You cannot select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
6159 lines
267 KiB
Python
6159 lines
267 KiB
Python
"""Integration tests for the full backtest pipeline (no DB/HTTP — uses SnapshotStore directly)."""
|
|
from __future__ import annotations
|
|
|
|
import datetime as dt
|
|
import json
|
|
from pathlib import Path
|
|
from zoneinfo import ZoneInfo
|
|
|
|
import pytest
|
|
import pyarrow as pa
|
|
import pyarrow.parquet as pq
|
|
|
|
_UTC = ZoneInfo("UTC")
|
|
|
|
|
|
def _build_synthetic_store() -> object:
|
|
"""Build a SnapshotStore with synthetic data for end-to-end testing."""
|
|
from libs.backtest.snapshot_store import SnapshotStore
|
|
|
|
# 5 trading days, 2 symbols
|
|
dates = [dt.date(2026, 1, d) for d in [5, 6, 7, 8, 9]]
|
|
|
|
candidates = {
|
|
dt.date(2026, 1, 5): [
|
|
{
|
|
"event_id": "EVT::001",
|
|
"symbol": "AAPL",
|
|
"execution_date": dt.date(2026, 1, 5),
|
|
"entry_date": "2026-01-05",
|
|
"entry_price": 150.0,
|
|
"score": 0.85,
|
|
"sector": "Technology",
|
|
"event_type": "earnings",
|
|
"event_timestamp": "2026-01-02T21:00:00+00:00",
|
|
"filing_time_bucket": "post_market",
|
|
"reaction_date": "2026-01-02",
|
|
"avg_dollar_volume": 5_000_000.0,
|
|
"atr_14": 3.0,
|
|
},
|
|
],
|
|
dt.date(2026, 1, 6): [
|
|
{
|
|
"event_id": "EVT::002",
|
|
"symbol": "MSFT",
|
|
"execution_date": dt.date(2026, 1, 6),
|
|
"entry_date": "2026-01-06",
|
|
"entry_price": 300.0,
|
|
"score": 0.70,
|
|
"sector": "Technology",
|
|
"event_type": "guidance",
|
|
"event_timestamp": "2026-01-05T21:00:00+00:00",
|
|
"filing_time_bucket": "post_market",
|
|
"reaction_date": "2026-01-05",
|
|
"avg_dollar_volume": 10_000_000.0,
|
|
"atr_14": 5.0,
|
|
},
|
|
],
|
|
}
|
|
|
|
bars = {
|
|
"AAPL": {
|
|
dt.date(2026, 1, 5): {"date": dt.date(2026, 1, 5), "open": 150.0, "high": 160.0, "low": 148.0, "close": 158.0, "volume": 1_000_000},
|
|
dt.date(2026, 1, 6): {"date": dt.date(2026, 1, 6), "open": 158.0, "high": 170.0, "low": 155.0, "close": 165.0, "volume": 900_000},
|
|
dt.date(2026, 1, 7): {"date": dt.date(2026, 1, 7), "open": 165.0, "high": 175.0, "low": 160.0, "close": 170.0, "volume": 800_000},
|
|
dt.date(2026, 1, 8): {"date": dt.date(2026, 1, 8), "open": 170.0, "high": 180.0, "low": 165.0, "close": 175.0, "volume": 750_000},
|
|
dt.date(2026, 1, 9): {"date": dt.date(2026, 1, 9), "open": 175.0, "high": 185.0, "low": 170.0, "close": 180.0, "volume": 700_000},
|
|
},
|
|
"MSFT": {
|
|
dt.date(2026, 1, 6): {"date": dt.date(2026, 1, 6), "open": 300.0, "high": 305.0, "low": 280.0, "close": 282.0, "volume": 500_000},
|
|
dt.date(2026, 1, 7): {"date": dt.date(2026, 1, 7), "open": 282.0, "high": 290.0, "low": 270.0, "close": 272.0, "volume": 480_000},
|
|
dt.date(2026, 1, 8): {"date": dt.date(2026, 1, 8), "open": 272.0, "high": 280.0, "low": 260.0, "close": 265.0, "volume": 450_000},
|
|
dt.date(2026, 1, 9): {"date": dt.date(2026, 1, 9), "open": 265.0, "high": 270.0, "low": 255.0, "close": 258.0, "volume": 420_000},
|
|
},
|
|
}
|
|
|
|
return SnapshotStore(
|
|
candidates_by_exec_date=candidates,
|
|
bars_by_symbol_date=bars,
|
|
)
|
|
|
|
|
|
def _build_multi_engine_store() -> object:
|
|
from libs.backtest.snapshot_store import SnapshotStore
|
|
|
|
candidates = {
|
|
dt.date(2026, 1, 7): [
|
|
{
|
|
"event_id": "EVT::SD::SHORT",
|
|
"symbol": "NFLX",
|
|
"execution_date": dt.date(2026, 1, 7),
|
|
"entry_date": "2026-01-07",
|
|
"event_date": "2026-01-06",
|
|
"event_close": 400.0,
|
|
"entry_price": 399.0,
|
|
"score": 0.90,
|
|
"sector": "Communication Services",
|
|
"event_type": "earnings_release",
|
|
"event_timestamp": "2026-01-06T21:00:00+00:00",
|
|
"filing_time_bucket": "post_market",
|
|
"reaction_date": "2026-01-06",
|
|
"reaction_day_return": -0.12,
|
|
"avg_dollar_volume": 8_000_000.0,
|
|
"atr_14": 4.0,
|
|
},
|
|
{
|
|
"event_id": "EVT::AC::LONG",
|
|
"symbol": "AMD",
|
|
"execution_date": dt.date(2026, 1, 7),
|
|
"entry_date": "2026-01-07",
|
|
"event_date": "2026-01-06",
|
|
"event_close": 122.0,
|
|
"entry_price": 123.0,
|
|
"score": 0.88,
|
|
"sector": "Technology",
|
|
"event_type": "earnings_release",
|
|
"event_timestamp": "2026-01-06T21:00:00+00:00",
|
|
"filing_time_bucket": "post_market",
|
|
"reaction_date": "2026-01-07",
|
|
"reaction_day_return": 0.14,
|
|
"avg_dollar_volume": 9_000_000.0,
|
|
"atr_14": 3.0,
|
|
},
|
|
],
|
|
}
|
|
|
|
bars = {
|
|
"NFLX": {
|
|
dt.date(2026, 1, 6): {"date": dt.date(2026, 1, 6), "open": 395.0, "high": 405.0, "low": 390.0, "close": 400.0, "volume": 1_200_000},
|
|
dt.date(2026, 1, 7): {"date": dt.date(2026, 1, 7), "open": 390.0, "high": 392.0, "low": 380.0, "close": 382.0, "volume": 1_100_000},
|
|
dt.date(2026, 1, 8): {"date": dt.date(2026, 1, 8), "open": 382.0, "high": 384.0, "low": 370.0, "close": 372.0, "volume": 1_000_000},
|
|
},
|
|
"AMD": {
|
|
dt.date(2026, 1, 7): {"date": dt.date(2026, 1, 7), "open": 123.0, "high": 130.0, "low": 122.0, "close": 129.0, "volume": 1_500_000},
|
|
dt.date(2026, 1, 8): {"date": dt.date(2026, 1, 8), "open": 129.0, "high": 135.0, "low": 128.0, "close": 134.0, "volume": 1_300_000},
|
|
dt.date(2026, 1, 9): {"date": dt.date(2026, 1, 9), "open": 134.0, "high": 138.0, "low": 133.0, "close": 137.0, "volume": 1_250_000},
|
|
},
|
|
}
|
|
|
|
return SnapshotStore(candidates_by_exec_date=candidates, bars_by_symbol_date=bars)
|
|
|
|
|
|
def _make_config(strategy_engines=None):
|
|
from libs.backtest.domain import (
|
|
BacktestConfig,
|
|
ExecutionConfig,
|
|
ReportingConfig,
|
|
RiskConfig,
|
|
SignalConfig,
|
|
UniverseConfig,
|
|
)
|
|
|
|
return BacktestConfig(
|
|
strategy_name="test_strategy",
|
|
dataset_snapshot_id="test_snapshot",
|
|
universe=UniverseConfig(min_price=5.0, min_avg_dollar_volume=100_000),
|
|
signal=SignalConfig(score_threshold=0.5, max_candidates_per_day=5),
|
|
risk=RiskConfig(
|
|
per_trade_risk_pct=0.01,
|
|
max_daily_new_risk_pct=0.05,
|
|
max_positions=10,
|
|
max_positions_per_sector=5,
|
|
),
|
|
execution=ExecutionConfig(
|
|
entry_fill_model="next_open",
|
|
exit_fill_model="daily_bar_approximation",
|
|
slippage_bps_base=10.0,
|
|
commission_per_share=0.005,
|
|
same_bar_priority="stop_first_conservative",
|
|
max_holding_days=10,
|
|
),
|
|
reporting=ReportingConfig(
|
|
write_trade_blotter=True,
|
|
write_equity_curve=True,
|
|
write_metrics_summary=True,
|
|
generate_plots=False,
|
|
),
|
|
strategy_engines=strategy_engines or [],
|
|
)
|
|
|
|
|
|
@pytest.mark.integration
|
|
class TestBacktestRunIntegration:
|
|
def test_run_completes(self, tmp_path):
|
|
"""Full run completes without error."""
