from __future__ import annotations

import re

GLFM_CHUNK_LIMIT = 65000

_FENCED_BLOCK_RE = re.compile(r"^```")


def sanitize_glfm(text: str) -> str:
    return text.replace("\x00", "")


def chunk_glfm_markdown(text: str, limit: int = GLFM_CHUNK_LIMIT) -> list[str]:
    if len(text) <= limit:
        return [text]

    chunks: list[str] = []
    in_fenced_block = False
    current = ""

    for line in text.split("\n"):
        if _FENCED_BLOCK_RE.match(line):
            in_fenced_block = not in_fenced_block

        candidate = current + line + "\n"
        if len(candidate) > limit and not in_fenced_block:
            chunks.append(current.rstrip())
            current = line + "\n"
        else:
            current = candidate

    if current.strip():
        chunks.append(current.rstrip())

    return chunks if chunks else [text]


def safe_glfm_body(text: str, max_length: int = 950000) -> str:
    if len(text) > max_length:
        return text[: max_length - 3] + "..."
    return text