"""Stable internal identifiers: prefixed ULIDs (`co_01J…`). Slugs, names and URLs change; ids never do (spec §173)."""
from __future__ import annotations

import hashlib
import re

from slugify import slugify as _slugify
from ulid import ULID

PREFIXES: dict[str, str] = {
    "company": "co",
    "domain": "dom",
    "sensor": "sen",
    "observation": "obs",
    "snapshot": "snap",
    "change": "chg",
    "event": "evt",
    "cluster": "cl",
    "job": "job",
    "person": "person",
    "product": "prod",
    "plan": "plan",
    "location": "loc",
    "news": "news",
    "relationship": "rel",
    "queue_job": "qj",
    "llm_job": "llm",
    "review": "rev",
    "failure": "fail",
    "crawl_run": "run",
    "watchlist": "wl",
    "alert": "alert",
    "api_key": "key",
    "signal": "sig",
}


def new_id(kind: str) -> str:
    prefix = PREFIXES.get(kind)
    if prefix is None:
        raise ValueError(f"unknown id kind {kind!r}")
    return f"{prefix}_{ULID()}"


def kind_of(entity_id: str) -> str | None:
    prefix = entity_id.split("_", 1)[0]
    for kind, p in PREFIXES.items():
        if p == prefix:
            return kind
    return None


_slug_clean = re.compile(r"[^a-z0-9.-]+")


def slugify(text: str, *, max_length: int = 80) -> str:
    """URL slug (`stripe`, `jpmorgan-chase`, `l-oreal`)."""
    s = _slugify(text, lowercase=True, regex_pattern=r"[^a-z0-9-]+", replacements=[("&", " and "), ("/", "-"), ("_", "-"), ("@", "-at-")])
    s = _slug_clean.sub("-", s).strip("-.")
    return s[:max_length].rstrip("-.") or "company"


def normalize_alias(text: str) -> str:
    """Deterministic alias key: lowercase ASCII, punctuation collapsed, legal suffixes dropped. Resolution only — never display."""
    s = _slugify(text, lowercase=True, regex_pattern=r"[^a-z0-9]+")
    s = s.replace("-", "")
    for suffix in ("incorporated", "corporation", "limited", "company", "holdings", "group", "inc", "corp", "ltd", "plc", "llc",
                   "gmbh", "ag", "sa", "nv", "bv", "se", "ab", "oyj", "asa", "spa", "srl", "kk", "co"):
        if s.endswith(suffix) and len(s) > len(suffix) + 2:
            s = s[: -len(suffix)]
            break
    return s


def stable_hash(*parts: str, length: int = 24) -> str:
    """Deterministic key for idempotent rows (queue jobs `sensor_id + window`, job fingerprints, cluster keys)."""
    h = hashlib.sha256("\x1f".join(parts).encode("utf-8")).hexdigest()
    return h[:length]


__all__ = ["PREFIXES", "kind_of", "new_id", "normalize_alias", "slugify", "stable_hash"]