qtalker---/backend/tests/test_ws_reward_replay.py

"""WS integration tests for the OTHER two reward faucets in app.ws — the
ritual milestone and the judgment verdict.

Same class of bug, and the same reason it survived a 400-test suite:
`tests/test_judgment.py` covers app/judgment.py, which is pure — it decides
outcomes and returns them. It cannot see the handler that banks them.
`tests/test_ws_ritual_judgment.py` drives the handlers over a real socket but
only ever performs each rite ONCE, so it could not see what happens on the
second pass.

The faucets:

  * `ritual_start` rewinds `ritual_completed` to False at any time, and the UI
    offers exactly that button (RitualPanel's "attempt again", which honest
    play genuinely needs after a failed roll). Every re-walk of
    ritual_start + 4x ritual_step re-credited RITUAL_SUCCESS_ESSENCE and
    re-rolled an item.
  * `judgment` had no once-per-presence guard at all except for `cross_over`.
    Replaying {"verdict": "trust"} against a benevolent spirit paid essence
    AND favor AND an item roll on every single frame — favor pinning to +1.0,
    which then biases every future mint.

Both are bounded in RATE by their limiters and unbounded in TOTAL, off a
single summon. These tests are written against the exploit path, not the
implementation: they send the frames a browser sends and then read the ledger
straight out of the database.
"""

import uuid

import pytest
from sqlalchemy import select

import app.judgment as judgment_module
import app.ws
from app.entities import fallback_profile
from app.inventory import (
    CORRECT_JUDGMENT_ESSENCE,
    RITUAL_SUCCESS_ESSENCE,
    SUMMON_ESSENCE_TRICKLE,
)
from app.models.inventory_item import InventoryItem
from app.models.user import User
from app.rate_limit import RateLimiter


class FakeSpiritService:
    async def mint_profile(
        self, signature, channel, anomalies, language="en", entropy=None, sky=None
    ):
        return fallback_profile(signature)

    async def fragment(self, source, anomaly, language="en"):
        return "listen"

    async def wire_whisper(self, telemetry, language="en"):
        return "the wire hums"

    def chat_stream(self, entity, question, history, language="en"):
        async def gen():
            yield "here."

        return gen()

    def ambient_ready(self):
        return False


async def _fake_synth(text, voice, profile, instability=0.0):
    return b"RIFFfake wav bytes"


@pytest.fixture(autouse=True)
def _fake_spirits(monkeypatch):
    monkeypatch.setattr(app.ws, "spirit_service", FakeSpiritService())
    monkeypatch.setattr(app.ws, "synthesize_spirit_voice", _fake_synth)
    # Module-level limiters are shared singletons accumulating real hit counts
    # across the whole session (precedent: test_ws_ritual_judgment.py). The
    # ritual/judgment ones matter most here: the exploits replayed below are
    # deliberately high-volume, and the real 6/60s and 10/60s caps would mask
    # the faucet behind a rate limit instead of letting us observe the total.
    for name in (
        "summon_limiter",
        "summon_ip_limiter",
        "question_limiter",
        "question_ip_limiter",
        "fragment_limiter",
        "fragment_ip_limiter",
        "ritual_limiter",
        "ritual_ip_limiter",
        "judgment_limiter",
        "judgment_ip_limiter",
        "passage_limiter",
        "passage_ip_limiter",
    ):
        monkeypatch.setattr(app.ws, name, RateLimiter(max_requests=10_000, window_seconds=60))


REPLAYS = 10


def _read_until(ws, msg_type, max_frames=80, **match):
    for _ in range(max_frames):
        frame = ws.receive_json()
        if frame.get("type") != msg_type:
            continue
        if all(frame.get(key) == value for key, value in match.items()):
            return frame
    raise AssertionError(f"never saw frame of type {msg_type!r} matching {match!r}")


def _login(sync_client, username):
    sync_client.post("/auth/register", json={"username": username, "password": "spookyspooky"})
    sync_client.post("/auth/login", json={"username": username, "password": "spookyspooky"})
    return sync_client.cookies.get("qm_session")


def _user_id(sync_client, token):
    return uuid.UUID(
        sync_client.get("/auth/me", headers={"cookie": f"qm_session={token}"}).json()["id"]
    )


def _ws_connect(sync_client, token):
    return sync_client.websocket_connect(
        "/ws/session", headers={"cookie": f"qm_session={token}"}
    )


def _summon(ws):
    ws.send_json({"type": "summon"})
    frame = _read_until(ws, "entity")
    _read_until(ws, "utterance", kind="greeting")
    return frame


def _settle(ws):
    """Force a full round trip so any post-frame DB write has landed.