|
|
from apps.backtester.run import BacktestRunner
|
|
from libs.backtest.domain import ExperimentManifest
|
|
|
|
store = _build_synthetic_store()
|
|
manifest = ExperimentManifest(
|
|
experiment_name="test_exp",
|
|
dataset_snapshot_id="test_snapshot",
|
|
base_config="configs/backtest/defaults.json",
|
|
overrides={},
|
|
)
|
|
config = _make_config()
|
|
runner = BacktestRunner(manifest=manifest, config=config, store=store, initial_equity=100_000.0)
|
|
result = runner.run(output_root=tmp_path)
|
|
|
|
assert result.run_id.startswith("bt_")
|
|
assert result.total_trading_days >= 0
|
|
assert result.metrics.trade_count >= 0
|
|
|
|
def test_engine_allowed_macro_regimes_filters_engine_by_date(self, tmp_path):
|
|
from apps.backtester.run import BacktestRunner
|
|
from libs.backtest.domain import (
|
|
BacktestConfig,
|
|
EventTypeProfile,
|
|
ExecutionConfig,
|
|
ExperimentManifest,
|
|
ReportingConfig,
|
|
RiskConfig,
|
|
SignalConfig,
|
|
StrategyEngineConfig,
|
|
UniverseConfig,
|
|
)
|
|
from libs.backtest.snapshot_store import SnapshotStore
|
|
import pyarrow.parquet as pq
|
|
|
|
store = SnapshotStore(
|
|
candidates_by_exec_date={
|
|
dt.date(2026, 1, 5): [
|
|
{
|
|
"event_id": "EVT::ROFF",
|
|
"symbol": "AAPL",
|
|
"execution_date": dt.date(2026, 1, 5),
|
|
"entry_date": "2026-01-05",
|
|
"entry_price": 100.0,
|
|
"score": 0.80,
|
|
"sector": "Technology",
|
|
"event_type": "earnings_release",
|
|
"event_timestamp": "2026-01-02T21:00:00+00:00",
|
|
"filing_time_bucket": "post_market",
|
|
"reaction_date": "2026-01-02",
|
|
"avg_dollar_volume": 5_000_000.0,
|
|
"atr_14": 3.0,
|
|
"reaction_day_return": 0.02,
|
|
},
|
|
],
|
|
dt.date(2026, 1, 6): [
|
|
{
|
|
"event_id": "EVT::RON",
|
|
"symbol": "MSFT",
|
|
"execution_date": dt.date(2026, 1, 6),
|
|
"entry_date": "2026-01-06",
|
|
"entry_price": 110.0,
|
|
"score": 0.82,
|
|
"sector": "Technology",
|
|
"event_type": "earnings_release",
|
|
"event_timestamp": "2026-01-05T21:00:00+00:00",
|
|
"filing_time_bucket": "post_market",
|
|
"reaction_date": "2026-01-05",
|
|
"avg_dollar_volume": 5_000_000.0,
|
|
"atr_14": 3.0,
|
|
"reaction_day_return": 0.02,
|
|
},
|
|
],
|
|
},
|
|
bars_by_symbol_date={
|
|
"AAPL": {
|
|
dt.date(2026, 1, 5): {"date": dt.date(2026, 1, 5), "open": 100.0, "high": 104.0, "low": 98.0, "close": 103.0, "volume": 1_000_000},
|
|
dt.date(2026, 1, 6): {"date": dt.date(2026, 1, 6), "open": 103.0, "high": 105.0, "low": 102.0, "close": 104.0, "volume": 1_000_000},
|
|
},
|
|
"MSFT": {
|
|
dt.date(2026, 1, 6): {"date": dt.date(2026, 1, 6), "open": 110.0, "high": 114.0, "low": 109.0, "close": 113.0, "volume": 1_000_000},
|
|
dt.date(2026, 1, 7): {"date": dt.date(2026, 1, 7), "open": 113.0, "high": 115.0, "low": 111.0, "close": 114.0, "volume": 1_000_000},
|
|
},
|
|
},
|
|
macro_by_date={
|
|
dt.date(2026, 1, 5): {
|
|
"spy_close": 90.0,
|
|
"spy_sma_20": 100.0,
|
|
"qqq_close": 180.0,
|
|
"qqq_sma_20": 200.0,
|
|
},
|
|
dt.date(2026, 1, 6): {
|
|
"spy_close": 110.0,
|
|
"spy_sma_20": 100.0,
|
|
"qqq_close": 210.0,
|
|
"qqq_sma_20": 200.0,
|
|
},
|
|
},
|
|
)
|
|
|
|
manifest = ExperimentManifest(
|
|
experiment_name="macro_gate_engine_test",
|
|
dataset_snapshot_id="test_snapshot",
|
|
base_config="configs/backtest/defaults.json",
|
|
overrides={},
|
|
strategy_engines=[
|
|
StrategyEngineConfig(
|
|
engine_id="macro_gated_engine",
|
|
event_types=["earnings_release"],
|
|
timing_class="any",
|
|
direction="long_only",
|
|
entry_timing_policy="next_open",
|
|
allowed_macro_regimes=["risk_on"],
|
|
),
|
|
],
|
|
)
|
|
config = BacktestConfig(
|
|
strategy_name="return_max_long_v1",
|
|
dataset_snapshot_id="test_snapshot",
|
|
universe=UniverseConfig(min_price=5.0, min_avg_dollar_volume=0.0),
|
|
signal=SignalConfig(score_threshold=0.5, max_candidates_per_day=5),
|
|
risk=RiskConfig(
|
|
per_trade_risk_pct=0.01,
|
|
max_daily_new_risk_pct=0.10,
|
|
max_positions=5,
|
|
max_positions_per_sector=5,
|
|
macro_regime_enabled=True,
|
|
macro_regime_mode="spy_qqq_scaler",
|
|
macro_regime_neutral_size_scaler=1.0,
|
|
macro_regime_risk_off_size_scaler=1.0,
|
|
macro_regime_risk_off_a_tier_only=False,
|
|
veto_unknown_direction=False,
|
|
veto_bearish_direction=False,
|
|
),
|
|
execution=ExecutionConfig(
|
|
entry_fill_model="next_open",
|
|
exit_fill_model="daily_bar_approximation",
|
|
slippage_bps_base=0.0,
|
|
commission_per_share=0.0,
|
|
same_bar_priority="stop_first_conservative",
|
|
max_holding_days=1,
|
|
),
|
|
reporting=ReportingConfig(
|
|
write_trade_blotter=True,
|
|
write_equity_curve=True,
|
|
write_metrics_summary=True,
|
|
generate_plots=False,
|
|
),
|
|
event_type_profiles={
|
|
"earnings_release": EventTypeProfile(enabled=True),
|
|
},
|
|
strategy_engines=manifest.strategy_engines,
|
|
)
|
|
|
|
runner = BacktestRunner(manifest=manifest, config=config, store=store, initial_equity=100_000.0)
|
|
result = runner.run(output_root=tmp_path)
|
|
trade_blotter = pq.read_table(tmp_path / result.run_id / "artifacts" / "trade_blotter.parquet").to_pylist()
|
|
|
|
assert [row["symbol"] for row in trade_blotter] == ["MSFT"]
|
|
|
|
def test_output_files_created(self, tmp_path):
|
|
"""All expected output files are written."""
|
|
from apps.backtester.run import BacktestRunner
|
|
from libs.backtest.domain import ExperimentManifest
|
|
|
|
store = _build_synthetic_store()
|
|
manifest = ExperimentManifest(
|
|
experiment_name="test_exp",
|
|
dataset_snapshot_id="test_snapshot",
|
|
base_config="configs/backtest/defaults.json",
|
|
overrides={},
|
|
)
|
|
config = _make_config()
|
|
runner = BacktestRunner(manifest=manifest, config=config, store=store, initial_equity=100_000.0)
|
|
result = runner.run(output_root=tmp_path)
|
|
|
|
run_dir = tmp_path / result.run_id
|
|
assert run_dir.exists()
|
|
assert (run_dir / "metadata.json").exists()
|
|
assert (run_dir / "manifest.json").exists()
|
|
|
|
def test_same_day_cash_recycle_replaces_stale_core(self, tmp_path):
|
|
from apps.backtester.run import BacktestRunner
|
|
from libs.backtest.domain import (
|
|
BacktestConfig,
|
|
ExecutionConfig,
|
|
ExperimentManifest,
|
|
ReportingConfig,
|
|
RiskConfig,
|
|
SignalConfig,
|
|
StrategyEngineConfig,
|
|
UniverseConfig,
|
|
)
|
|
from libs.backtest.snapshot_store import SnapshotStore
|
|
|
|
d1 = dt.date(2026, 1, 5)
|
|
d2 = dt.date(2026, 1, 6)
|
|
d3 = dt.date(2026, 1, 7)
|
|
|
|
store = SnapshotStore(
|
|
candidates_by_exec_date={
|
|
d1: [
|
|
{
|
|
"event_id": "EVT::OLD",
|
|
"symbol": "OLD",
|
|
"execution_date": d1,
|
|
"entry_date": d1.isoformat(),
|
|
"event_date": d1.isoformat(),
|
|
"event_close": 100.0,
|
|
"entry_price": 100.0,
|
|
"score": 0.60,
|
|
"sector": "Technology",
|
|
"event_type": "earnings_release",
|
|
"event_direction": "bullish",
|
|
"guidance_status": "not_provided",
|
|
"event_timestamp": "2026-01-05T21:00:00+00:00",
|
|
"filing_time_bucket": "post_market",
|
|
"reaction_date": d1.isoformat(),
|
|
"reaction_day_return": 0.08,
|
|
"close_location": 0.74,
|
|
"volume_ratio_20d": 2.1,
|
|
"gap_size": 0.01,
|
|
"avg_dollar_volume": 5_000_000.0,
|
|
"atr_14": 2.0,
|
|
},
|
|
],
|
|
d2: [
|
|
{
|
|
"event_id": "EVT::NEW",
|
|
"symbol": "NEW",
|
|
"execution_date": d2,
|
|
"entry_date": d2.isoformat(),
|
|
"event_date": d2.isoformat(),
|
|
"event_close": 50.0,
|
|
"entry_price": 50.0,
|
|
"score": 0.80,
|
|
"sector": "Technology",
|
|
"event_type": "earnings_release",
|
|
"event_direction": "bullish",
|
|
"guidance_status": "not_provided",
|
|
"event_timestamp": "2026-01-06T21:00:00+00:00",
|
|
"filing_time_bucket": "post_market",
|
|
"reaction_date": d2.isoformat(),
|
|
"reaction_day_return": 0.09,
|
|
"close_location": 0.76,
|
|
"volume_ratio_20d": 2.2,
|
|
"gap_size": 0.01,
|
|
"avg_dollar_volume": 5_000_000.0,
|
|
"atr_14": 2.0,
|
|
},
|
|
],
|
|
},
|
|
bars_by_symbol_date={
|
|
"OLD": {
|
|
d1: {"date": d1, "open": 100.0, "high": 100.0, "low": 99.0, "close": 100.0, "volume": 1_000_000},
|
|
d2: {"date": d2, "open": 100.0, "high": 102.0, "low": 100.0, "close": 101.0, "volume": 1_000_000},
|
|
d3: {"date": d3, "open": 101.0, "high": 101.0, "low": 101.0, "close": 101.0, "volume": 1_000_000},
|
|
},
|
|
"NEW": {
|
|
d2: {"date": d2, "open": 50.0, "high": 50.0, "low": 50.0, "close": 50.0, "volume": 1_000_000},
|
|
d3: {"date": d3, "open": 50.0, "high": 55.0, "low": 50.0, "close": 55.0, "volume": 1_000_000},
|
|
},
|
|
},
|
|
)
|
|
|
|
engine = StrategyEngineConfig(
|
|
engine_id="core",
|
|
event_types=["earnings_release"],
|
|
timing_class="same_day",
|
|
direction="long_only",
|
|
entry_timing_policy="reaction_close",
|
|
score_threshold_override=0.5,
|
|
recycle_on_cash_block=True,
|
|
recycle_min_days_held=1,
|
|
recycle_min_score_delta=0.05,
|
|
recycle_allowed_victim_engine_ids=["core"],
|
|
recycle_positive_pnl_only=True,
|
|
enabled=True,
|
|
)
|
|
manifest = ExperimentManifest(
|
|
experiment_name="test_same_day_recycle",
|
|
dataset_snapshot_id="test_snapshot",
|
|
base_config="configs/backtest/defaults.json",
|
|
overrides={},
|
|
strategy_engines=[engine],
|
|
)
|
|
config = BacktestConfig(
|
|
strategy_name="test_strategy",
|
|
dataset_snapshot_id="test_snapshot",
|
|
universe=UniverseConfig(min_price=5.0, min_avg_dollar_volume=100_000),
|
|
signal=SignalConfig(score_threshold=0.5, max_candidates_per_day=5),
|
|
risk=RiskConfig(
|
|
per_trade_risk_pct=1.0,
|
|
max_daily_new_risk_pct=1.0,
|
|
max_positions=10,
|
|
max_positions_per_sector=5,
|
|
max_position_value_pct=1.0,
|
|
),
|
|
execution=ExecutionConfig(max_holding_days=10),
|
|
reporting=ReportingConfig(),
|
|
strategy_engines=[engine],
|
|
)
|
|
|
|
runner = BacktestRunner(manifest=manifest, config=config, store=store, initial_equity=1_000.0)
