feat: autonomous trading engine — full implementation
- Database migration 018 with 13 tables for trading engine state - Trading engine service (services/trading/) with 12 pure computation modules: position sizer, stop-loss manager, reserve pool, circuit breaker, risk tier controller, correlation matrix, tax lots, trading window, gradual entry, notifications, micro-trading, backtester - Core TradingEngine with pre-trade evaluation pipeline and integration wiring - FastAPI HTTP service with 14 endpoints (health, config, decisions, metrics, backtest) - Performance tracker with Sharpe ratio, drawdown, profit factor computation - 194 Python tests (165 property-based + 29 integration) - Frontend: 13 TanStack Query hooks, 7 dashboard panels, tabbed Trading Engine page - Helm chart entry, network policy, nginx proxy, ingress for trading-engine - Shared infrastructure: enums, Redis keys, TradingConfig in AppConfig
This commit is contained in:
Celes Renata
@@ -0,0 +1,317 @@
|
||||
"""Property-based tests for Trading Window and Gradual Entry.
|
||||
|
||||
Feature: autonomous-trading-engine
|
||||
|
||||
Tests properties 20 and 21 from the design specification, covering
|
||||
trading window determination (9:45 AM – 3:45 PM ET on weekdays) and
|
||||
gradual entry tranche splitting.
|
||||
"""
|
||||
|
||||
from __future__ import annotations
|
||||
|
||||
from datetime import datetime, time
|
||||
|
||||
from hypothesis import given, settings
|
||||
from hypothesis import strategies as st
|
||||
|
||||
from services.trading.gradual_entry import (
|
||||
create_tranches,
|
||||
should_use_gradual_entry,
|
||||
split_into_tranches,
|
||||
)
|
||||
from services.trading.trading_window import (
|
||||
ET,
|
||||
MARKET_OPEN,
|
||||
WINDOW_CLOSE,
|
||||
WINDOW_OPEN,
|
||||
is_market_open,
|
||||
is_within_trading_window,
|
||||
next_window_open,
|
||||
)
|
||||
|
||||
# ---------------------------------------------------------------------------
|
||||
# Hypothesis strategies
|
||||
# ---------------------------------------------------------------------------
|
||||
|
||||
_WEEKDAYS = range(0, 5) # Mon–Fri
|
||||
|
||||
|
||||
def _et_datetime_in_window() -> st.SearchStrategy[datetime]:
|
||||
"""Generate a timezone-aware datetime that is inside the trading window."""
|
||||
return (
|
||||
st.dates(
|
||||
min_value=datetime(2024, 1, 1).date(),
|
||||
max_value=datetime(2025, 12, 31).date(),
|
||||
)
|
||||
.filter(lambda d: d.weekday() in _WEEKDAYS)
|
||||
.flatmap(
|
||||
lambda d: st.times(
|
||||
min_value=WINDOW_OPEN,
|
||||
max_value=time(15, 44, 59),
|
||||
).map(lambda t: datetime.combine(d, t, tzinfo=ET))
|
||||
)
|
||||
)
|
||||
|
||||
|
||||
def _et_datetime_outside_window_weekday() -> st.SearchStrategy[datetime]:
|
||||
"""Generate a weekday datetime that is outside the trading window.
|
||||
|
||||
Either before 9:45 AM ET or at/after 3:45 PM ET.
|
||||
"""
|
||||
before_open = st.times(min_value=time(0, 0), max_value=time(9, 44, 59))
|
||||
after_close = st.times(min_value=WINDOW_CLOSE, max_value=time(23, 59, 59))
|
||||
|
||||
return (
|
||||
st.dates(
|
||||
min_value=datetime(2024, 1, 1).date(),
|
||||
max_value=datetime(2025, 12, 31).date(),
|
||||
)
|
||||
.filter(lambda d: d.weekday() in _WEEKDAYS)
|
||||
.flatmap(
|
||||
lambda d: st.one_of(before_open, after_close).map(
|
||||
lambda t: datetime.combine(d, t, tzinfo=ET)
|
||||
)
|
||||
)
|
||||
)
|
||||
|
||||
|
||||
def _et_datetime_weekend() -> st.SearchStrategy[datetime]:
|
||||
"""Generate a weekend datetime (Saturday or Sunday)."""
