from __future__ import annotations

import html
import re

_MD_TO_HTML_MAP = [
    (r"\*\*(.+?)\*\*", r"<b>\1</b>"),
    (r"__(.+?)__", r"<b>\1</b>"),
    (r"(?<!\*)\*(?!\*)(.+?)(?<!\*)\*(?!\*)", r"<i>\1</i>"),
    (r"(?<!_)_(?!_)(.+?)(?<!_)_(?!_)", r"<i>\1</i>"),
    (r"~~(.+?)~~", r"<del>\1</del>"),
    (r"`([^`\n]+?)`", r"<code>\1</code>"),
]

_HEADING_PATTERN = re.compile(r"^(#{1,6})\s+(.+)$", re.MULTILINE)
_LINK_PATTERN = re.compile(r"\[(.+?)\]\((.+?)\)")
_CODE_BLOCK_PATTERN = re.compile(r"```(?:\w+)?\n(.+?)```", re.DOTALL)
_BLOCKQUOTE_PATTERN = re.compile(r"^>\s?(.*)$", re.MULTILINE)
_HR_PATTERN = re.compile(r"^(?:[-*_]\s*){3,}$", re.MULTILINE)
_UNORDERED_LIST_PATTERN = re.compile(r"^[\-\*\+]\s+(.+)$", re.MULTILINE)
_ORDERED_LIST_PATTERN = re.compile(r"^\d+\.\s+(.+)$", re.MULTILINE)

_UNSAFE_TAGS = re.compile(
    r"<\s*(script|iframe|object|embed|form|input|svg|style|link|meta|base|applet)[^>]*>",
    re.IGNORECASE,
)
_UNSAFE_ATTRS = re.compile(
    r'\s(on\w+|href\s*=\s*["\']javascript:|style\s*=\s*["\']\s*expression)',
    re.IGNORECASE,
)


def markdown_to_matrix_html(text: str) -> str:
    result = text

    result = _CODE_BLOCK_PATTERN.sub(_code_block_replacement, result)

    result = _HEADING_PATTERN.sub(_heading_replacement, result)

    for pattern, replacement in _MD_TO_HTML_MAP:
        result = re.sub(pattern, replacement, result)

    result = _LINK_PATTERN.sub(r'<a href="\2">\1</a>', result)

    result = _UNORDERED_LIST_PATTERN.sub(r"<li>\1</li>", result)
    result = _ORDERED_LIST_PATTERN.sub(r"<li>\1</li>", result)

    result = _BLOCKQUOTE_PATTERN.sub(r"<blockquote>\1</blockquote>", result)

    result = _HR_PATTERN.sub(r"<hr>", result)

    result = result.replace("\n", "<br>")
    result = _sanitize_html(result)

    return result


def _code_block_replacement(match: re.Match) -> str:
    content = match.group(1)
    safe_content = html.escape(content)
    return f"<pre><code>{safe_content}</code></pre>"


def _heading_replacement(match: re.Match) -> str:
    level = len(match.group(1))
    text_content = match.group(2)
    return f"<h{level}>{text_content}</h{level}>"


def _sanitize_html(text: str) -> str:
    text = _UNSAFE_TAGS.sub("", text)
    text = _UNSAFE_ATTRS.sub("", text)
    return text