# scripts/lib/source_state.py
"""Read/write STATE.json (committed source-snapshot state) and compute content fingerprints."""
from __future__ import annotations

import hashlib
import json
from pathlib import Path
from typing import Optional

from lib.paths import repo_root


SCHEMA_VERSION = 1


def _default_state_path() -> Path:
    return repo_root() / "STATE.json"


_STATE_PATH_FN = _default_state_path


def write(data: dict) -> None:
    """Write STATE.json (sorted keys, 2-space indent — diff-friendly)."""
    data = {**data, "schema_version": SCHEMA_VERSION}
    _STATE_PATH_FN().write_text(json.dumps(data, indent=2, sort_keys=True), encoding="utf-8")


def read() -> Optional[dict]:
    path = _STATE_PATH_FN()
    if not path.is_file():
        return None
    return json.loads(path.read_text(encoding="utf-8"))


def compute_directory_fingerprint(directory: Path) -> str:
    """sha256 over (relpath, content) tuples for every file in the tree, sorted.

    relpath uses POSIX separators so fingerprints stay stable across Windows/POSIX.
    Empty directories ignored — matches git's content-addressable semantics.
    """
    if not directory.is_dir():
        raise FileNotFoundError(
            f"compute_directory_fingerprint: not a directory: {directory}"
        )
    h = hashlib.sha256()
    for fp in sorted(directory.rglob("*"), key=lambda p: p.as_posix()):
        # Empty directories ignored — matches git's content-addressable semantics.
        if not fp.is_file():
            continue
        rel = fp.relative_to(directory).as_posix()
        h.update(rel.encode("utf-8"))
        h.update(b"\x00")
        h.update(fp.read_bytes())
        h.update(b"\x00")
    return f"sha256:{h.hexdigest()}"