    Rewards are credited before/around the frame being queued, but the sender
    task drains that queue concurrently with the handler's remaining awaits.
    The connection's message loop is sequential, so a pong is a hard guarantee
    that the previous handler returned — not a poll-and-hope.
    """
    ws.send_json({"type": "ping"})
    _read_until(ws, "pong")


def _run_ritual(ws, steps=4):
    """One full attempt: the rewind button, then every step."""
    ws.send_json({"type": "ritual_start"})
    for i in range(1, steps + 1):
        ws.send_json({"type": "ritual_step", "step": i})
    result = _read_until(ws, "ritual_complete")
    _settle(ws)
    return result


def _judge(ws, verdict):
    ws.send_json({"type": "judgment", "verdict": verdict})
    result = _read_until(ws, "judgment_result")
    _settle(ws)
    return result


async def _ledger(db_session, user_id):
    db_session.expire_all()
    user = await db_session.get(User, user_id)
    items = (
        await db_session.execute(
            select(InventoryItem).where(InventoryItem.user_id == user_id)
        )
    ).scalars().all()
    return user.essence, user.favor, len(items)


def _always_wins(monkeypatch):
    monkeypatch.setattr(
        app.ws.judgment, "roll_ritual_success", lambda traits, rng=None: True
    )


def _verdict_outcomes(monkeypatch, table):
    """Pin `judge_verdict` to a fixed verdict -> JudgmentOutcome table.

    The real roll depends on the minted entity's hidden traits; these tests
    are about the ledger, not about which spirit answered, so the outcome is
    made deterministic exactly the way test_ws_ritual_judgment.py does it.
    """
    monkeypatch.setattr(
        app.ws.judgment, "judge_verdict", lambda verdict, traits, **kw: table[verdict]
    )


# --- the ritual faucet -----------------------------------------------------


@pytest.mark.asyncio
async def test_replaying_the_ritual_never_mints_essence_twice(
    sync_client, db_session, monkeypatch
):
    """The exploit, executed literally.

    Complete the rite, hit `ritual_start` to rewind, complete it again — ten
    times over. The seeker must finish with exactly one milestone's worth of
    essence, because there was only ever one spirit.

    Before the fix this credited RITUAL_SUCCESS_ESSENCE on every pass: eleven
    completions paid 165 instead of 15.
    """
    _always_wins(monkeypatch)

    token = _login(sync_client, "ritual-faucet")
    user_id = _user_id(sync_client, token)

    with _ws_connect(sync_client, token) as ws:
        _read_until(ws, "session")
        _summon(ws)
        baseline, _, _ = await _ledger(db_session, user_id)

        first = _run_ritual(ws)
        after_one, _, items_after_one = await _ledger(db_session, user_id)

        for _ in range(REPLAYS):
            result = _run_ritual(ws)
            assert result["success"] is True, "the replay did not even complete"

        after_replays, _, items_after_replays = await _ledger(db_session, user_id)

    assert first["success"] is True
    earned = after_one - baseline
    assert earned == RITUAL_SUCCESS_ESSENCE, "the rite paid nothing — the test proves nothing"
    assert after_replays == after_one, (
        f"replaying the ritual minted {after_replays - after_one} extra essence "
        f"across {REPLAYS} rewinds; the ritual is an unbounded faucet again"
    )
    assert items_after_replays == items_after_one, (
        f"replaying the ritual rolled {items_after_replays - items_after_one} extra "
        "item drops off a single summon"
    )


@pytest.mark.asyncio
async def test_replaying_the_ritual_still_reveals_the_traits(
    sync_client, db_session, monkeypatch
):
    """The guard must bite the ledger only.

