Files
signal-platform/tests/unit/test_prod_strategy_parity.py
T
dennisthiessen b0e33e1606 fix: align production defaults and close review parity gaps
Ship greenfield min_rr=2.0 and conf=0, read-only Structural S/R, indicator
cache invalidation, and UI/gate language that treats GTL as screening not exit.
Align strategy_rank missing-vol fallback live vs backtest, single-source
PRIMARY_TARGET_MIN_RR, expand prod parity tests, and drop dead FE clients.
2026-07-18 13:03:22 +02:00

256 lines
10 KiB
Python

"""Parity guards: the backtest's production strategy must equal the live setup.
The portfolio monitor's production row replays the live qualification flag and
the runtime Admin exit policy, but several constants are still defined on both
sides (defaults, trail width, ordering weights). These tests fail if the two
sides drift, so a change to the live strategy forces the backtest — and vice
versa — to move with it.
"""
import pytest
from app.services import paper_trade_service
from app.services.admin_service import ACTIVATION_DEFAULTS
from app.services.backtest_service import (
ATR_MULTIPLIER,
ATR_TRAIL_MULTIPLIER,
LIVE_EXIT_MODE_TO_SIM,
PORTFOLIO_MONITOR_STRATEGIES,
PRODUCTION_PERCENTILE_KEY,
RESIDUAL_HIGH_VOL_BLEND_80_20_KEY,
TIME_EXIT_DAYS,
_entry_variant_config,
_momentum_qualifies,
_qualifies_strategy_variant,
)
from app.services.momentum_service import (
STRATEGY_RANK_MOMENTUM_WEIGHT,
STRATEGY_RANK_VOL_WEIGHT,
blend_strategy_rank,
)
from app.services.qualification import MIN_TARGET_PROBABILITY
from app.services.recommendation_service import PRIMARY_TARGET_MIN_RR
from app.services import rr_scanner_service
def _production_monitor_row() -> dict:
return next(s for s in PORTFOLIO_MONITOR_STRATEGIES if s.get("is_production"))
def test_exit_defaults_match_the_simulated_exit() -> None:
assert paper_trade_service.DEFAULT_EXIT_MODE == "atr_trailing"
assert LIVE_EXIT_MODE_TO_SIM[paper_trade_service.DEFAULT_EXIT_MODE] == "atr_trail3"
assert paper_trade_service.DEFAULT_ATR_MULTIPLIER == ATR_TRAIL_MULTIPLIER
assert paper_trade_service.DEFAULT_HOLD_DAYS == max(TIME_EXIT_DAYS)
def test_every_live_exit_mode_has_a_sim_mapping() -> None:
assert set(paper_trade_service._VALID_EXIT_MODES) == set(LIVE_EXIT_MODE_TO_SIM)
def test_setup_stop_width_matches_the_frontend_constant() -> None:
"""The UI recovers ATR from a setup as |entry - stop| / 1.5 to render the real
exit plan (frontend/src/lib/exitPlan.ts: SETUP_STOP_ATR_MULTIPLIER). Nothing
else transmits ATR, so if the scanner's stop width changes here the UI would
silently draw the trailing stop in the wrong place."""
import inspect
from app.services import rr_scanner_service
frontend_constant = 1.5
assert ATR_MULTIPLIER == frontend_constant
for fn in (rr_scanner_service.scan_ticker, rr_scanner_service.scan_all_tickers):
signature = inspect.signature(fn)
assert signature.parameters["atr_multiplier"].default == frontend_constant
def test_gate_default_matches_the_promoted_cutoff() -> None:
prod = _production_monitor_row()
entry_cfg = _entry_variant_config(str(prod["entry_variant"]))
assert entry_cfg is not None
assert float(entry_cfg["cutoff"]) == float(ACTIVATION_DEFAULTS["min_momentum_percentile"])
def test_production_ordering_weights_are_single_sourced() -> None:
# The promoted ordering is 80/20 momentum/vol; the backtest imports the
# weight, so equality here pins the *value* the promotion was validated at.
assert STRATEGY_RANK_MOMENTUM_WEIGHT == 0.8
assert STRATEGY_RANK_VOL_WEIGHT == pytest.approx(0.2)
prod = _production_monitor_row()
entry_cfg = _entry_variant_config(str(prod["entry_variant"]))
assert entry_cfg is not None
assert entry_cfg["ranking_key"] == RESIDUAL_HIGH_VOL_BLEND_80_20_KEY
def test_production_monitor_row_replays_the_live_config() -> None:
prod = _production_monitor_row()
assert prod.get("use_live_config") is True
assert prod["exit_policy"] == "atr_trail3"
def test_live_gate_equals_the_production_variant_gate() -> None:
"""The monitor's live-gate switch relies on the runtime `qualified` flag
(_momentum_qualifies) selecting exactly what the frozen production variant
gate selects at the default cutoff."""
prod = _production_monitor_row()
entry_cfg = _entry_variant_config(str(prod["entry_variant"]))
assert entry_cfg is not None
cutoff = float(ACTIVATION_DEFAULTS["min_momentum_percentile"])
for cand in (
{"meets_core": True, "direction": "long", PRODUCTION_PERCENTILE_KEY: 92.0},
{"meets_core": True, "direction": "long", PRODUCTION_PERCENTILE_KEY: 80.0},
{"meets_core": True, "direction": "long", PRODUCTION_PERCENTILE_KEY: 79.9},
{"meets_core": True, "direction": "long", PRODUCTION_PERCENTILE_KEY: None},
{"meets_core": True, "direction": "short", PRODUCTION_PERCENTILE_KEY: 95.0},
{"meets_core": False, "direction": "long", PRODUCTION_PERCENTILE_KEY: 95.0},
):
assert _momentum_qualifies(cand, cutoff) == _qualifies_strategy_variant(
cand, entry_cfg
), cand
def test_activation_defaults_match_promoted_production_gate() -> None:
"""Greenfield Admin must ship the researched gate, not the old trough defaults."""
