"""Content-addressable object store for raw observations, normalized text and block structures (spec §26, §96, §97).

    objects/<sha256[0:2]>/<sha256[2:4]>/<sha256>.zst

Objects are written once (dedupe by hash) and never mutated; many observations may point at the same key. zstd level 9 gives
~8× on HTML. `key` = sha256 hex of the *uncompressed* bytes; the relative path is derived, so the store can move as a whole.
"""
from __future__ import annotations

import hashlib
import os
from pathlib import Path

import zstandard as zstd

from companyatlas.config import settings

_cctx = zstd.ZstdCompressor(level=9)
_dctx = zstd.ZstdDecompressor()


def object_path(key: str, root: Path | None = None) -> Path:
    root = root or settings.objects_dir
    return root / key[:2] / key[2:4] / f"{key}.zst"


def sha256_hex(data: bytes) -> str:
    return hashlib.sha256(data).hexdigest()


def put_bytes(data: bytes, *, key: str | None = None) -> tuple[str, int, bool]:
    """Store bytes; return (key, stored_size_bytes, created). Idempotent."""
    key = key or sha256_hex(data)
    path = object_path(key)
    if path.exists():
        return key, path.stat().st_size, False
    path.parent.mkdir(parents=True, exist_ok=True)
    tmp = path.with_suffix(f".zst.tmp{os.getpid()}")
    with open(tmp, "wb") as fh:
        fh.write(_cctx.compress(data))
    tmp.replace(path)
    return key, path.stat().st_size, True


def put_text(text: str) -> tuple[str, int, bool]:
    return put_bytes(text.encode("utf-8"))


def get_bytes(key: str) -> bytes:
    with open(object_path(key), "rb") as fh:
        return _dctx.decompress(fh.read(), max_output_size=64 * 1024 * 1024)


def get_text(key: str) -> str:
    return get_bytes(key).decode("utf-8", errors="replace")


def exists(key: str) -> bool:
    return object_path(key).exists()


def store_stats(root: Path | None = None) -> dict[str, int]:
    root = root or settings.objects_dir
    total = count = 0
    if root.exists():
        for dirpath, _dirs, files in os.walk(root):
            for f in files:
                if f.endswith(".zst"):
                    try:
                        total += os.stat(os.path.join(dirpath, f)).st_size
                        count += 1
                    except OSError:
                        pass
    return {"objects": count, "bytes": total}


__all__ = ["exists", "get_bytes", "get_text", "object_path", "put_bytes", "put_text", "sha256_hex", "store_stats"]