|
||||
return (
|
||||
st.dates(
|
||||
min_value=datetime(2024, 1, 1).date(),
|
||||
max_value=datetime(2025, 12, 31).date(),
|
||||
)
|
||||
.filter(lambda d: d.weekday() >= 5)
|
||||
.flatmap(
|
||||
lambda d: st.times().map(
|
||||
lambda t: datetime.combine(d, t, tzinfo=ET)
|
||||
)
|
||||
)
|
||||
)
|
||||
|
||||
|
||||
# ---------------------------------------------------------------------------
|
||||
# Property 20: Trading window determination
|
||||
# **Validates: Requirements 11.1**
|
||||
# ---------------------------------------------------------------------------
|
||||
|
||||
|
||||
class TestProperty20TradingWindowDetermination:
|
||||
"""Property 20: Trading window determination.
|
||||
|
||||
**Validates: Requirements 11.1**
|
||||
"""
|
||||
|
||||
@settings(max_examples=100)
|
||||
@given(dt=_et_datetime_in_window())
|
||||
def test_within_window_on_weekday(self, dt: datetime) -> None:
|
||||
"""Timestamps between 9:45 AM and 3:45 PM ET on weekdays are within the window."""
|
||||
assert is_within_trading_window(dt) is True, (
|
||||
f"Expected within window: {dt} (ET weekday={dt.weekday()}, time={dt.time()})"
|
||||
)
|
||||
|
||||
@settings(max_examples=100)
|
||||
@given(dt=_et_datetime_outside_window_weekday())
|
||||
def test_outside_window_on_weekday(self, dt: datetime) -> None:
|
||||
"""Weekday timestamps before 9:45 AM or at/after 3:45 PM ET are outside the window."""
|
||||
assert is_within_trading_window(dt) is False, (
|
||||
f"Expected outside window: {dt} (ET time={dt.time()})"
|
||||
)
|
||||
|
||||
@settings(max_examples=100)
|
||||
@given(dt=_et_datetime_weekend())
|
||||
def test_outside_window_on_weekend(self, dt: datetime) -> None:
|
||||
"""Weekend timestamps are always outside the trading window."""
|
||||
assert is_within_trading_window(dt) is False, (
|
||||
f"Expected outside window on weekend: {dt} (weekday={dt.weekday()})"
|
||||
)
|
||||
|
||||
@settings(max_examples=100)
|
||||
@given(dt=_et_datetime_outside_window_weekday())
|
||||
def test_next_window_open_is_in_future(self, dt: datetime) -> None:
|
||||
"""next_window_open always returns a time >= the input when outside the window."""
|
||||
nwo = next_window_open(dt)
|
||||
et_dt = dt.astimezone(ET)
|
||||
et_nwo = nwo.astimezone(ET)
|
||||
# If we're past today's open, next open must be a future day
|
||||
if et_dt.time() >= WINDOW_OPEN:
|
||||
assert et_nwo.date() > et_dt.date(), (
|
||||
f"Expected future date: nwo={et_nwo}, dt={et_dt}"
|
||||
)
|
||||
assert et_nwo.time().hour == WINDOW_OPEN.hour
|
||||
assert et_nwo.time().minute == WINDOW_OPEN.minute
|
||||
|
||||
@settings(max_examples=100)
|
||||
@given(dt=_et_datetime_weekend())
|
||||
def test_next_window_open_skips_weekends(self, dt: datetime) -> None:
|
||||
"""next_window_open from a weekend returns a weekday."""
|
||||
nwo = next_window_open(dt)
|
||||
et_nwo = nwo.astimezone(ET)
|
||||
assert et_nwo.weekday() in _WEEKDAYS, (
|
||||
f"Expected weekday, got {et_nwo.weekday()} for {et_nwo}"
|
||||
)
|
||||
|
||||
@settings(max_examples=100)
|
||||
@given(
|
||||
dt=st.dates(
|
||||
min_value=datetime(2024, 1, 1).date(),
|
||||
max_value=datetime(2025, 12, 31).date(),
|
||||
)
|
||||
.filter(lambda d: d.weekday() in _WEEKDAYS)
|
||||
.flatmap(
|
||||
lambda d: st.times(
|
||||
min_value=MARKET_OPEN,
|
||||
max_value=time(15, 59, 59),
|
||||
).map(lambda t: datetime.combine(d, t, tzinfo=ET))
|
||||
),
|
||||
)
|
||||
def test_is_market_open_during_market_hours(self, dt: datetime) -> None:
|
||||
"""Timestamps between 9:30 AM and 4:00 PM ET on weekdays are market-open."""