|
|
result = runner.run(output_root=tmp_path)
|
|
run_dir = tmp_path / result.run_id
|
|
|
|
recycle_trades = [trade for trade in runner._closed_trades if trade.exit_reason.value == "RECYCLE"]
|
|
assert recycle_trades
|
|
assert recycle_trades[0].symbol == "OLD"
|
|
assert any(trade.symbol == "NEW" for trade in runner._closed_trades)
|
|
assert result.metrics.trade_count >= 2
|
|
assert (run_dir / "resolved_config.json").exists()
|
|
assert (run_dir / "metrics" / "metrics_summary.json").exists()
|
|
assert (run_dir / "plots").exists() # empty dir
|
|
|
|
def test_next_open_cash_recycle_replaces_stale_next_open_position(self, tmp_path):
|
|
from apps.backtester.run import BacktestRunner
|
|
from libs.backtest.domain import (
|
|
BacktestConfig,
|
|
ExecutionConfig,
|
|
ExperimentManifest,
|
|
ReportingConfig,
|
|
RiskConfig,
|
|
SignalConfig,
|
|
StrategyEngineConfig,
|
|
UniverseConfig,
|
|
)
|
|
from libs.backtest.snapshot_store import SnapshotStore
|
|
|
|
d1 = dt.date(2026, 1, 5)
|
|
d2 = dt.date(2026, 1, 6)
|
|
d3 = dt.date(2026, 1, 7)
|
|
|
|
store = SnapshotStore(
|
|
candidates_by_exec_date={
|
|
d1: [
|
|
{
|
|
"event_id": "EVT::OLD::NEXT",
|
|
"symbol": "OLDN",
|
|
"execution_date": d1,
|
|
"entry_date": d1.isoformat(),
|
|
"event_date": d1.isoformat(),
|
|
"event_close": 100.0,
|
|
"entry_price": 100.0,
|
|
"score": 0.60,
|
|
"sector": "Technology",
|
|
"event_type": "earnings_release",
|
|
"event_direction": "bullish",
|
|
"guidance_status": "raised",
|
|
"event_timestamp": "2026-01-05T13:00:00+00:00",
|
|
"filing_time_bucket": "pre_market",
|
|
"reaction_date": d1.isoformat(),
|
|
"reaction_day_return": 0.01,
|
|
"close_location": 0.65,
|
|
"volume_ratio_20d": 1.2,
|
|
"gap_size": 0.01,
|
|
"avg_dollar_volume": 5_000_000.0,
|
|
"atr_14": 2.0,
|
|
},
|
|
],
|
|
d2: [
|
|
{
|
|
"event_id": "EVT::NEW::NEXT",
|
|
"symbol": "NEWN",
|
|
"execution_date": d2,
|
|
"entry_date": d2.isoformat(),
|
|
"event_date": d2.isoformat(),
|
|
"event_close": 50.0,
|
|
"entry_price": 50.0,
|
|
"score": 0.80,
|
|
"sector": "Technology",
|
|
"event_type": "earnings_release",
|
|
"event_direction": "bullish",
|
|
"guidance_status": "raised",
|
|
"event_timestamp": "2026-01-06T13:00:00+00:00",
|
|
"filing_time_bucket": "pre_market",
|
|
"reaction_date": d2.isoformat(),
|
|
"reaction_day_return": 0.01,
|
|
"close_location": 0.66,
|
|
"volume_ratio_20d": 1.2,
|
|
"gap_size": 0.01,
|
|
"avg_dollar_volume": 5_000_000.0,
|
|
"atr_14": 2.0,
|
|
},
|
|
],
|
|
},
|
|
bars_by_symbol_date={
|
|
"OLDN": {
|
|
d1: {"date": d1, "open": 100.0, "high": 101.0, "low": 99.0, "close": 100.0, "volume": 1_000_000},
|
|
d2: {"date": d2, "open": 100.0, "high": 102.0, "low": 100.0, "close": 101.0, "volume": 1_000_000},
|
|
d3: {"date": d3, "open": 101.0, "high": 101.0, "low": 100.0, "close": 100.5, "volume": 1_000_000},
|
|
},
|
|
"NEWN": {
|
|
d2: {"date": d2, "open": 50.0, "high": 50.0, "low": 50.0, "close": 50.0, "volume": 1_000_000},
|
|
d3: {"date": d3, "open": 50.0, "high": 54.0, "low": 50.0, "close": 53.0, "volume": 1_000_000},
|
|
},
|
|
},
|
|
)
|
|
|
|
engine = StrategyEngineConfig(
|
|
engine_id="next_open_guidance",
|
|
event_types=["earnings_release"],
|
|
timing_class="any",
|
|
direction="long_only",
|
|
entry_timing_policy="next_open",
|
|
score_threshold_override=0.5,
|
|
recycle_on_cash_block=True,
|
|
recycle_min_days_held=1,
|
|
recycle_min_score_delta=0.05,
|
|
recycle_allowed_victim_engine_ids=["next_open_guidance"],
|
|
recycle_positive_pnl_only=True,
|
|
enabled=True,
|
|
)
|
|
manifest = ExperimentManifest(
|
|
experiment_name="test_next_open_recycle",
|
|
dataset_snapshot_id="test_snapshot",
|
|
base_config="configs/backtest/defaults.json",
|
|
overrides={},
|
|
strategy_engines=[engine],
|
|
)
|
|
config = BacktestConfig(
|
|
strategy_name="test_strategy",
|
|
dataset_snapshot_id="test_snapshot",
|
|
universe=UniverseConfig(min_price=5.0, min_avg_dollar_volume=100_000),
|
|
signal=SignalConfig(score_threshold=0.5, max_candidates_per_day=5),
|
|
risk=RiskConfig(
|
|
per_trade_risk_pct=1.0,
|
|
max_daily_new_risk_pct=1.0,
|
|
max_positions=10,
|
|
max_positions_per_sector=5,
|
|
max_position_value_pct=1.0,
|
|
),
|
|
execution=ExecutionConfig(max_holding_days=10),
|
|
reporting=ReportingConfig(),
|
|
strategy_engines=[engine],
|
|
)
|
|
|
|
runner = BacktestRunner(manifest=manifest, config=config, store=store, initial_equity=1_000.0)
|
|
result = runner.run(output_root=tmp_path)
|
|
|
|
recycle_trades = [trade for trade in runner._closed_trades if trade.exit_reason.value == "RECYCLE"]
|
|
assert recycle_trades
|
|
assert recycle_trades[0].symbol == "OLDN"
|
|
assert any(trade.symbol == "NEWN" for trade in runner._closed_trades)
|
|
assert result.metrics.trade_count >= 2
|
|
|
|
def test_next_open_cash_recycle_can_target_stale_cross_engine_victim(self, tmp_path):
|
|
from apps.backtester.run import BacktestRunner
|
|
from libs.backtest.domain import (
|
|
BacktestConfig,
|
|
ExecutionConfig,
|
|
ExperimentManifest,
|
|
ReportingConfig,
|
|
RiskConfig,
|
|
SignalConfig,
|
|
StrategyEngineConfig,
|
|
UniverseConfig,
|
|
)
|
|
from libs.backtest.snapshot_store import SnapshotStore
|
|
|
|
d1 = dt.date(2026, 1, 5)
|
|
d2 = dt.date(2026, 1, 6)
|
|
d3 = dt.date(2026, 1, 7)
|
|
d4 = dt.date(2026, 1, 8)
|
|
d5 = dt.date(2026, 1, 9)
|
|
|
|
store = SnapshotStore(
|
|
candidates_by_exec_date={
|
|
d1: [
|
|
{
|
|
"event_id": "EVT::OLD::CROSS",
|
|
"symbol": "OLDX",
|
|
"execution_date": d1,
|
|
"entry_date": d1.isoformat(),
|
|
"event_date": d1.isoformat(),
|
|
"event_close": 100.0,
|
|
"entry_price": 100.0,
|
|
"score": 0.55,
|
|
"sector": "Technology",
|
|
"event_type": "earnings_release",
|
|
"event_direction": "bullish",
|
|
"guidance_status": "raised",
|
|
"event_timestamp": "2026-01-05T13:00:00+00:00",
|
|
"filing_time_bucket": "pre_market",
|
|
"reaction_date": d1.isoformat(),
|
|
"reaction_day_return": 0.01,
|
|
"close_location": 0.60,
|
|
"volume_ratio_20d": 1.2,
|
|
"gap_size": 0.01,
|
|
"avg_dollar_volume": 5_000_000.0,
|
|
"atr_14": 2.0,
|
|
},
|
|
],
|
|
d5: [
|
|
{
|
|
"event_id": "EVT::NEW::CROSS",
|
|
"symbol": "NEWX",
|
|
"execution_date": d5,
|
|
"entry_date": d5.isoformat(),
|
|
"event_date": d5.isoformat(),
|
|
"event_close": 50.0,
|
|
"entry_price": 50.0,
|
|
"score": 0.80,
|
|
"sector": "Technology",
|
|
"event_type": "other_material_event",
|
|
"event_direction": "bullish",
|
|
"guidance_status": "not_provided",
|
|
"event_timestamp": "2026-01-09T13:00:00+00:00",
|
|
"filing_time_bucket": "post_market",
|
|
"reaction_date": d5.isoformat(),
|
|
"reaction_day_return": 0.01,
|
|
"close_location": 0.70,
|
|
"volume_ratio_20d": 1.2,
|
|
"gap_size": 0.01,
|
|
"avg_dollar_volume": 5_000_000.0,
|
|
"atr_14": 2.0,
|
|
},
|
|
],
|
|
},
|
|
bars_by_symbol_date={
|
|
"OLDX": {
|
|
d1: {"date": d1, "open": 100.0, "high": 101.0, "low": 99.0, "close": 100.0, "volume": 1_000_000},
|
|
d2: {"date": d2, "open": 100.0, "high": 101.0, "low": 99.5, "close": 100.2, "volume": 1_000_000},
|
|
d3: {"date": d3, "open": 100.2, "high": 100.8, "low": 99.8, "close": 100.3, "volume": 1_000_000},
|
|
d4: {"date": d4, "open": 100.3, "high": 100.7, "low": 100.0, "close": 100.4, "volume": 1_000_000},
|
|
d5: {"date": d5, "open": 100.4, "high": 100.8, "low": 100.2, "close": 100.5, "volume": 1_000_000},
|
|
},
|
|
"NEWX": {
|
|
d5: {"date": d5, "open": 50.0, "high": 50.0, "low": 50.0, "close": 50.0, "volume": 1_000_000},
|
|
},
|
|
},
|
|
)
|
|
|
|
old_engine = StrategyEngineConfig(
|
|
engine_id="stale_next",
|
|
event_types=["earnings_release"],
|
|
timing_class="any",
|
|
direction="long_only",
|
|
entry_timing_policy="next_open",
|
|
score_threshold_override=0.5,
|
|
enabled=True,
|
|
)
|
|
new_engine = StrategyEngineConfig(
|
|
engine_id="fresh_next",
|
|
event_types=["other_material_event"],
|
|
timing_class="any",
|
|
direction="long_only",
|
|
entry_timing_policy="next_open",
|
|
score_threshold_override=0.5,
|
|
recycle_on_cash_block=True,
|
|
recycle_min_days_held=3,
|
|
recycle_min_score_delta=0.05,
|
|
recycle_allow_any_victim_engine=True,
|
|
recycle_max_victim_fitness=0.45,
|
|
recycle_max_victim_unrealized_r=0.30,
|
|
recycle_positive_pnl_only=True,
|
|
enabled=True,
|
|
)
|
|
manifest = ExperimentManifest(
|
|
experiment_name="test_cross_engine_recycle",
|
|
dataset_snapshot_id="test_snapshot",
|
|
base_config="configs/backtest/defaults.json",
|
|
overrides={},
|
|
strategy_engines=[old_engine, new_engine],
|
|
)
|
|
config = BacktestConfig(
|
|
strategy_name="test_strategy",
|
|
dataset_snapshot_id="test_snapshot",
|
|
universe=UniverseConfig(min_price=5.0, min_avg_dollar_volume=100_000),
|
|
signal=SignalConfig(score_threshold=0.5, max_candidates_per_day=5),
|
|
risk=RiskConfig(
|
|
per_trade_risk_pct=1.0,
|
|
max_daily_new_risk_pct=1.0,
|
|
max_positions=10,
|
|
max_positions_per_sector=5,
|
|
max_position_value_pct=1.0,
|
|
),
|
|
execution=ExecutionConfig(max_holding_days=5),
|
|
reporting=ReportingConfig(),
|
|
strategy_engines=[old_engine, new_engine],
|
|
)
|
|
|
|
runner = BacktestRunner(manifest=manifest, config=config, store=store, initial_equity=1_000.0)
|
|
result = runner.run(output_root=tmp_path)
|
|
|
|
recycle_trades = [trade for trade in runner._closed_trades if trade.exit_reason.value == "RECYCLE"]
|
|
assert recycle_trades
|
|
assert recycle_trades[0].symbol == "OLDX"
|
|
assert any(trade.symbol == "NEWX" for trade in runner._closed_trades)
|
|
assert result.metrics.trade_count >= 2
|
|
|
|
def test_rotation_can_skip_large_unrealized_winner(self, tmp_path):
|
|
from apps.backtester.run import BacktestRunner
|
|
from libs.backtest.domain import (
|
|
BacktestConfig,
|
|
ExecutionConfig,