    Re-attempting is a real, UI-offered action (it is how honest play recovers
    from a failed roll), so a repeat must still roll and still reveal the true
    traits on success — it just must not pay again. A guard that silently
    swallowed the attempt would be a worse bug than the faucet.
    """
    _always_wins(monkeypatch)

    token = _login(sync_client, "ritual-rewalker")
    _user_id(sync_client, token)

    with _ws_connect(sync_client, token) as ws:
        _read_until(ws, "session")
        _summon(ws)
        first = _run_ritual(ws)
        second = _run_ritual(ws)

    assert second["success"] is True
    assert second["revealed"] == first["revealed"] is not None


@pytest.mark.asyncio
async def test_a_genuinely_new_presence_reopens_the_ritual_purse(
    sync_client, db_session, monkeypatch
):
    """The guard must not overshoot: each spirit is worth its own rite.

    A `ritual_paid` latch that survived a summon would silently make every
    spirit after the first worthless.
    """
    _always_wins(monkeypatch)

    token = _login(sync_client, "ritual-second-spirit")
    user_id = _user_id(sync_client, token)

    with _ws_connect(sync_client, token) as ws:
        _read_until(ws, "session")
        _summon(ws)
        baseline, _, _ = await _ledger(db_session, user_id)
        _run_ritual(ws)
        after_first, _, _ = await _ledger(db_session, user_id)

        _summon(ws)
        _run_ritual(ws)
        after_second, _, _ = await _ledger(db_session, user_id)

    first_rite = after_first - baseline
    # The second summon also pays its own trickle, so compare the rite only.
    second_rite = after_second - after_first - SUMMON_ESSENCE_TRICKLE
    assert first_rite == RITUAL_SUCCESS_ESSENCE
    assert second_rite == first_rite, (
        "a fresh presence did not re-open the purse — every spirit after the "
        "first has a worthless ritual"
    )


# --- the judgment faucet ---------------------------------------------------


_CORRECT_TRUST = judgment_module.JudgmentOutcome(
    True, judgment_module.FAVOR_CORRECT_TRUST, CORRECT_JUDGMENT_ESSENCE, False, "reward"
)
_WRONG_TRUST = judgment_module.JudgmentOutcome(
    False, judgment_module.FAVOR_WRONG_TRUST, 0, False, "escalation"
)
_CORRECT_BANISH = judgment_module.JudgmentOutcome(
    True, judgment_module.FAVOR_CORRECT_BANISH, CORRECT_JUDGMENT_ESSENCE, False, "reward"
)


@pytest.mark.asyncio
async def test_replaying_a_correct_judgment_never_pays_twice(
    sync_client, db_session, monkeypatch
):
    """One spirit, one verdict, one payment.

    Before the fix, eleven `trust` frames against one benevolent spirit paid
    11 x CORRECT_JUDGMENT_ESSENCE (132 instead of 12) and 11 x +0.05 favor
    (0.55, clamped toward the +1.0 ceiling instead of 0.05), plus eleven item
    rolls — all off a single summon.
    """
    _verdict_outcomes(monkeypatch, {"trust": _CORRECT_TRUST})

    token = _login(sync_client, "judgment-faucet")
    user_id = _user_id(sync_client, token)

    with _ws_connect(sync_client, token) as ws:
        _read_until(ws, "session")
        _summon(ws)
        baseline_essence, baseline_favor, _ = await _ledger(db_session, user_id)

        _judge(ws, "trust")
        essence_one, favor_one, items_one = await _ledger(db_session, user_id)

        for _ in range(REPLAYS):
            _judge(ws, "trust")

        essence_end, favor_end, items_end = await _ledger(db_session, user_id)

    assert essence_one - baseline_essence == CORRECT_JUDGMENT_ESSENCE
    assert favor_one - baseline_favor == pytest.approx(judgment_module.FAVOR_CORRECT_TRUST)
    assert essence_end == essence_one, (
        f"replaying one verdict minted {essence_end - essence_one} extra essence "
        f"across {REPLAYS} frames; judgment is an unbounded faucet again"
    )
    assert favor_end == pytest.approx(favor_one), (
        f"replaying one verdict moved favor by a further {favor_end - favor_one}; "
        "a scripted client can pin favor at the +1.0 ceiling"
    )
    assert items_end == items_one, (
        f"replaying one verdict rolled {items_end - items_one} extra item drops"
    )