|
||||
assert is_market_open(dt) is True, (
|
||||
f"Expected market open: {dt} (time={dt.time()})"
|
||||
)
|
||||
|
||||
@settings(max_examples=100)
|
||||
@given(dt=_et_datetime_weekend())
|
||||
def test_is_market_closed_on_weekends(self, dt: datetime) -> None:
|
||||
"""Weekend timestamps always have market closed."""
|
||||
assert is_market_open(dt) is False
|
||||
|
||||
|
||||
# ---------------------------------------------------------------------------
|
||||
# Property 21: Gradual entry tranche splitting
|
||||
# **Validates: Requirements 11.3, 11.5**
|
||||
# ---------------------------------------------------------------------------
|
||||
|
||||
|
||||
class TestProperty21GradualEntryTrancheSplitting:
|
||||
"""Property 21: Gradual entry tranche splitting.
|
||||
|
||||
**Validates: Requirements 11.3, 11.5**
|
||||
"""
|
||||
|
||||
# -- should_use_gradual_entry ------------------------------------------
|
||||
|
||||
@settings(max_examples=100)
|
||||
@given(
|
||||
active_pool=st.floats(min_value=100.0, max_value=100_000.0, allow_nan=False, allow_infinity=False),
|
||||
threshold_dollars=st.floats(min_value=1.0, max_value=1000.0, allow_nan=False, allow_infinity=False),
|
||||
excess=st.floats(min_value=0.01, max_value=1000.0, allow_nan=False, allow_infinity=False),
|
||||
)
|
||||
def test_gradual_entry_triggered_above_threshold(
|
||||
self, active_pool: float, threshold_dollars: float, excess: float,
|
||||
) -> None:
|
||||
"""Gradual entry is used when position size exceeds min(threshold, 5% of pool)."""
|
||||
effective = min(threshold_dollars, 0.05 * active_pool)
|
||||
position_size = effective + excess
|
||||
assert should_use_gradual_entry(position_size, active_pool, threshold_dollars) is True
|
||||
|
||||
@settings(max_examples=100)
|
||||
@given(
|
||||
active_pool=st.floats(min_value=100.0, max_value=100_000.0, allow_nan=False, allow_infinity=False),
|
||||
threshold_dollars=st.floats(min_value=1.0, max_value=1000.0, allow_nan=False, allow_infinity=False),
|
||||
fraction=st.floats(min_value=0.0, max_value=1.0, allow_nan=False, allow_infinity=False),
|
||||
)
|
||||
def test_gradual_entry_not_triggered_at_or_below_threshold(
|
||||
self, active_pool: float, threshold_dollars: float, fraction: float,
|
||||
) -> None:
|
||||
"""Gradual entry is NOT used when position size <= effective threshold."""
|
||||
effective = min(threshold_dollars, 0.05 * active_pool)
|
||||
position_size = effective * fraction
|
||||
assert should_use_gradual_entry(position_size, active_pool, threshold_dollars) is False
|
||||
|
||||
# -- split_into_tranches -----------------------------------------------
|
||||
|
||||
@settings(max_examples=100)
|
||||
@given(
|
||||
total_quantity=st.integers(min_value=1, max_value=10_000),
|
||||
num_tranches=st.integers(min_value=1, max_value=20),
|
||||
)
|
||||
def test_tranche_sum_equals_total(
|
||||
self, total_quantity: int, num_tranches: int,
|
||||
) -> None:
|
||||
"""Sum of all tranches must equal the original total quantity."""
|
||||
tranches = split_into_tranches(total_quantity, num_tranches)
|
||||
assert sum(tranches) == total_quantity
|
||||
|
||||
@settings(max_examples=100)
|
||||
@given(
|
||||
total_quantity=st.integers(min_value=1, max_value=10_000),
|
||||
num_tranches=st.integers(min_value=1, max_value=20),
|
||||
)
|
||||
def test_tranche_sizes_approximately_equal(
|
||||
self, total_quantity: int, num_tranches: int,
|
||||
) -> None:
|
||||
"""All tranche sizes differ by at most 1."""