|
|
ExperimentManifest,
|
|
ReportingConfig,
|
|
RiskConfig,
|
|
SignalConfig,
|
|
StrategyEngineConfig,
|
|
UniverseConfig,
|
|
)
|
|
from libs.backtest.snapshot_store import SnapshotStore
|
|
|
|
d1 = dt.date(2026, 1, 5)
|
|
d2 = dt.date(2026, 1, 6)
|
|
d3 = dt.date(2026, 1, 7)
|
|
d4 = dt.date(2026, 1, 8)
|
|
d5 = dt.date(2026, 1, 9)
|
|
|
|
store = SnapshotStore(
|
|
candidates_by_exec_date={
|
|
d1: [
|
|
{
|
|
"event_id": "EVT::OLD::ROT",
|
|
"symbol": "OLDR",
|
|
"execution_date": d1,
|
|
"entry_date": d1.isoformat(),
|
|
"event_date": d1.isoformat(),
|
|
"event_close": 100.0,
|
|
"entry_price": 100.0,
|
|
"score": 0.60,
|
|
"sector": "Technology",
|
|
"event_type": "earnings_release",
|
|
"event_direction": "bullish",
|
|
"guidance_status": "raised",
|
|
"event_timestamp": "2026-01-05T21:00:00+00:00",
|
|
"filing_time_bucket": "post_market",
|
|
"reaction_date": d1.isoformat(),
|
|
"reaction_day_return": 0.03,
|
|
"close_location": 0.70,
|
|
"volume_ratio_20d": 1.5,
|
|
"gap_size": 0.01,
|
|
"avg_dollar_volume": 5_000_000.0,
|
|
"atr_14": 2.0,
|
|
},
|
|
],
|
|
d5: [
|
|
{
|
|
"event_id": "EVT::NEW::ROT",
|
|
"symbol": "NEWR",
|
|
"execution_date": d5,
|
|
"entry_date": d5.isoformat(),
|
|
"event_date": d5.isoformat(),
|
|
"event_close": 50.0,
|
|
"entry_price": 50.0,
|
|
"score": 0.85,
|
|
"sector": "Technology",
|
|
"event_type": "earnings_release",
|
|
"event_direction": "bullish",
|
|
"guidance_status": "raised",
|
|
"event_timestamp": "2026-01-09T21:00:00+00:00",
|
|
"filing_time_bucket": "post_market",
|
|
"reaction_date": d5.isoformat(),
|
|
"reaction_day_return": 0.04,
|
|
"close_location": 0.72,
|
|
"volume_ratio_20d": 1.6,
|
|
"gap_size": 0.01,
|
|
"avg_dollar_volume": 5_000_000.0,
|
|
"atr_14": 2.0,
|
|
},
|
|
],
|
|
},
|
|
bars_by_symbol_date={
|
|
"OLDR": {
|
|
d1: {"date": d1, "open": 100.0, "high": 100.0, "low": 99.0, "close": 100.0, "volume": 1_000_000},
|
|
d2: {"date": d2, "open": 100.0, "high": 101.0, "low": 99.8, "close": 100.8, "volume": 1_000_000},
|
|
d3: {"date": d3, "open": 100.8, "high": 102.5, "low": 100.6, "close": 102.4, "volume": 1_000_000},
|
|
d4: {"date": d4, "open": 102.4, "high": 103.4, "low": 102.2, "close": 103.0, "volume": 1_000_000},
|
|
d5: {"date": d5, "open": 103.0, "high": 103.8, "low": 102.8, "close": 103.5, "volume": 1_000_000},
|
|
},
|
|
"NEWR": {
|
|
d5: {"date": d5, "open": 50.0, "high": 50.0, "low": 50.0, "close": 50.0, "volume": 1_000_000},
|
|
},
|
|
},
|
|
)
|
|
|
|
engine = StrategyEngineConfig(
|
|
engine_id="rot_next",
|
|
event_types=["earnings_release"],
|
|
timing_class="any",
|
|
direction="long_only",
|
|
entry_timing_policy="next_open",
|
|
score_threshold_override=0.5,
|
|
target_1_r_override=6.0,
|
|
rotation_enabled=True,
|
|
rotation_min_days_held=3,
|
|
rotation_fitness_threshold=0.45,
|
|
rotation_min_candidate_score=0.8,
|
|
rotation_max_unrealized_r=0.5,
|
|
enabled=True,
|
|
)
|
|
manifest = ExperimentManifest(
|
|
experiment_name="test_rotation_max_unrealized_r",
|
|
dataset_snapshot_id="test_snapshot",
|
|
base_config="configs/backtest/defaults.json",
|
|
overrides={},
|
|
strategy_engines=[engine],
|
|
)
|
|
config = BacktestConfig(
|
|
strategy_name="test_strategy",
|
|
dataset_snapshot_id="test_snapshot",
|
|
universe=UniverseConfig(min_price=5.0, min_avg_dollar_volume=100_000),
|
|
signal=SignalConfig(score_threshold=0.5, max_candidates_per_day=5),
|
|
risk=RiskConfig(
|
|
per_trade_risk_pct=1.0,
|
|
max_daily_new_risk_pct=1.0,
|
|
max_positions=10,
|
|
max_positions_per_sector=5,
|
|
max_position_value_pct=1.0,
|
|
),
|
|
execution=ExecutionConfig(max_holding_days=5),
|
|
reporting=ReportingConfig(),
|
|
strategy_engines=[engine],
|
|
)
|
|
|
|
runner = BacktestRunner(manifest=manifest, config=config, store=store, initial_equity=1_000.0)
|
|
result = runner.run(output_root=tmp_path)
|
|
|
|
rotation_trades = [trade for trade in runner._closed_trades if trade.exit_reason.value == "ROTATION"]
|
|
assert not rotation_trades
|
|
assert not any(trade.symbol == "NEWR" for trade in runner._closed_trades)
|
|
assert result.metrics.trade_count == 1
|
|
|
|
def test_equity_curve_has_all_days(self, tmp_path):
|
|
"""Equity curve has one entry per candidate date."""
|
|
from apps.backtester.run import BacktestRunner
|
|
from libs.backtest.domain import ExperimentManifest
|
|
|
|
store = _build_synthetic_store()
|
|
manifest = ExperimentManifest(
|
|
experiment_name="test_exp",
|
|
dataset_snapshot_id="test_snapshot",
|
|
base_config="configs/backtest/defaults.json",
|
|
overrides={},
|
|
)
|
|
config = _make_config()
|
|
runner = BacktestRunner(manifest=manifest, config=config, store=store)
|
|
result = runner.run()
|
|
|
|
# Should have simulated days covering the range (all_trading_days between
|
|
# first and last execution date), plus the initial equity state
|
|
assert result.total_trading_days >= 2
|
|
|
|
def test_deterministic_results(self, tmp_path):
|
|
"""Two runs with same inputs produce identical metrics."""
|
|
from apps.backtester.run import BacktestRunner
|
|
from libs.backtest.domain import ExperimentManifest
|
|
|
|
manifest = ExperimentManifest(
|
|
experiment_name="test_exp",
|
|
dataset_snapshot_id="test_snapshot",
|
|
base_config="configs/backtest/defaults.json",
|
|
overrides={},
|
|
)
|
|
config = _make_config()
|
|
|
|
store1 = _build_synthetic_store()
|
|
runner1 = BacktestRunner(manifest=manifest, config=config, store=store1)
|
|
result1 = runner1.run()
|
|
|
|
store2 = _build_synthetic_store()
|
|
runner2 = BacktestRunner(manifest=manifest, config=config, store=store2)
|
|
result2 = runner2.run()
|
|
|
|
assert result1.metrics.trade_count == result2.metrics.trade_count
|
|
assert result1.metrics.win_rate == result2.metrics.win_rate
|
|
assert result1.metrics.total_return_pct == result2.metrics.total_return_pct
|
|
assert result1.total_candidates_seen == result2.total_candidates_seen
|
|
assert result1.total_orders_rejected == result2.total_orders_rejected
|
|
|
|
def test_no_future_data_used(self, tmp_path):
|
|
"""Candidates for day D should not appear in a simulation of day D-1."""
|
|
from libs.backtest.snapshot_store import SnapshotStore
|
|
|
|
candidates = {
|
|
dt.date(2026, 1, 5): [
|
|
{
|
|
"event_id": "EVT::001",
|
|
"symbol": "AAPL",
|
|
"execution_date": dt.date(2026, 1, 5),
|
|
"entry_date": "2026-01-05",
|
|
"entry_price": 150.0,
|
|
"score": 0.85,
|
|
"sector": "Technology",
|
|
"event_type": "earnings",
|
|
"event_timestamp": "2026-01-02T21:00:00+00:00",
|
|
"filing_time_bucket": "post_market",
|
|
"reaction_date": "2026-01-02",
|
|
"avg_dollar_volume": 5_000_000.0,
|
|
"atr_14": 3.0,
|
|
}
|
|
],
|
|
dt.date(2026, 1, 6): [
|
|
{
|
|
"event_id": "EVT::FUTURE",
|
|
"symbol": "FUTURE_TICKER",
|
|
"execution_date": dt.date(2026, 1, 6),
|
|
"entry_date": "2026-01-06",
|
|
"entry_price": 50.0,
|
|
"score": 0.99,
|
|
"sector": "Technology",
|
|
"event_type": "earnings",
|
|
"event_timestamp": "2026-01-05T21:00:00+00:00",
|
|
"filing_time_bucket": "post_market",
|
|
"reaction_date": "2026-01-05",
|
|
"avg_dollar_volume": 1_000_000.0,
|
|
"atr_14": 1.0,
|
|
}
|
|
],
|
|
}
|
|
bars = {
|
|
"AAPL": {dt.date(2026, 1, 5): {"date": dt.date(2026, 1, 5), "open": 150.0, "high": 160.0, "low": 148.0, "close": 158.0, "volume": 1_000_000}},
|
|
"FUTURE_TICKER": {dt.date(2026, 1, 6): {"date": dt.date(2026, 1, 6), "open": 50.0, "high": 55.0, "low": 48.0, "close": 52.0, "volume": 500_000}},
|
|
}
|
|
store = SnapshotStore(candidates_by_exec_date=candidates, bars_by_symbol_date=bars)
|
|
|
|
# Querying Jan 5 should NOT return FUTURE_TICKER candidate
|
|
rows = store.get_candidates_for_date(dt.date(2026, 1, 5))
|
|
symbols = [r["symbol"] for r in rows]
|
|
assert "FUTURE_TICKER" not in symbols
|
|
assert "AAPL" in symbols
|
|
|
|
def test_multi_engine_run_writes_per_engine_metrics(self, tmp_path):
|
|
from apps.backtester.run import BacktestRunner
|
|
from libs.backtest.domain import ExperimentManifest, StrategyEngineConfig
|
|
import json
|
|
import pyarrow.parquet as pq
|
|
|
|
store = _build_multi_engine_store()
|
|
manifest = ExperimentManifest(
|
|
experiment_name="portfolio_v2",
|
|
dataset_snapshot_id="test_snapshot",
|
|
base_config="configs/backtest/defaults.json",
|
|
overrides={},
|
|
strategy_engines=[
|
|
StrategyEngineConfig(
|
|
engine_id="earnings_same_day_short_v1",
|
|
event_types=["earnings_release"],
|
|
timing_class="same_day",
|
|
direction="short_only",
|
|
entry_timing_policy="next_open",
|
|
max_holding_days=3,
|
|
engine_risk_budget_pct=0.40,
|
|
),
|
|
StrategyEngineConfig(
|
|
engine_id="earnings_after_close_long_v1",
|
|
event_types=["earnings_release"],
|
|
timing_class="after_close",
|
|
direction="long_only",
|
|
entry_timing_policy="next_open",
|
|
max_holding_days=5,
|
|
engine_risk_budget_pct=0.25,
|
|
),
|
|
StrategyEngineConfig(
|
|
engine_id="earnings_after_close_short_v1",
|
|
event_types=["earnings_release"],
|
|
timing_class="after_close",
|
|
direction="short_only",
|
|
entry_timing_policy="next_open",
|
|
max_holding_days=3,
|
|
engine_risk_budget_pct=0.10,
|
|
shadow_only=True,
|
|
),
|
|
],
|
|
)
|
|
config = _make_config(strategy_engines=manifest.strategy_engines)
|
|
runner = BacktestRunner(manifest=manifest, config=config, store=store, initial_equity=100_000.0)
|
|
result = runner.run(output_root=tmp_path)
|
|
|
|
run_dir = tmp_path / result.run_id
|
|
per_engine_metrics = run_dir / "metrics" / "per_engine_metrics.json"
|
|
attribution_by_engine = run_dir / "metrics" / "attribution_by_engine.csv"
|
|
|
|
assert per_engine_metrics.exists()
|
|
assert attribution_by_engine.exists()
|
|
|
|
payload = per_engine_metrics.read_text()
|
|