@pytest.mark.asyncio
async def test_a_replayed_judgment_reports_zero_deltas(
    sync_client, db_session, monkeypatch
):
    """The client sums `essence_delta`/`favor_delta` into its displayed
    totals (frontend/src/state/seance.tsx's `judgment_result` case). A frame
    that advertises the verdict's face value while the ledger credits nothing
    shows the seeker currency their account never received — the most
    corrosive kind of bug in a game about trust.
    """
    _verdict_outcomes(monkeypatch, {"trust": _CORRECT_TRUST})

    token = _login(sync_client, "judgment-honest-frame")
    user_id = _user_id(sync_client, token)

    with _ws_connect(sync_client, token) as ws:
        _read_until(ws, "session")
        _summon(ws)
        baseline_essence, baseline_favor, _ = await _ledger(db_session, user_id)

        frames = [_judge(ws, "trust") for _ in range(3)]
        essence_end, favor_end, _ = await _ledger(db_session, user_id)

    assert sum(f["essence_delta"] for f in frames) == essence_end - baseline_essence
    assert sum(f["favor_delta"] for f in frames) == pytest.approx(favor_end - baseline_favor)
    for frame in frames[1:]:
        assert frame["essence_delta"] == 0 and frame["favor_delta"] == 0, (
            "a replayed verdict advertised essence/favor it did not pay"
        )


@pytest.mark.asyncio
async def test_an_honest_correction_still_resolves(sync_client, db_session, monkeypatch):
    """The guard must not overshoot: a seeker who calls it wrong and then
    calls it right is doing something real, not replaying. A blanket
    "judged once" latch would make that second, different verdict a no-op.
    """
    _verdict_outcomes(monkeypatch, {"trust": _WRONG_TRUST, "banish": _CORRECT_BANISH})

    token = _login(sync_client, "judgment-corrector")
    user_id = _user_id(sync_client, token)

    with _ws_connect(sync_client, token) as ws:
        _read_until(ws, "session")
        _summon(ws)
        baseline_essence, baseline_favor, _ = await _ledger(db_session, user_id)

        wrong = _judge(ws, "trust")
        right = _judge(ws, "banish")
        essence_end, favor_end, _ = await _ledger(db_session, user_id)

    assert wrong["consequence"] == "escalation"
    assert right["consequence"] == "reward"
    assert essence_end - baseline_essence == CORRECT_JUDGMENT_ESSENCE, (
        "the corrected verdict paid nothing — the replay guard is too broad"
    )
    assert favor_end - baseline_favor == pytest.approx(
        judgment_module.FAVOR_WRONG_TRUST + judgment_module.FAVOR_CORRECT_BANISH
    )


@pytest.mark.asyncio
async def test_a_genuinely_new_presence_reopens_the_judgment_purse(
    sync_client, db_session, monkeypatch
):
    _verdict_outcomes(monkeypatch, {"trust": _CORRECT_TRUST})

    token = _login(sync_client, "judgment-second-spirit")
    user_id = _user_id(sync_client, token)

    with _ws_connect(sync_client, token) as ws:
        _read_until(ws, "session")
        _summon(ws)
        _judge(ws, "trust")
        after_first, _, _ = await _ledger(db_session, user_id)

        _summon(ws)
        _judge(ws, "trust")
        after_second, _, _ = await _ledger(db_session, user_id)

    second_verdict = after_second - after_first - SUMMON_ESSENCE_TRICKLE
    assert second_verdict == CORRECT_JUDGMENT_ESSENCE, (
        "a fresh presence did not re-open the purse — every spirit after the "
        "first is unjudgeable for reward"
    )


@pytest.mark.asyncio
async def test_the_test_verdict_stays_freely_repeatable(sync_client, db_session, monkeypatch):
    """`test` is a diagnostic pulse that never touches the ledger, so it is
    deliberately exempt from the once-per-presence guard — the panel lets a
    seeker pulse a spirit as often as they like.
    """
    token = _login(sync_client, "judgment-tester")
    user_id = _user_id(sync_client, token)

    with _ws_connect(sync_client, token) as ws:
        _read_until(ws, "session")
        _summon(ws)
        baseline, baseline_favor, _ = await _ledger(db_session, user_id)
        results = [_judge(ws, "test") for _ in range(3)]
        essence_end, favor_end, _ = await _ledger(db_session, user_id)

    assert all(r["consequence"] == "neutral" for r in results)
    assert essence_end == baseline
    assert favor_end == baseline_favor