from __future__ import annotations

from collections.abc import Iterable, Iterator
from dataclasses import dataclass

from .parsers import CSVParser


def _unescape_dumpling_csv_char(char: str) -> str:
    if char == "0":
        return "\0"
    if char == "r":
        return "\r"
    if char == "n":
        return "\n"
    return char


_SENTINEL = "\ue000"
_ESCAPED_SENTINEL = f"{_SENTINEL}s"
_ESCAPED_NEWLINE = f"{_SENTINEL}n"
_ESCAPED_CARRIAGE_RETURN = f"{_SENTINEL}r"
_ESCAPED_NUL = f"{_SENTINEL}0"


def _append_adapted_char(parts: list[str], char: str) -> None:
    if char == _SENTINEL:
        parts.append(_ESCAPED_SENTINEL)
    elif char == "\n":
        parts.append(_ESCAPED_NEWLINE)
    elif char == "\r":
        parts.append(_ESCAPED_CARRIAGE_RETURN)
    elif char == "\0":
        parts.append(_ESCAPED_NUL)
    else:
        parts.append(char)


def _encode_sentinels(value: str) -> str:
    return value.replace(_SENTINEL, _ESCAPED_SENTINEL)


def _adapt_line(line: str) -> str:
    if not line:
        return '""'
    if "\\" not in line or '"' not in line:
        return _encode_sentinels(line) if _SENTINEL in line else line
    if _SENTINEL not in line and _line_has_no_quoted_escape(line):
        return line

    parts: list[str] = []
    field_empty = True
    in_quotes = False
    quoted_field = False
    i = 0
    length = len(line)

    while i < length:
        char = line[i]
        if in_quotes:
            if char == "\\":
                i += 1
                if i >= length:
                    parts.append("\\")
                    break
                escaped = _unescape_dumpling_csv_char(line[i])
                if escaped == '"':
                    parts.append('""')
                else:
                    _append_adapted_char(parts, escaped)
                field_empty = False
            elif char == '"':
                if i + 1 < length and line[i + 1] == '"':
                    parts.append('""')
                    i += 1
                    field_empty = False
                else:
                    parts.append('"')
                    in_quotes = False
            else:
                _append_adapted_char(parts, char)
                field_empty = False
        else:
            if char == ",":
                parts.append(char)
                field_empty = True
                quoted_field = False
            elif char == '"' and field_empty and not quoted_field:
                parts.append(char)
                in_quotes = True
                quoted_field = True
            else:
                _append_adapted_char(parts, char)
                field_empty = False
        i += 1

    if in_quotes:
        raise ValueError("CSV row has an unclosed quoted field")
    return "".join(parts)


def _line_has_no_quoted_escape(line: str) -> bool:
    in_quotes = False
    quoted_field = False
    field_empty = True
    i = 0
    length = len(line)

    while i < length:
        char = line[i]
        if in_quotes:
            if char == "\\":
                return False
            if char == '"':
                if i + 1 < length and line[i + 1] == '"':
                    i += 1
                    field_empty = False
                else:
                    in_quotes = False
            else:
                field_empty = False
        else:
            if char == ",":
                field_empty = True
                quoted_field = False
            elif char == '"' and field_empty and not quoted_field:
                in_quotes = True
                quoted_field = True
            else:
                field_empty = False
        i += 1

    if in_quotes:
        raise ValueError("CSV row has an unclosed quoted field")
    return True


def _restore_value(value: str) -> str:
    if _SENTINEL not in value:
        return value

    parts: list[str] = []
    i = 0
    length = len(value)
    while i < length:
        char = value[i]
        if char != _SENTINEL or i + 1 >= length:
            parts.append(char)
            i += 1
            continue

        code = value[i + 1]
        if code == "s":
            parts.append(_SENTINEL)
            i += 2
        elif code == "n":
            parts.append("\n")
            i += 2
        elif code == "r":
            parts.append("\r")
            i += 2
        elif code == "0":
            parts.append("\0")
            i += 2
        else:
            parts.append(char)
            i += 1

    return "".join(parts)


def _restore_row(row: list[str]) -> list[str]:
    return [_restore_value(value) for value in row]


def _strip_line_ending(line: str) -> str:
    if line.endswith("\n"):
        line = line[:-1]
    if line.endswith("\r"):
        line = line[:-1]
    return line


def _adapt_lines(handle: Iterable[str]) -> Iterator[str]:
    for line in handle:
        yield _adapt_line(_strip_line_ending(line))


@dataclass(frozen=True)
class DumplingBackslashCSVParser:
    parser: CSVParser

    def iter_rows(self, handle: Iterable[str]) -> Iterator[list[str]]:
        try:
            for row in self.parser.iter_rows(_adapt_lines(handle)):
                yield _restore_row(row)
        except ValueError as exc:
            raise self.parser.error_type(f"{self.parser.context} CSV row has an unclosed quoted field") from exc


__all__ = ["DumplingBackslashCSVParser"]