assert "earnings_same_day_short_v1" in payload
|
|
assert "earnings_after_close_long_v1" in payload
|
|
assert "earnings_after_close_short_v1" in payload
|
|
assert result.metrics.trade_count >= 1
|
|
|
|
trade_blotter = pq.read_table(run_dir / "artifacts" / "trade_blotter.parquet").to_pylist()
|
|
engine_ids = {row["engine_id"] for row in trade_blotter}
|
|
assert "earnings_after_close_short_v1" not in engine_ids
|
|
|
|
equity_curve = pq.read_table(run_dir / "artifacts" / "daily_equity_curve.parquet").to_pylist()
|
|
assert any(float(row["net_exposure"]) < 0 for row in equity_curve if float(row["gross_exposure"]) > 0)
|
|
|
|
metrics_summary = json.loads((run_dir / "metrics" / "metrics_summary.json").read_text())
|
|
assert "avg_gross_exposure_pct" in metrics_summary
|
|
assert "avg_net_exposure_pct" in metrics_summary
|
|
assert "days_in_market_pct" in metrics_summary
|
|
|
|
def test_engine_execution_overrides_flow_into_effective_execution_config(self):
|
|
from apps.backtester.run import BacktestRunner
|
|
from libs.backtest.domain import Candidate, ExperimentManifest, StrategyEngineConfig
|
|
|
|
store = _build_multi_engine_store()
|
|
manifest = ExperimentManifest(
|
|
experiment_name="portfolio_exec_overrides",
|
|
dataset_snapshot_id="test_snapshot",
|
|
base_config="configs/backtest/defaults.json",
|
|
overrides={},
|
|
strategy_engines=[
|
|
StrategyEngineConfig(
|
|
engine_id="earnings_same_day_long_trend_v1",
|
|
event_types=["earnings_release"],
|
|
timing_class="same_day",
|
|
direction="long_only",
|
|
entry_timing_policy="reaction_close",
|
|
max_holding_days=12,
|
|
engine_risk_budget_pct=0.25,
|
|
target_atr_multiplier_override=2.5,
|
|
target_1_r_override=3.25,
|
|
target_1_fraction_override=0.33,
|
|
trailing_model_override="pct_10",
|
|
trailing_warmup_days_override=2,
|
|
),
|
|
],
|
|
)
|
|
config = _make_config(strategy_engines=manifest.strategy_engines)
|
|
runner = BacktestRunner(manifest=manifest, config=config, store=store, initial_equity=100_000.0)
|
|
|
|
candidate = Candidate(
|
|
event_id="EVT::SD::LONG",
|
|
symbol="AMD",
|
|
issuer_id="ISSUER::AMD",
|
|
score=0.92,
|
|
sector="Technology",
|
|
event_type="earnings_release",
|
|
event_timestamp=dt.datetime(2026, 1, 6, 21, 0, tzinfo=_UTC),
|
|
event_date=dt.date(2026, 1, 6),
|
|
filing_time_bucket="post_market",
|
|
timing_class="same_day",
|
|
reaction_date=dt.date(2026, 1, 6),
|
|
execution_date=dt.date(2026, 1, 6),
|
|
entry_price_est=122.0,
|
|
avg_dollar_volume=9_000_000.0,
|
|
atr_14=3.0,
|
|
score_bucket="high",
|
|
engine_id="earnings_same_day_long_trend_v1",
|
|
entry_timing_policy="reaction_close",
|
|
shadow_only=False,
|
|
engine_max_holding_days=12,
|
|
engine_risk_budget_pct=0.25,
|
|
engine_target_atr_multiplier=2.5,
|
|
engine_target_1_r=3.25,
|
|
engine_target_1_fraction=0.33,
|
|
engine_trailing_model="pct_10",
|
|
engine_trailing_warmup_days=2,
|
|
trade_direction="long",
|
|
)
|
|
effective_exec = runner._build_effective_execution_config(candidate)
|
|
|
|
assert candidate.engine_id == "earnings_same_day_long_trend_v1"
|
|
assert effective_exec.max_holding_days == 12
|
|
assert effective_exec.target_atr_multiplier == pytest.approx(2.5)
|
|
assert effective_exec.target_1_r == pytest.approx(3.25)
|
|
assert effective_exec.target_1_fraction == pytest.approx(0.33)
|
|
assert effective_exec.trailing_model == "pct_10"
|
|
assert effective_exec.trailing_warmup_days == 2
|
|
|
|
def test_engine_target_overrides_take_precedence_over_tiered_targets(self):
|
|
from apps.backtester.run import BacktestRunner
|
|
from libs.backtest.domain import Candidate, ExecutionConfig, ExperimentManifest, SignalConfig, StrategyEngineConfig
|
|
|
|
store = _build_multi_engine_store()
|
|
manifest = ExperimentManifest(
|
|
experiment_name="portfolio_exec_target_override_precedence",
|
|
dataset_snapshot_id="test_snapshot",
|
|
base_config="configs/backtest/defaults.json",
|
|
overrides={},
|
|
strategy_engines=[
|
|
StrategyEngineConfig(
|
|
engine_id="earnings_same_day_long_trend_v1",
|
|
event_types=["earnings_release"],
|
|
timing_class="same_day",
|
|
direction="long_only",
|
|
entry_timing_policy="reaction_close",
|
|
target_1_r_override=3.5,
|
|
target_1_fraction_override=0.1,
|
|
),
|
|
],
|
|
)
|
|
config = _make_config(strategy_engines=manifest.strategy_engines)
|
|
config = config.model_copy(
|
|
update={
|
|
"signal": SignalConfig(score_threshold=0.5, max_candidates_per_day=5, a_tier_score_threshold=0.8),
|
|
"execution": ExecutionConfig(
|
|
entry_fill_model="next_open",
|
|
exit_fill_model="daily_bar_approximation",
|
|
slippage_bps_base=10.0,
|
|
commission_per_share=0.005,
|
|
same_bar_priority="stop_first_conservative",
|
|
max_holding_days=10,
|
|
target_1_r=1.5,
|
|
target_1_fraction=0.5,
|
|
use_tiered_targets=True,
|
|
a_tier_target_1_r=2.0,
|
|
a_tier_target_1_fraction=0.25,
|
|
non_a_tier_target_1_r=1.25,
|
|
non_a_tier_target_1_fraction=0.6,
|
|
),
|
|
}
|
|
)
|
|
runner = BacktestRunner(manifest=manifest, config=config, store=store, initial_equity=100_000.0)
|
|
|
|
candidate = Candidate(
|
|
event_id="EVT::SD::LONG::TIER",
|
|
symbol="AMD",
|
|
issuer_id="ISSUER::AMD",
|
|
score=0.92,
|
|
sector="Technology",
|
|
event_type="earnings_release",
|
|
event_timestamp=dt.datetime(2026, 1, 6, 21, 0, tzinfo=_UTC),
|
|
event_date=dt.date(2026, 1, 6),
|
|
filing_time_bucket="post_market",
|
|
timing_class="same_day",
|
|
reaction_date=dt.date(2026, 1, 6),
|
|
execution_date=dt.date(2026, 1, 6),
|
|
entry_price_est=122.0,
|
|
avg_dollar_volume=9_000_000.0,
|
|
atr_14=3.0,
|
|
score_bucket="high",
|
|
engine_id="earnings_same_day_long_trend_v1",
|
|
entry_timing_policy="reaction_close",
|
|
shadow_only=False,
|
|
engine_target_1_r=3.5,
|
|
engine_target_1_fraction=0.1,
|
|
trade_direction="long",
|
|
)
|
|
effective_exec = runner._build_effective_execution_config(candidate)
|
|
|
|
assert effective_exec.target_1_r == pytest.approx(3.5)
|
|
assert effective_exec.target_1_fraction == pytest.approx(0.1)
|
|
|
|
def test_tail_exit_adjuster_tightens_execution_for_hot_candidate(self):
|
|
from apps.backtester.run import BacktestRunner
|
|
from libs.backtest.domain import Candidate, ExecutionConfig, ExperimentManifest, RiskConfig
|
|
|
|
store = _build_multi_engine_store()
|
|
manifest = ExperimentManifest(
|
|
experiment_name="portfolio_tail_exit_adjuster",
|
|
dataset_snapshot_id="test_snapshot",
|
|
base_config="configs/backtest/defaults.json",
|
|
overrides={},
|
|
)
|
|
config = _make_config()
|
|
config = config.model_copy(
|
|
update={
|
|
"risk": RiskConfig(
|
|
per_trade_risk_pct=0.01,
|
|
max_daily_new_risk_pct=0.05,
|
|
max_positions=10,
|
|
max_positions_per_sector=5,
|
|
tail_exit_adjuster_enabled=True,
|
|
tail_exit_threshold=0.60,
|
|
tail_exit_min_signals=2,
|
|
tail_exit_max_holding_days=8,
|
|
tail_exit_no_progress_days=1,
|
|
tail_exit_no_progress_r=0.25,
|
|
tail_exit_no_progress_fraction=1.0,
|
|
),
|
|
"execution": ExecutionConfig(
|
|
entry_fill_model="next_open",
|
|
exit_fill_model="daily_bar_approximation",
|
|
slippage_bps_base=10.0,
|
|
commission_per_share=0.005,
|
|
same_bar_priority="stop_first_conservative",
|
|
max_holding_days=12,
|
|
early_failure_no_progress_days=2,
|
|
early_failure_no_progress_r=0.15,
|
|
early_failure_no_progress_fraction=0.5,
|
|
),
|
|
}
|
|
)
|
|
runner = BacktestRunner(manifest=manifest, config=config, store=store, initial_equity=100_000.0)
|
|
|
|
candidate = Candidate(
|
|
event_id="EVT::TAIL::HOT",
|
|
symbol="AMD",
|
|
issuer_id="ISSUER::AMD",
|
|
score=0.88,
|
|
sector="Technology",
|
|
event_type="earnings_release",
|
|
event_timestamp=dt.datetime(2026, 1, 6, 21, 0, tzinfo=_UTC),
|
|
event_date=dt.date(2026, 1, 6),
|
|
filing_time_bucket="post_market",
|
|
timing_class="same_day",
|
|
reaction_date=dt.date(2026, 1, 6),
|
|
execution_date=dt.date(2026, 1, 6),
|
|
entry_price_est=122.0,
|
|
avg_dollar_volume=9_000_000.0,
|
|
atr_14=3.0,
|
|
score_bucket="high",
|
|
engine_id="tail_hot",
|
|
entry_timing_policy="reaction_close",
|
|
shadow_only=False,
|
|
engine_max_holding_days=12,
|
|
trade_direction="long",
|
|
features={
|
|
"reaction_day_return": 0.14,
|
|
"oneoff_penalty": 0.4,
|
|
"pre_event_market_temperature": 1.4,
|
|
"pre_event_entropy_60d": 2.0,
|
|
},
|
|
)
|
|
effective_exec = runner._build_effective_execution_config(candidate)
|
|
|
|
assert effective_exec.max_holding_days == 8
|
|
assert effective_exec.early_failure_no_progress_days == 1
|
|
assert effective_exec.early_failure_no_progress_r == pytest.approx(0.25)
|
|
assert effective_exec.early_failure_no_progress_fraction == pytest.approx(1.0)
|
|
|
|
def test_funding_optimizer_prefers_more_capital_efficient_candidate(self):
|
|
from apps.backtester.run import BacktestRunner
|
|
from libs.backtest.domain import Candidate, DailyPortfolioState, ExperimentManifest
|
|
|
|
store = _build_multi_engine_store()
|
|
manifest = ExperimentManifest(
|
|
experiment_name="portfolio_cap_efficiency_order",
|
|
dataset_snapshot_id="test_snapshot",
|
|
base_config="configs/backtest/defaults.json",
|
|
overrides={},
|
|
)
|
|
config = _make_config()
|
|
config = config.model_copy(update={"strategy_engine_selection_mode": "interleave_cap_efficiency_soft"})
|
|
runner = BacktestRunner(manifest=manifest, config=config, store=store, initial_equity=100_000.0)
|
|
|
|
expensive = Candidate(
|
|
event_id="EVT::EXPENSIVE",
|
|
symbol="EXP",
|
|
issuer_id="ISSUER::EXP",
|
|
score=0.9,
|
|
sector="Technology",
|
|
event_type="earnings_release",
|
|
event_timestamp=dt.datetime(2026, 1, 6, 21, 0, tzinfo=_UTC),
|
|
event_date=dt.date(2026, 1, 6),
|
|
filing_time_bucket="post_market",
|
|
timing_class="same_day",
|
|