|
||||
tranches = split_into_tranches(total_quantity, num_tranches)
|
||||
assert max(tranches) - min(tranches) <= 1
|
||||
|
||||
@settings(max_examples=100)
|
||||
@given(
|
||||
total_quantity=st.integers(min_value=1, max_value=10_000),
|
||||
num_tranches=st.integers(min_value=1, max_value=20),
|
||||
)
|
||||
def test_tranche_count_matches_requested(
|
||||
self, total_quantity: int, num_tranches: int,
|
||||
) -> None:
|
||||
"""Number of tranches returned matches the requested count."""
|
||||
tranches = split_into_tranches(total_quantity, num_tranches)
|
||||
assert len(tranches) == num_tranches
|
||||
|
||||
# -- create_tranches ---------------------------------------------------
|
||||
|
||||
@settings(max_examples=100)
|
||||
@given(
|
||||
total_quantity=st.integers(min_value=1, max_value=10_000),
|
||||
parent_id=st.text(min_size=1, max_size=36, alphabet=st.characters(whitelist_categories=("L", "N", "Pd"))),
|
||||
num_tranches=st.integers(min_value=1, max_value=10),
|
||||
)
|
||||
def test_all_tranches_reference_same_parent_decision_id(
|
||||
self, total_quantity: int, parent_id: str, num_tranches: int,
|
||||
) -> None:
|
||||
"""Every tranche references the same parent decision ID."""
|
||||
tranches = create_tranches(total_quantity, parent_id, num_tranches)
|
||||
for t in tranches:
|
||||
assert t.parent_decision_id == parent_id
|
||||
|
||||
@settings(max_examples=100)
|
||||
@given(
|
||||
total_quantity=st.integers(min_value=1, max_value=10_000),
|
||||
parent_id=st.text(min_size=1, max_size=36, alphabet=st.characters(whitelist_categories=("L", "N", "Pd"))),
|
||||
num_tranches=st.integers(min_value=1, max_value=10),
|
||||
)
|
||||
def test_create_tranches_quantity_sum(
|
||||
self, total_quantity: int, parent_id: str, num_tranches: int,
|
||||
) -> None:
|
||||
"""Sum of tranche quantities from create_tranches equals total."""
|
||||
tranches = create_tranches(total_quantity, parent_id, num_tranches)
|
||||
assert sum(t.quantity for t in tranches) == total_quantity
|
||||
|
||||
@settings(max_examples=100)
|
||||
@given(
|
||||
total_quantity=st.integers(min_value=1, max_value=10_000),
|
||||
parent_id=st.text(min_size=1, max_size=36, alphabet=st.characters(whitelist_categories=("L", "N", "Pd"))),
|
||||
num_tranches=st.integers(min_value=1, max_value=10),
|
||||
)
|
||||
def test_create_tranches_indices_sequential(
|
||||
self, total_quantity: int, parent_id: str, num_tranches: int,
|
||||
) -> None:
|
||||
"""Tranche indices are sequential starting from 0."""
|
||||
tranches = create_tranches(total_quantity, parent_id, num_tranches)
|
||||
for i, t in enumerate(tranches):
|
||||
assert t.tranche_index == i
|
||||
|
||||
@settings(max_examples=100)
|
||||
@given(
|
||||
total_quantity=st.integers(min_value=1, max_value=10_000),
|
||||
parent_id=st.text(min_size=1, max_size=36, alphabet=st.characters(whitelist_categories=("L", "N", "Pd"))),
|
||||
num_tranches=st.integers(min_value=1, max_value=10),
|
||||
)
|
||||
def test_create_tranches_default_status_pending(
|
||||
self, total_quantity: int, parent_id: str, num_tranches: int,
|
||||
) -> None:
|
||||
"""All tranches start with status 'pending'."""
|
||||
tranches = create_tranches(total_quantity, parent_id, num_tranches)
|
||||
for t in tranches:
|
||||
assert t.status == "pending"
|
||||
Reference in New Issue
Block a user