"""Raw source archive — content-addressed, gzip-compressed, deduplicated. Historical snapshots are never discarded.

Layout (under AIA_DATA_DIR):
    raw/<sha256[0:2]>/<sha256[2:4]>/<sha256>.gz      original bytes (HTML, JSON, XML, PDF…)
    text/<sha256[0:2]>/<sha256[2:4]>/<sha256>.txt.gz cleaned text
"""
from __future__ import annotations

import gzip
import hashlib
from pathlib import Path

from aiatlas.config import settings


def _path(root: Path, digest: str, suffix: str) -> Path:
    return root / digest[:2] / digest[2:4] / f"{digest}{suffix}"


def store_raw(content: bytes, *, sha256: str | None = None) -> str:
    """Write bytes once (dedupe by hash); return the path relative to raw_dir."""
    digest = sha256 or hashlib.sha256(content).hexdigest()
    path = _path(settings.raw_dir, digest, ".gz")
    if not path.exists():
        path.parent.mkdir(parents=True, exist_ok=True)
        tmp = path.with_suffix(".gz.tmp")
        with gzip.open(tmp, "wb", compresslevel=6) as fh:
            fh.write(content)
        tmp.replace(path)
    return str(path.relative_to(settings.raw_dir))


def store_text(text: str) -> tuple[str, str]:
    """Write cleaned text; return (relative path, sha256 of the text)."""
    data = text.encode("utf-8")
    digest = hashlib.sha256(data).hexdigest()
    path = _path(settings.text_dir, digest, ".txt.gz")
    if not path.exists():
        path.parent.mkdir(parents=True, exist_ok=True)
        tmp = path.with_suffix(".gz.tmp")
        with gzip.open(tmp, "wb", compresslevel=6) as fh:
            fh.write(data)
        tmp.replace(path)
    return str(path.relative_to(settings.text_dir)), digest


def load_raw(rel_path: str) -> bytes:
    with gzip.open(settings.raw_dir / rel_path, "rb") as fh:
        return fh.read()


def load_text(rel_path: str) -> str:
    with gzip.open(settings.text_dir / rel_path, "rb") as fh:
        return fh.read().decode("utf-8")


def archive_size() -> dict[str, int]:
    out: dict[str, int] = {}
    for name, root in (("raw", settings.raw_dir), ("text", settings.text_dir)):
        total = 0
        count = 0
        if root.exists():
            for p in root.rglob("*.gz"):
                try:
                    total += p.stat().st_size
                    count += 1
                except OSError:
                    pass
        out[f"{name}_bytes"] = total
        out[f"{name}_files"] = count
    return out


__all__ = ["archive_size", "load_raw", "load_text", "store_raw", "store_text"]