reaction_date=dt.date(2026, 1, 6),
|
|
execution_date=dt.date(2026, 1, 6),
|
|
entry_price_est=100.0,
|
|
avg_dollar_volume=20_000_000.0,
|
|
atr_14=2.0,
|
|
score_bucket="high",
|
|
engine_id="core",
|
|
entry_timing_policy="reaction_close",
|
|
shadow_only=False,
|
|
trade_direction="long",
|
|
forced_shares=500,
|
|
)
|
|
cheap = Candidate(
|
|
event_id="EVT::CHEAP",
|
|
symbol="CHP",
|
|
issuer_id="ISSUER::CHP",
|
|
score=0.9,
|
|
sector="Technology",
|
|
event_type="earnings_release",
|
|
event_timestamp=dt.datetime(2026, 1, 6, 21, 0, tzinfo=_UTC),
|
|
event_date=dt.date(2026, 1, 6),
|
|
filing_time_bucket="post_market",
|
|
timing_class="same_day",
|
|
reaction_date=dt.date(2026, 1, 6),
|
|
execution_date=dt.date(2026, 1, 6),
|
|
entry_price_est=100.0,
|
|
avg_dollar_volume=20_000_000.0,
|
|
atr_14=2.0,
|
|
score_bucket="high",
|
|
engine_id="core",
|
|
entry_timing_policy="reaction_close",
|
|
shadow_only=False,
|
|
trade_direction="long",
|
|
forced_shares=50,
|
|
)
|
|
portfolio_state = DailyPortfolioState(
|
|
date=dt.date(2026, 1, 6),
|
|
equity=100_000.0,
|
|
cash_available=100_000.0,
|
|
gross_exposure=0.0,
|
|
net_exposure=0.0,
|
|
reserved_risk_budget=0.0,
|
|
unrealized_pnl=0.0,
|
|
realized_pnl=0.0,
|
|
open_positions=[],
|
|
daily_new_risk_used=0.0,
|
|
peak_equity=100_000.0,
|
|
current_drawdown_pct=0.0,
|
|
)
|
|
|
|
ordered = runner._reorder_candidates_for_funding([expensive, cheap], portfolio_state, macro_data=None)
|
|
assert [candidate.symbol for candidate in ordered[:2]] == ["CHP", "EXP"]
|
|
|
|
def test_attention_gate_filters_engine_candidates(self):
|
|
from apps.backtester.run import BacktestRunner
|
|
from libs.backtest.domain import ExperimentManifest, StrategyEngineConfig
|
|
from libs.backtest.snapshot_store import SnapshotStore
|
|
from libs.oracle_client.models import EntityInfo, EventAttentionResponse, NewsFeatures, WikiFeatures
|
|
|
|
date = dt.date(2026, 1, 7)
|
|
store = SnapshotStore(
|
|
candidates_by_exec_date={
|
|
date: [
|
|
{
|
|
"event_id": "EVT::ATTN::PASS",
|
|
"symbol": "PASS",
|
|
"execution_date": date,
|
|
"entry_date": "2026-01-07",
|
|
"event_date": "2026-01-06",
|
|
"event_close": 100.0,
|
|
"gap_size": 0.12,
|
|
"entry_price": 100.0,
|
|
"score": 0.90,
|
|
"sector": "Technology",
|
|
"event_type": "earnings_release",
|
|
"event_timestamp": "2026-01-06T21:00:00+00:00",
|
|
"filing_time_bucket": "post_market",
|
|
"reaction_date": "2026-01-06",
|
|
"reaction_day_return": -0.12,
|
|
"avg_dollar_volume": 8_000_000.0,
|
|
"atr_14": 4.0,
|
|
},
|
|
{
|
|
"event_id": "EVT::ATTN::FAIL",
|
|
"symbol": "FAIL",
|
|
"execution_date": date,
|
|
"entry_date": "2026-01-07",
|
|
"event_date": "2026-01-06",
|
|
"event_close": 101.0,
|
|
"gap_size": 0.11,
|
|
"entry_price": 101.0,
|
|
"score": 0.89,
|
|
"sector": "Technology",
|
|
"event_type": "earnings_release",
|
|
"event_timestamp": "2026-01-06T21:00:00+00:00",
|
|
"filing_time_bucket": "post_market",
|
|
"reaction_date": "2026-01-06",
|
|
"reaction_day_return": -0.11,
|
|
"avg_dollar_volume": 7_500_000.0,
|
|
"atr_14": 4.0,
|
|
},
|
|
]
|
|
},
|
|
bars_by_symbol_date={
|
|
"PASS": {date: {"date": date, "open": 100.0, "high": 100.0, "low": 95.0, "close": 96.0, "volume": 1_000_000}},
|
|
"FAIL": {date: {"date": date, "open": 101.0, "high": 102.0, "low": 98.0, "close": 99.0, "volume": 900_000}},
|
|
},
|
|
)
|
|
manifest = ExperimentManifest(
|
|
experiment_name="attention_gate_test",
|
|
dataset_snapshot_id="test_snapshot",
|
|
base_config="configs/backtest/defaults.json",
|
|
overrides={},
|
|
strategy_engines=[
|
|
StrategyEngineConfig(
|
|
engine_id="same_day_short_attention",
|
|
event_types=["earnings_release"],
|
|
timing_class="same_day",
|
|
direction="short_only",
|
|
attention_min_wiki_spike_10d=1.2,
|
|
)
|
|
],
|
|
)
|
|
config = _make_config(strategy_engines=manifest.strategy_engines)
|
|
runner = BacktestRunner(manifest=manifest, config=config, store=store)
|
|
|
|
def _fake_attention(candidate):
|
|
spike = 1.5 if candidate.symbol == "PASS" else 0.9
|
|
return EventAttentionResponse(
|
|
ticker=candidate.symbol,
|
|
event_date="2026-01-06",
|
|
entity=EntityInfo(
|
|
ticker=candidate.symbol,
|
|
canonical_name=candidate.symbol,
|
|
resolver_confidence=0.9,
|
|
),
|
|
wiki=WikiFeatures(spike_10d=spike, zscore_20d=1.0),
|
|
news=NewsFeatures(),
|
|
metadata={},
|
|
)
|
|
|
|
runner._attention_service._get_event_attention = _fake_attention # type: ignore[method-assign]
|
|
selected = runner._select_candidates_for_date(date)
|
|
|
|
assert [candidate.symbol for candidate in selected] == ["PASS"]
|
|
assert selected[0].features["attention_wiki_spike_10d"] == pytest.approx(1.5)
|
|
|
|
def test_attention_max_gate_allows_missing_payload(self):
|
|
from apps.backtester.run import BacktestRunner
|
|
from libs.backtest.domain import ExperimentManifest, StrategyEngineConfig
|
|
from libs.backtest.snapshot_store import SnapshotStore
|
|
|
|
date = dt.date(2026, 1, 6)
|
|
store = SnapshotStore(
|
|
candidates_by_exec_date={
|
|
date: [
|
|
{
|
|
"event_id": "EVT::ATTN::KNOWN",
|
|
"symbol": "KNOWN",
|
|
"execution_date": date,
|
|
"entry_date": "2026-01-06",
|
|
"event_date": "2026-01-06",
|
|
"event_close": 100.0,
|
|
"gap_size": 0.12,
|
|
"entry_price": 100.0,
|
|
"score": 0.90,
|
|
"sector": "Technology",
|
|
"event_type": "earnings_release",
|
|
"event_timestamp": "2026-01-06T21:00:00+00:00",
|
|
"filing_time_bucket": "post_market",
|
|
"reaction_date": "2026-01-06",
|
|
"reaction_day_return": 0.15,
|
|
"avg_dollar_volume": 8_000_000.0,
|
|
"atr_14": 4.0,
|
|
},
|
|
{
|
|
"event_id": "EVT::ATTN::MISSING",
|
|
"symbol": "MISSING",
|
|
"execution_date": date,
|
|
"entry_date": "2026-01-06",
|
|
"event_date": "2026-01-06",
|
|
"event_close": 101.0,
|
|
"gap_size": 0.11,
|
|
"entry_price": 101.0,
|
|
"score": 0.89,
|
|
"sector": "Technology",
|
|
"event_type": "earnings_release",
|
|
"event_timestamp": "2026-01-06T21:00:00+00:00",
|
|
"filing_time_bucket": "post_market",
|
|
"reaction_date": "2026-01-06",
|
|
"reaction_day_return": 0.14,
|
|
"avg_dollar_volume": 7_500_000.0,
|
|
"atr_14": 4.0,
|
|
},
|
|
]
|
|
},
|
|
bars_by_symbol_date={
|
|
"KNOWN": {date: {"date": date, "open": 100.0, "high": 105.0, "low": 99.0, "close": 103.0, "volume": 1_000_000}},
|
|
"MISSING": {date: {"date": date, "open": 101.0, "high": 104.0, "low": 100.0, "close": 102.0, "volume": 900_000}},
|
|
},
|
|
)
|
|
manifest = ExperimentManifest(
|
|
experiment_name="attention_max_gate_test",
|
|
dataset_snapshot_id="test_snapshot",
|
|
base_config="configs/backtest/defaults.json",
|
|
overrides={},
|
|
strategy_engines=[
|
|
StrategyEngineConfig(
|
|
engine_id="same_day_long_attention_cap",
|
|
event_types=["earnings_release"],
|
|
timing_class="same_day",
|
|
direction="long_only",
|
|
entry_timing_policy="reaction_close",
|
|
gap_size_min=0.10,
|
|
attention_max_wiki_spike_10d=1.5,
|
|
)
|
|
],
|
|
)
|
|
config = _make_config(strategy_engines=manifest.strategy_engines)
|
|
runner = BacktestRunner(manifest=manifest, config=config, store=store)
|
|
|
|
def _fake_attention(candidate):
|
|
if candidate.symbol == "KNOWN":
|
|
from libs.oracle_client.models import EntityInfo, EventAttentionResponse, NewsFeatures, WikiFeatures
|
|
|
|
return EventAttentionResponse(
|
|
ticker=candidate.symbol,
|
|
event_date="2026-01-06",
|
|
entity=EntityInfo(
|
|
ticker=candidate.symbol,
|
|
canonical_name=candidate.symbol,
|
|
resolver_confidence=0.9,
|
|
),
|
|
wiki=WikiFeatures(spike_10d=1.4, zscore_20d=0.5),
|
|
news=NewsFeatures(),
|
|
metadata={},
|
|
)
|
|
return None
|
|
|
|
runner._attention_service._get_event_attention = _fake_attention # type: ignore[method-assign]
|
|
selected = runner._select_candidates_for_date(date)
|
|
|
|
assert [candidate.symbol for candidate in selected] == ["KNOWN", "MISSING"]
|
|
assert selected[0].features["attention_wiki_spike_10d"] == pytest.approx(1.4)
|
|
assert "attention_wiki_spike_10d" not in selected[1].features
|
|
|
|
def test_attention_path_honors_custom_ranking_fields(self):
|
|
from apps.backtester.run import BacktestRunner
|
|
from libs.backtest.domain import ExperimentManifest, StrategyEngineConfig
|
|
from libs.backtest.snapshot_store import SnapshotStore
|
|
from libs.oracle_client.models import EntityInfo, EventAttentionResponse, NewsFeatures, WikiFeatures
|
|
|
|
date = dt.date(2026, 1, 8)
|
|
store = SnapshotStore(
|
|
candidates_by_exec_date={
|
|
date: [
|
|
{
|
|
"event_id": "EVT::ATTN::DOCWIN",
|
|
"symbol": "DOCWIN",
|
|
"execution_date": date,
|
|
"entry_date": date.isoformat(),
|
|
"event_date": date.isoformat(),
|
|
"event_close": 100.0,
|
|
"entry_price": 100.0,
|
|
"score": 0.90,
|
|
"sector": "Technology",
|
|
"event_type": "earnings_release",
|
|
"event_direction": "bullish",
|
|
"guidance_status": "raised",
|
|
"event_timestamp": "2026-01-08T21:00:00+00:00",
|
|
"filing_time_bucket": "post_market",
|
|
"reaction_date": date.isoformat(),
|
|
"reaction_day_return": 0.09,
|
|
"close_location": 0.72,
|
|
"volume_ratio_20d": 1.8,
|
|
"gap_size": 0.01,
|
|
"parse_confidence_overall": 0.82,
|
|
"parse_confidence_event_direction": 0.82,
|
|
"parse_confidence_guidance": 0.82,
|
|
"document_quality_score": 0.95,
|
|
"signal_strength_score": 0.95,
|
|
"guidance_direction_score": 1.0,
|
|
"oneoff_penalty": 0.05,
|
|
"avg_dollar_volume": 8_000_000.0,
|
|
"atr_14": 4.0,
|
|
},
|
|
{
|
|
"event_id": "EVT::ATTN::CLOSEWIN",
|
|
"symbol": "CLOSEWIN",
|
|
"execution_date": date,
|
|
"entry_date": date.isoformat(),
|
|
"event_date": date.isoformat(),
|
|
"event_close": 101.0,
|
|
"entry_price": 101.0,
|
|
"score": 0.89,
|
|
"sector": "Technology",
|
|
"event_type": "earnings_release",
|
|
