"""Stable internal identifiers: prefixed ULIDs (`model_01J…`). URLs and names change; ids never do."""
from __future__ import annotations

import re

from slugify import slugify as _slugify
from ulid import ULID

PREFIXES: dict[str, str] = {
    "model": "model",
    "model_family": "family",       # Llama 4, Qwen3.6, Claude — groups model releases
    "artifact": "artifact",         # checkpoint / quantisation / conversion of a model (never a model of its own)
    "company": "company",
    "organization": "org",
    "researcher": "person",
    "paper": "paper",
    "dataset": "dataset",
    "benchmark": "bench",
    "provider": "provider",
    "framework": "framework",
    "library": "lib",
    "repository": "repo",
    "tool": "tool",
    "agent": "agent",
    "application": "app",
    "hardware": "hw",
    "runtime": "runtime",
    "quantization": "quant",
    "license": "license",
    "conference": "conf",
    "university": "univ",
    "lab": "lab",
    "product": "product",
    "release": "release",
    "regulation": "reg",
    "incident": "incident",
    "country": "country",
    "mcp_server": "mcp",
    "robot": "robot",
    "data_center": "dc",
    "job": "job",
    "course": "course",
    "standard": "std",
    "funding_round": "funding",
    "acquisition": "acq",
    # infrastructure records
    "source": "src",
    "document": "doc",
    "snapshot": "snap",
    "claim": "claim",
    "relation": "rel",
    "change_event": "evt",
    "connector_run": "run",
    "queue_job": "qj",
    "llm_job": "llm",
    "review": "rev",
    "price": "price",
    "result": "res",
    "api_key": "key",
}

ENTITY_TYPES: tuple[str, ...] = tuple(k for k in PREFIXES if k not in {
    "source", "document", "snapshot", "claim", "relation", "change_event", "connector_run", "queue_job", "llm_job",
    "review", "price", "result", "api_key"})


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 = 96) -> str:
    """URL slug that keeps dots and plus signs (model names like `qwen3-8b`, `gpt-4.1`, `c++`)."""
    s = _slugify(text, lowercase=True, regex_pattern=r"[^a-z0-9.+-]+", replacements=[("/", "-"), ("_", "-"), ("@", "-at-")])
    s = _slug_clean.sub("-", s).strip("-.")
    return s[:max_length].rstrip("-.") or "item"


def normalize_alias(text: str) -> str:
    """Deterministic alias key: lowercase, ASCII, punctuation collapsed. Used for entity resolution — never for display."""
    s = _slugify(text, lowercase=True, regex_pattern=r"[^a-z0-9]+")
    return s.replace("-", "")