app/problems.py

"""The judge endpoint's problem-id whitelist.

``POST /v1/judge`` must reject an unknown ``problem_id`` before anything is
queued (ERA2_DESIGN.md: "rejects unknown problem_id — must be in public-200 ∪
canonical set"). public-200 is a subset of the canonical set — the eval Space
enforces that invariant on its side (``evaluator._public_200_ids_ok``) — so the
union IS the canonical 1,669 and this module is the whole answer.

The set ships as data (``app/data/problem_ids.json``, regenerated by
``scripts/gen_problem_ids.py`` from the judge repo's four canonical JSONLs)
rather than as a regex, because a regex is wrong: ``hard1_0070`` matches
``^(normal|hard[123])_\\d{4}$`` and does not exist. It carries IDS ONLY — no
statements, no answers — since era 2 deliberately removes the oracle surface
from the service path and this app is the public face of the collab.

Loaded once per process (per path) behind ``lru_cache``: ~1,669 short strings,
read at first use rather than at import, so a broken data file surfaces in the
one route that needs it instead of as a Space that will not boot. The path is
threaded in from ``Settings.problem_ids_path`` by the caller rather than read
from the global settings here, so tests can point at a fixture file the same way
they override every other dependency.
"""
from __future__ import annotations

import json
from functools import lru_cache
from pathlib import Path


DEFAULT_PATH = Path(__file__).resolve().parent / "data" / "problem_ids.json"


@lru_cache
def problem_ids(path: str | None = None) -> frozenset[str]:
    """Every judgeable problem id. Raises if the data file is missing or does
    not parse — an id set we cannot read must never degrade to "accept
    everything", and that is a deploy bug, not something the caller can fix."""
    doc = json.loads((Path(path) if path else DEFAULT_PATH).read_text())
    ids = frozenset(doc["ids"])
    count = doc.get("count")
    if isinstance(count, int) and count != len(ids):
        raise ValueError(
            f"{path or DEFAULT_PATH}: declares count={count} "
            f"but carries {len(ids)} unique ids"
        )
    if not ids:
        raise ValueError(f"{path or DEFAULT_PATH}: empty problem-id set")
    return ids


def is_known_problem(problem_id: str, path: str | None = None) -> bool:
    return problem_id in problem_ids(path)