"event_direction": "bullish",
|
|
"guidance_status": "raised",
|
|
"event_timestamp": "2026-01-08T21:00:00+00:00",
|
|
"filing_time_bucket": "post_market",
|
|
"reaction_date": date.isoformat(),
|
|
"reaction_day_return": 0.08,
|
|
"close_location": 0.92,
|
|
"volume_ratio_20d": 1.8,
|
|
"gap_size": 0.01,
|
|
"parse_confidence_overall": 0.70,
|
|
"parse_confidence_event_direction": 0.70,
|
|
"parse_confidence_guidance": 0.70,
|
|
"document_quality_score": 0.60,
|
|
"signal_strength_score": 0.60,
|
|
"guidance_direction_score": 0.5,
|
|
"oneoff_penalty": 0.05,
|
|
"avg_dollar_volume": 7_500_000.0,
|
|
"atr_14": 4.0,
|
|
},
|
|
]
|
|
},
|
|
bars_by_symbol_date={},
|
|
)
|
|
manifest = ExperimentManifest(
|
|
experiment_name="attention_ranking_fields_test",
|
|
dataset_snapshot_id="test_snapshot",
|
|
base_config="configs/backtest/defaults.json",
|
|
overrides={},
|
|
strategy_engines=[
|
|
StrategyEngineConfig(
|
|
engine_id="same_day_long_attention",
|
|
event_types=["earnings_release"],
|
|
timing_class="same_day",
|
|
direction="long_only",
|
|
entry_timing_policy="reaction_close",
|
|
attention_max_wiki_spike_10d=2.0,
|
|
)
|
|
],
|
|
)
|
|
config = _make_config(strategy_engines=manifest.strategy_engines)
|
|
config.signal.scoring_model = "return_max_long_v2"
|
|
config.signal.score_threshold = 0.45
|
|
config.signal.max_candidates_per_day = 1
|
|
config.signal.ranking_fields = ["-close_location", "-score"]
|
|
runner = BacktestRunner(manifest=manifest, config=config, store=store)
|
|
|
|
def _fake_attention(candidate):
|
|
return EventAttentionResponse(
|
|
ticker=candidate.symbol,
|
|
event_date=date.isoformat(),
|
|
entity=EntityInfo(
|
|
ticker=candidate.symbol,
|
|
canonical_name=candidate.symbol,
|
|
resolver_confidence=0.9,
|
|
),
|
|
wiki=WikiFeatures(spike_10d=1.0, zscore_20d=0.5),
|
|
news=NewsFeatures(),
|
|
metadata={},
|
|
)
|
|
|
|
runner._get_event_attention = _fake_attention # type: ignore[method-assign]
|
|
selected = runner._select_candidates_for_date(date)
|
|
|
|
assert [candidate.symbol for candidate in selected] == ["CLOSEWIN"]
|
|
|
|
def test_return_max_long_strategy_schedules_add_on_and_writes_benchmark_metrics(self, tmp_path):
|
|
from apps.backtester.run import BacktestRunner
|
|
from libs.backtest.domain import (
|
|
BacktestConfig,
|
|
EventTypeProfile,
|
|
ExecutionConfig,
|
|
ExperimentManifest,
|
|
ReportingConfig,
|
|
RiskConfig,
|
|
SignalConfig,
|
|
StrategyEngineConfig,
|
|
UniverseConfig,
|
|
)
|
|
from libs.backtest.snapshot_store import SnapshotStore
|
|
import json
|
|
import pyarrow.parquet as pq
|
|
|
|
dates = [dt.date(2026, 1, d) for d in [5, 6, 7, 8]]
|
|
store = SnapshotStore(
|
|
candidates_by_exec_date={
|
|
dt.date(2026, 1, 5): [
|
|
{
|
|
"event_id": "EVT::LONG::001",
|
|
"symbol": "NVDA",
|
|
"execution_date": dt.date(2026, 1, 5),
|
|
"entry_date": "2026-01-05",
|
|
"event_date": "2026-01-05",
|
|
"event_close": 101.0,
|
|
"entry_price": 101.0,
|
|
"score": 0.82,
|
|
"sector": "Technology",
|
|
"event_type": "earnings_release",
|
|
"event_timestamp": "2026-01-05T21:00:00+00:00",
|
|
"filing_time_bucket": "post_market",
|
|
"reaction_date": "2026-01-05",
|
|
"reaction_day_return": 0.08,
|
|
"close_location": 0.82,
|
|
"volume_ratio_20d": 1.8,
|
|
"gap_size": 0.01,
|
|
"event_direction": "bullish",
|
|
"guidance_status": "raised",
|
|
"parse_confidence_overall": 0.9,
|
|
"parse_confidence_event_direction": 0.85,
|
|
"parse_confidence_guidance": 0.8,
|
|
"document_quality_score": 0.8,
|
|
"guidance_direction_score": 1.0,
|
|
"oneoff_penalty": 0.1,
|
|
"avg_dollar_volume": 120_000_000.0,
|
|
"market_cap_proxy": 10_000_000_000.0,
|
|
"atr_14": 3.0,
|
|
}
|
|
],
|
|
dt.date(2026, 1, 8): [],
|
|
},
|
|
bars_by_symbol_date={
|
|
"NVDA": {
|
|
dt.date(2026, 1, 5): {"date": dt.date(2026, 1, 5), "open": 100.0, "high": 102.0, "low": 99.0, "close": 101.0, "volume": 2_000_000},
|
|
dt.date(2026, 1, 6): {"date": dt.date(2026, 1, 6), "open": 102.0, "high": 108.0, "low": 101.0, "close": 107.0, "volume": 1_900_000},
|
|
dt.date(2026, 1, 7): {"date": dt.date(2026, 1, 7), "open": 108.0, "high": 112.0, "low": 107.0, "close": 111.0, "volume": 1_800_000},
|
|
dt.date(2026, 1, 8): {"date": dt.date(2026, 1, 8), "open": 112.0, "high": 115.0, "low": 111.0, "close": 114.0, "volume": 1_700_000},
|
|
}
|
|
},
|
|
macro_by_date={
|
|
dates[0]: {"spy_close": 500.0, "spy_sma_20": 495.0, "qqq_close": 420.0, "qqq_sma_20": 415.0},
|
|
dates[1]: {"spy_close": 503.0, "spy_sma_20": 496.0, "qqq_close": 425.0, "qqq_sma_20": 416.0},
|
|
dates[2]: {"spy_close": 505.0, "spy_sma_20": 497.0, "qqq_close": 430.0, "qqq_sma_20": 417.0},
|
|
dates[3]: {"spy_close": 507.0, "spy_sma_20": 498.0, "qqq_close": 435.0, "qqq_sma_20": 418.0},
|
|
},
|
|
)
|
|
|
|
manifest = ExperimentManifest(
|
|
experiment_name="return_max_long_v1",
|
|
dataset_snapshot_id="midlarge-liquid-long-v1",
|
|
base_config="configs/backtest/return_max_long_v1.json",
|
|
overrides={},
|
|
strategy_engines=[
|
|
StrategyEngineConfig(
|
|
engine_id="reaction_close_long_core",
|
|
event_types=["earnings_release", "guidance_update"],
|
|
timing_class="same_day",
|
|
direction="long_only",
|
|
entry_timing_policy="reaction_close",
|
|
max_holding_days=20,
|
|
engine_risk_budget_pct=0.5,
|
|
reaction_day_return_min=0.04,
|
|
reaction_day_return_max=0.12,
|
|
close_location_min=0.70,
|
|
volume_ratio_min=1.5,
|
|
score_threshold_override=0.62,
|
|
),
|
|
StrategyEngineConfig(
|
|
engine_id="delayed_add_on_long",
|
|
event_types=["earnings_release", "guidance_update"],
|
|
timing_class="any",
|
|
direction="long_only",
|
|
entry_timing_policy="next_open",
|
|
synthetic_only=True,
|
|
max_holding_days=15,
|
|
engine_risk_budget_pct=1.0,
|
|
),
|
|
],
|
|
)
|
|
config = BacktestConfig(
|
|
strategy_name="return_max_long_v1",
|
|
dataset_snapshot_id="midlarge-liquid-long-v1",
|
|
universe=UniverseConfig(
|
|
min_price=15.0,
|
|
min_avg_dollar_volume=75_000_000.0,
|
|
min_market_cap_proxy=2_000_000_000.0,
|
|
),
|
|
signal=SignalConfig(
|
|
score_threshold=0.62,
|
|
max_candidates_per_day=6,
|
|
scoring_model="return_max_long_v1",
|
|
a_tier_score_threshold=0.75,
|
|
),
|
|
risk=RiskConfig(
|
|
per_trade_risk_pct=0.004,
|
|
per_trade_risk_pct_a_tier=0.005,
|
|
max_daily_new_risk_pct=0.015,
|
|
max_positions=6,
|
|
max_positions_per_sector=2,
|
|
macro_regime_enabled=True,
|
|
macro_regime_mode="spy_qqq_scaler",
|
|
macro_regime_neutral_size_scaler=0.6,
|
|
macro_regime_risk_off_size_scaler=0.35,
|
|
macro_regime_risk_off_a_tier_only=True,
|
|
stop_atr_multiplier=2.25,
|
|
veto_oneoff_penalty=0.4,
|
|
veto_parse_confidence_min=0.6,
|
|
veto_unknown_direction=True,
|
|
veto_bearish_direction=True,
|
|
),
|
|
execution=ExecutionConfig(
|
|
entry_fill_model="next_open",
|
|
exit_fill_model="daily_bar_approximation",
|
|
slippage_bps_base=0.0,
|
|
commission_per_share=0.0,
|
|
same_bar_priority="stop_first_conservative",
|
|
target_model="fixed_r",
|
|
target_1_r=1.5,
|
|
target_1_fraction=0.5,
|
|
use_tiered_targets=True,
|
|
a_tier_target_1_r=2.0,
|
|
a_tier_target_1_fraction=0.25,
|
|
non_a_tier_target_1_r=1.5,
|
|
non_a_tier_target_1_fraction=0.5,
|
|
trailing_model="pct_6",
|
|
trailing_warmup_days=3,
|
|
max_holding_days=15,
|
|
early_failure_close_below_entry_and_reaction_close=True,
|
|
early_failure_no_progress_days=2,
|
|
early_failure_no_progress_r=0.5,
|
|
early_failure_no_progress_fraction=0.5,
|
|
),
|
|
reporting=ReportingConfig(
|
|
write_trade_blotter=True,
|
|
write_equity_curve=True,
|
|
write_metrics_summary=True,
|
|
generate_plots=False,
|
|
),
|
|
event_type_profiles={
|
|
"earnings_release": EventTypeProfile(enabled=True, max_holding_days_override=20),
|
|
"guidance_update": EventTypeProfile(enabled=True, max_holding_days_override=12),
|
|
"unknown": EventTypeProfile(enabled=False),
|
|
},
|
|
strategy_engines=manifest.strategy_engines,
|
|
strategy_engine_selection_mode="interleave",
|
|
)
|
|
|
|
runner = BacktestRunner(manifest=manifest, config=config, store=store, initial_equity=100_000.0)
|
|
result = runner.run(output_root=tmp_path)
|
|
|
|
run_dir = tmp_path / result.run_id
|
|
trade_blotter = pq.read_table(run_dir / "artifacts" / "trade_blotter.parquet").to_pylist()
|
|
assert any(bool(row["is_add_on"]) for row in trade_blotter)
|
|
add_on_rows = [row for row in trade_blotter if row["engine_id"] == "delayed_add_on_long"]
|
|
assert add_on_rows
|
|
assert all(bool(row["is_add_on"]) for row in add_on_rows)
|
|
assert all(row["parent_position_id"] is not None for row in add_on_rows)
|
|
|
|
metrics_summary = json.loads((run_dir / "metrics" / "metrics_summary.json").read_text())
|
|
assert metrics_summary["qqq_benchmark_return_pct"] is not None
|
|
assert metrics_summary["excess_vs_qqq_pct"] is not None
|
|
assert metrics_summary["long_pnl_contribution_pct"] is not None
|
|
|
|
def test_residual_priority_engine_excludes_selected_names_from_core_queue(self):
|
|
from apps.backtester.run import BacktestRunner
|
|
from libs.backtest.domain import (
|
|
BacktestConfig,
|
|
ExecutionConfig,
|
|
ExperimentManifest,
|
|
ReportingConfig,
|
|
RiskConfig,
|
|
SignalConfig,
|
|
StrategyEngineConfig,
|
|
UniverseConfig,
|
|
)
|
|
from libs.backtest.snapshot_store import SnapshotStore
|
|
|
|
date = dt.date(2026, 1, 5)
|
|
store = SnapshotStore(
|
|
candidates_by_exec_date={
|
|
date: [
|
|
{
|
|
"event_id": "EVT::A",
|
|
"symbol": "AAPL",
|
|
"execution_date": date,
|
|
"entry_date": "2026-01-05",
|
|
"event_date": "2026-01-05",
|
|
"event_close": 101.0,
|
|
"entry_price": 101.0,
|
|
"score": 0.60,
|
|
"sector": "Technology",
|
|
"event_type": "earnings_release",
|
|
"event_timestamp": "2026-01-05T21:00:00+00:00",
|
|