assert float(ACTIVATION_DEFAULTS["min_rr"]) == 2.0
assert float(ACTIVATION_DEFAULTS["min_confidence"]) == 0.0
assert float(ACTIVATION_DEFAULTS["min_momentum_percentile"]) == 80.0
assert ACTIVATION_DEFAULTS["exclude_neutral"] is True
def test_primary_target_rr_floor_is_single_sourced() -> None:
assert rr_scanner_service.PRIMARY_TARGET_MIN_RR == PRIMARY_TARGET_MIN_RR
assert PRIMARY_TARGET_MIN_RR == 1.5
assert MIN_TARGET_PROBABILITY == 20.0
def test_strategy_rank_falls_back_to_momentum_when_vol_missing() -> None:
"""Live and backtest must not bury a name solely because vol history is short."""
assert blend_strategy_rank(80.0, 60.0) == 76.0
assert blend_strategy_rank(80.0, None) == 80.0
assert blend_strategy_rank(None, 60.0) is None
assert blend_strategy_rank(None, None) is None
from app.services import backtest_service as bt
cands = [
{bt.PRODUCTION_PERCENTILE_KEY: 80.0, bt.VOL_PERCENTILE_KEY: None},
{bt.PRODUCTION_PERCENTILE_KEY: 70.0, bt.VOL_PERCENTILE_KEY: 50.0},
]
bt._assign_residual_high_vol_blend(cands)
assert cands[0][bt.RESIDUAL_HIGH_VOL_BLEND_80_20_KEY] == 80.0
assert cands[1][bt.RESIDUAL_HIGH_VOL_BLEND_80_20_KEY] == 66.0
@pytest.mark.asyncio
async def test_live_scan_and_backtest_window_share_gtl_primary() -> None:
"""Same OHLCV + dims: live scan_ticker primary ≡ backtest _window_setups.
No gate_levels_override — both paths build the production GTL from bars.
Dimension scores are seeded to the values the backtest window computes so
probability ranking cannot diverge for that reason alone.
"""
from datetime import date, datetime, timedelta, timezone
from app.models.ohlcv import OHLCVRecord
from app.models.score import DimensionScore
from app.models.ticker import Ticker
from app.services import backtest_service as bt
from app.services.recommendation_service import DEFAULT_RECOMMENDATION_CONFIG
from app.services.rr_scanner_service import scan_ticker
from app.services.scoring_service import (
compute_momentum_from_closes,
compute_technical_from_arrays,
)
from tests.conftest import _test_session_factory
n = 120
base = date(2024, 1, 1)
# Oscillating range so GTL finds traffic-backed proposals above/below spot.
closes: list[float] = []
highs: list[float] = []
lows: list[float] = []
volumes: list[int] = []
price = 100.0
for i in range(n):
phase = i % 30
if phase < 12:
price = price + (94.0 - price) * 0.2
elif phase < 24:
price = price + (108.0 - price) * 0.2
else:
price = 100.0 + (i % 5) * 0.3
high = price + 1.2
low = price - 1.2
close = price
closes.append(close)
highs.append(high)
lows.append(low)
volumes.append(100_000 + i * 10)
tech = (compute_technical_from_arrays(highs, lows, closes, volumes)[0]) or 50.0
mom = (compute_momentum_from_closes(closes)[0]) or 50.0
async with _test_session_factory() as session:
ticker = Ticker(symbol="GTLPAR")
session.add(ticker)
await session.flush()
bars = [
OHLCVRecord(
ticker_id=ticker.id,
date=base + timedelta(days=i),
open=closes[i] - 0.2,
high=highs[i],
low=lows[i],
close=closes[i],
volume=volumes[i],
)
for i in range(n)
]
session.add_all(bars)
now = datetime.now(timezone.utc)
session.add_all([
DimensionScore(
ticker_id=ticker.id, dimension="technical", score=float(tech),
is_stale=False, computed_at=now,
),
DimensionScore(
ticker_id=ticker.id, dimension="momentum", score=float(mom),
is_stale=False, computed_at=now,
),
])
await session.commit()
live = await scan_ticker(session, "GTLPAR", rr_threshold=1.5, atr_multiplier=1.5)
# Re-load bars as plain ORM list for the pure backtest window path.
from sqlalchemy import select
records = list(
(
await session.execute(
select(OHLCVRecord)
.where(OHLCVRecord.ticker_id == ticker.id)
.order_by(OHLCVRecord.date.asc())
)
).scalars().all()
)
config = dict(DEFAULT_RECOMMENDATION_CONFIG)
activation = dict(ACTIVATION_DEFAULTS)
sim = bt._window_setups(records, config, activation)
live_by_dir = {s.direction: s for s in live}
sim_by_dir = {s["direction"]: s for s in sim}
assert set(live_by_dir) == set(sim_by_dir), (
f"direction mismatch live={set(live_by_dir)} sim={set(sim_by_dir)}"
)
assert live_by_dir, "expected at least one directional setup from GTL"
for direction, live_setup in live_by_dir.items():
sim_setup = sim_by_dir[direction]
assert live_setup.target == pytest.approx(float(sim_setup["target"]), abs=0.05), (
f"{direction}: live target {live_setup.target} != sim {sim_setup['target']}"
)
assert live_setup.rr_ratio == pytest.approx(float(sim_setup["rr"]), abs=0.05), (
f"{direction}: live rr {live_setup.rr_ratio} != sim {sim_setup['rr']}"
)