int:
    """
    Return the screen column width of text between start_offs and end_offs.

    text may be unicode or a byte string in the target _byte_encoding

    Some characters are wide (take two columns) and others affect the
    previous character (take zero columns), while others are grouped
    in sequence by "grapheme boundaries" (Emoji, Skin tones, flags, etc).
    """

    if start_offs > end_offs:
        msg = f"{start_offs=} > {end_offs=}"
        raise ValueError(msg)

    if isinstance(text, str):
        return wcwidth.width(text[start_offs:end_offs], control_codes="ignore")

    if _byte_encoding == "utf8":
        try:
            return wcwidth.width(text[start_offs:end_offs].decode("utf-8"), control_codes="ignore")
        except UnicodeDecodeError as exc:
            warnings.warn(
                "`calc_width` with text encoded to bytes can produce incorrect results"
                f"due to possible offset in the middle of character: {exc}",
                UnicodeWarning,
                stacklevel=2,
            )

        i = start_offs
        sc = 0
        while i < end_offs:
            o, i = decode_one(text, i)
            if (w := wcwidth.wcwidth(chr(o))) > 0:
                sc += w
        return sc
    # "wide", "narrow" or all printable ASCII, just return the character count
    return end_offs - start_offs


def is_wide_char(text: str | bytes, offs: int) -> bool:
    """
    Test if the grapheme cluster at offs within text is wide (2 columns).

    For Unicode strings, extracts the full grapheme cluster starting at offs
    and checks if it renders as wide. This correctly handles multi-codepoint
    graphemes like emoji ZWJ sequences and flags.

    text may be unicode or a byte string in the target _byte_encoding
    """
    if isinstance(text, str):
        grapheme = next(wcwidth.iter_graphemes(text[offs:]))
        return wcwidth.width(grapheme, control_codes="ignore") == 2
    if not isinstance(text, bytes):
        raise TypeError(text)
    if _byte_encoding == "utf8":
        grapheme, _ = _decode_grapheme_at(text, offs, len(text))
        return wcwidth.width(grapheme, control_codes="ignore") == 2
    if _byte_encoding == "wide":
        return within_double_byte(text, offs, offs) == 1
    return False


def move_prev_char(text: str | bytes, start_offs: int, end_offs: int) -> int:
    """
    Return the position of the grapheme cluster before end_offs.

    For Unicode strings, handle multi-codepoint, "grapheme clusters",
    to better measure emoji ZWJ, flags, combining characters, skin tones.
    """
    if start_offs >= end_offs:
        raise ValueError((start_offs, end_offs))
    if isinstance(text, str):
        return wcwidth.grapheme_boundary_before(text, end_offs)
    if not isinstance(text, bytes):
        raise TypeError(text)
    if _byte_encoding == "utf8":
        decoded = text[start_offs:end_offs].decode("utf-8")
        str_pos = len(decoded)
        prev_str_pos = wcwidth.grapheme_boundary_before(decoded, str_pos)
        prefix = decoded[:prev_str_pos]
        return start_offs + len(prefix.encode("utf-8"))
    if _byte_encoding == "wide" and within_double_byte(text, start_offs, end_offs - 1) == 2:
        return end_offs - 2
    return end_offs - 1


def move_next_char(text: str | bytes, start_offs: int, end_offs: int) -> int:
    """
    Return the position of the next grapheme cluster after start_offs.

    For Unicode strings, handle multi-codepoint, "grapheme clusters",
    to better measure emoji ZWJ, flags, combining characters, skin tones.
    """
    if start_offs >= end_offs:
        raise ValueError((start_offs, end_offs))
    if isinstance(text, str):
        grapheme = next(wcwidth.iter_graphemes(text[start_offs:end_offs]))
        return start_offs + len(grapheme)
    if not isinstance(text, bytes):
        raise TypeError(text)
    if _byte_encoding == "utf8":
        _, next_pos = _decode_grapheme_at(text, start_offs, end_offs)
        return next_pos
    if _byte_encoding == "wide" and within_double_byte(text, start_offs, start_offs) == 1:
        return start_offs + 2
    return start_offs + 1


def within_double_byte(text: bytes, line_start: int, pos: int) -> Literal[0, 1, 2]:
    """Return whether pos is within a double-byte encoded character.

    text -- byte string in question
    line_start -- offset of beginning of line (< pos)
    pos -- offset in question

    Return values:
    0 -- not within dbe char, or double_byte_encoding == False
    1 -- pos is on the 1st half of a dbe char
    2 -- pos is on the 2nd half of a dbe char
    """
    if not isinstance(text, bytes):
        raise TypeError(text)
    v = text[pos]

    if 0x40 <= v < 0x7F:
        # might be second half of big5, uhc or gbk encoding
        if pos == line_start:
            return 0

        if text[pos - 1] >= 0x81 and within_double_byte(text, line_start, pos - 1) == 1:
            return 2
        return 0

    if v < 0x80:
        return 0

    i = pos - 1
    while i >= line_start:
        if text[i] < 0x80:
            break
        i -= 1

    if (pos - i) & 1:
        return 1
    return 2