"filing_time_bucket": "regular_hours",
|
|
"reaction_date": "2026-01-05",
|
|
"reaction_day_return": 0.10,
|
|
"close_location": 0.75,
|
|
"volume_ratio_20d": 2.5,
|
|
"gap_size": 0.03,
|
|
"event_direction": "bullish",
|
|
"guidance_status": "raised",
|
|
"parse_confidence_overall": 0.90,
|
|
"parse_confidence_event_direction": 0.85,
|
|
"parse_confidence_guidance": 0.80,
|
|
"document_quality_score": 0.80,
|
|
"guidance_direction_score": 1.0,
|
|
"oneoff_penalty": 0.1,
|
|
"avg_dollar_volume": 120_000_000.0,
|
|
"market_cap_proxy": 10_000_000_000.0,
|
|
"atr_14": 3.0,
|
|
},
|
|
{
|
|
"event_id": "EVT::B",
|
|
"symbol": "MSFT",
|
|
"execution_date": date,
|
|
"entry_date": "2026-01-05",
|
|
"event_date": "2026-01-05",
|
|
"event_close": 201.0,
|
|
"entry_price": 201.0,
|
|
"score": 0.58,
|
|
"sector": "Technology",
|
|
"event_type": "earnings_release",
|
|
"event_timestamp": "2026-01-05T21:00:00+00:00",
|
|
"filing_time_bucket": "regular_hours",
|
|
"reaction_date": "2026-01-05",
|
|
"reaction_day_return": 0.07,
|
|
"close_location": 0.72,
|
|
"volume_ratio_20d": 2.1,
|
|
"gap_size": 0.02,
|
|
"event_direction": "bullish",
|
|
"guidance_status": "raised",
|
|
"parse_confidence_overall": 0.88,
|
|
"parse_confidence_event_direction": 0.84,
|
|
"parse_confidence_guidance": 0.79,
|
|
"document_quality_score": 0.79,
|
|
"guidance_direction_score": 1.0,
|
|
"oneoff_penalty": 0.1,
|
|
"avg_dollar_volume": 150_000_000.0,
|
|
"market_cap_proxy": 20_000_000_000.0,
|
|
"atr_14": 4.0,
|
|
},
|
|
],
|
|
},
|
|
bars_by_symbol_date={},
|
|
)
|
|
|
|
manifest = ExperimentManifest(
|
|
experiment_name="residual_priority_test",
|
|
dataset_snapshot_id="test_snapshot",
|
|
base_config="configs/backtest/return_max_long_v1.json",
|
|
overrides={},
|
|
strategy_engines=[
|
|
StrategyEngineConfig(
|
|
engine_id="priority",
|
|
event_types=["earnings_release"],
|
|
event_directions=["bullish"],
|
|
guidance_statuses=["raised"],
|
|
timing_class="same_day",
|
|
direction="long_only",
|
|
entry_timing_policy="reaction_close",
|
|
reaction_day_return_min=0.08,
|
|
reaction_day_return_max=0.18,
|
|
volume_ratio_min=2.0,
|
|
residual_reserve_selected=True,
|
|
score_threshold_override=0.50,
|
|
),
|
|
StrategyEngineConfig(
|
|
engine_id="core",
|
|
event_types=["earnings_release"],
|
|
timing_class="same_day",
|
|
direction="long_only",
|
|
entry_timing_policy="reaction_close",
|
|
reaction_day_return_min=0.03,
|
|
reaction_day_return_max=0.18,
|
|
volume_ratio_min=1.0,
|
|
score_threshold_override=0.50,
|
|
),
|
|
],
|
|
)
|
|
config = BacktestConfig(
|
|
strategy_name="return_max_long_v1",
|
|
dataset_snapshot_id="test_snapshot",
|
|
universe=UniverseConfig(min_price=15.0, min_avg_dollar_volume=75_000_000.0, min_market_cap_proxy=2_000_000_000.0),
|
|
signal=SignalConfig(
|
|
score_threshold=0.50,
|
|
max_candidates_per_day=2,
|
|
scoring_model="return_max_long_v2",
|
|
),
|
|
risk=RiskConfig(max_positions=4, max_positions_per_sector=4),
|
|
execution=ExecutionConfig(),
|
|
reporting=ReportingConfig(generate_plots=False),
|
|
strategy_engines=manifest.strategy_engines,
|
|
strategy_engine_selection_mode="interleave",
|
|
)
|
|
runner = BacktestRunner(manifest=manifest, config=config, store=store)
|
|
selected = runner._select_candidates_for_date(date)
|
|
|
|
assert [candidate.symbol for candidate in selected] == ["AAPL", "MSFT"]
|
|
assert [candidate.engine_id for candidate in selected] == ["priority", "core"]
|
|
|
|
def test_return_max_long_staged_add_on_can_scale_twice(self, tmp_path):
|
|
from apps.backtester.run import BacktestRunner
|
|
from libs.backtest.domain import (
|
|
BacktestConfig,
|
|
EventTypeProfile,
|
|
ExecutionConfig,
|
|
ExperimentManifest,
|
|
ReportingConfig,
|
|
RiskConfig,
|
|
SignalConfig,
|
|
StrategyEngineConfig,
|
|
UniverseConfig,
|
|
)
|
|
from libs.backtest.snapshot_store import SnapshotStore
|
|
import pyarrow.parquet as pq
|
|
|
|
dates = [dt.date(2026, 1, d) for d in [5, 6, 7, 8, 9, 12]]
|
|
store = SnapshotStore(
|
|
candidates_by_exec_date={
|
|
dt.date(2026, 1, 5): [
|
|
{
|
|
"event_id": "EVT::LONG::STAGED",
|
|
"symbol": "NVDA",
|
|
"execution_date": dt.date(2026, 1, 5),
|
|
"entry_date": "2026-01-05",
|
|
"event_date": "2026-01-05",
|
|
"event_close": 101.0,
|
|
"entry_price": 101.0,
|
|
"score": 0.84,
|
|
"sector": "Technology",
|
|
"event_type": "earnings_release",
|
|
"event_timestamp": "2026-01-05T21:00:00+00:00",
|
|
"filing_time_bucket": "regular_hours",
|
|
"reaction_date": "2026-01-05",
|
|
"reaction_day_return": 0.08,
|
|
"close_location": 0.82,
|
|
"volume_ratio_20d": 1.8,
|
|
"gap_size": 0.01,
|
|
"event_direction": "bullish",
|
|
"guidance_status": "raised",
|
|
"parse_confidence_overall": 0.9,
|
|
"parse_confidence_event_direction": 0.85,
|
|
"parse_confidence_guidance": 0.8,
|
|
"document_quality_score": 0.8,
|
|
"guidance_direction_score": 1.0,
|
|
"oneoff_penalty": 0.1,
|
|
"avg_dollar_volume": 120_000_000.0,
|
|
"market_cap_proxy": 10_000_000_000.0,
|
|
"atr_14": 3.0,
|
|
"reaction_day_high": 102.0,
|
|
}
|
|
],
|
|
dt.date(2026, 1, 12): [],
|
|
},
|
|
bars_by_symbol_date={
|
|
"NVDA": {
|
|
dt.date(2026, 1, 5): {"date": dt.date(2026, 1, 5), "open": 100.0, "high": 102.0, "low": 99.0, "close": 101.0, "volume": 2_000_000},
|
|
dt.date(2026, 1, 6): {"date": dt.date(2026, 1, 6), "open": 102.0, "high": 106.0, "low": 101.0, "close": 105.0, "volume": 1_900_000},
|
|
dt.date(2026, 1, 7): {"date": dt.date(2026, 1, 7), "open": 106.0, "high": 109.0, "low": 105.0, "close": 108.0, "volume": 1_850_000},
|
|
dt.date(2026, 1, 8): {"date": dt.date(2026, 1, 8), "open": 109.0, "high": 114.0, "low": 108.0, "close": 113.0, "volume": 1_800_000},
|
|
dt.date(2026, 1, 9): {"date": dt.date(2026, 1, 9), "open": 114.0, "high": 118.0, "low": 113.0, "close": 117.0, "volume": 1_750_000},
|
|
dt.date(2026, 1, 12): {"date": dt.date(2026, 1, 12), "open": 118.0, "high": 120.0, "low": 116.0, "close": 119.0, "volume": 1_700_000},
|
|
}
|
|
},
|
|
macro_by_date={
|
|
day: {"spy_close": 500.0 + idx, "spy_sma_20": 490.0, "qqq_close": 420.0 + idx, "qqq_sma_20": 410.0}
|
|
for idx, day in enumerate(dates)
|
|
},
|
|
)
|
|
|
|
manifest = ExperimentManifest(
|
|
experiment_name="return_max_long_v1_staged_add_on",
|
|
dataset_snapshot_id="midlarge-liquid-long-v1",
|
|
base_config="configs/backtest/return_max_long_v1.json",
|
|
overrides={},
|
|
strategy_engines=[
|
|
StrategyEngineConfig(
|
|
engine_id="reaction_close_long_core",
|
|
event_types=["earnings_release"],
|
|
timing_class="same_day",
|
|
direction="long_only",
|
|
entry_timing_policy="reaction_close",
|
|
max_holding_days=20,
|
|
engine_risk_budget_pct=1.0,
|
|
reaction_day_return_min=0.04,
|
|
reaction_day_return_max=0.12,
|
|
close_location_min=0.70,
|
|
volume_ratio_min=1.5,
|
|
score_threshold_override=0.62,
|
|
),
|
|
StrategyEngineConfig(
|
|
engine_id="delayed_add_on_long",
|
|
event_types=["earnings_release"],
|
|
timing_class="any",
|
|
direction="long_only",
|
|
entry_timing_policy="next_open",
|
|
synthetic_only=True,
|
|
max_holding_days=15,
|
|
engine_risk_budget_pct=1.0,
|
|
add_on_parent_score_min=0.75,
|
|
add_on_max_parent_days_held=4,
|
|
add_on_schedule_days=[1, 3],
|
|
add_on_progress_r_levels=[0.50, 1.50],
|
|
add_on_max_count=2,
|
|
add_on_size_fraction=0.25,
|
|
add_on_close_location_min=0.70,
|
|
add_on_require_above_reaction_high=True,
|
|
),
|
|
],
|
|
)
|
|
config = BacktestConfig(
|
|
strategy_name="return_max_long_v1",
|
|
dataset_snapshot_id="midlarge-liquid-long-v1",
|
|
universe=UniverseConfig(min_price=15.0, min_avg_dollar_volume=75_000_000.0, min_market_cap_proxy=2_000_000_000.0),
|
|
signal=SignalConfig(score_threshold=0.62, max_candidates_per_day=6, scoring_model="return_max_long_v1", a_tier_score_threshold=0.75),
|
|
risk=RiskConfig(
|
|
per_trade_risk_pct=0.01,
|
|
per_trade_risk_pct_a_tier=0.01,
|
|
max_daily_new_risk_pct=0.20,
|
|
max_positions=6,
|
|
max_positions_per_sector=3,
|
|
macro_regime_enabled=True,
|
|
macro_regime_mode="spy_qqq_scaler",
|
|
macro_regime_neutral_size_scaler=1.0,
|
|
macro_regime_risk_off_size_scaler=1.0,
|
|
macro_regime_risk_off_a_tier_only=False,
|
|
stop_atr_multiplier=2.25,
|
|
veto_oneoff_penalty=0.4,
|
|
veto_parse_confidence_min=0.6,
|
|
veto_unknown_direction=False,
|
|
veto_bearish_direction=True,
|
|
),
|
|
execution=ExecutionConfig(
|
|
entry_fill_model="next_open",
|
|
exit_fill_model="daily_bar_approximation",
|
|
slippage_bps_base=0.0,
|
|
commission_per_share=0.0,
|
|
same_bar_priority="stop_first_conservative",
|
|
target_model="fixed_r",
|
|
target_1_r=10.0,
|
|
target_1_fraction=0.0,
|
|
trailing_model=None,
|
|
trailing_warmup_days=20,
|
|
max_holding_days=20,
|
|
),
|
|
reporting=ReportingConfig(write_trade_blotter=True, write_equity_curve=True, write_metrics_summary=True, generate_plots=False),
|
|
event_type_profiles={"earnings_release": EventTypeProfile(enabled=True, max_holding_days_override=20)},
|
|
strategy_engines=manifest.strategy_engines,
|
|
strategy_engine_selection_mode="interleave",
|
|
)
|
|
|
|
runner = BacktestRunner(manifest=manifest, config=config, store=store, initial_equity=100_000.0)
|
|
result = runner.run(output_root=tmp_path)
|
|
run_dir = tmp_path / result.run_id
|
|
trade_blotter = pq.read_table(run_dir / "artifacts" / "trade_blotter.parquet").to_pylist()
|
|
add_on_rows = [row for row in trade_blotter if row["engine_id"] == "delayed_add_on_long"]
|
|
assert len(add_on_rows) >= 2
|
|
assert sorted(int(row["shares"]) for row in add_on_rows)[0] > 0
|
|
assert all(bool(row["is_add_on"]) for row in add_on_rows)
|
|
|
|
def test_return_max_long_add_on_parent_score_gate_blocks_synthetic_child(self, tmp_path):
|
|
from apps.backtester.run import BacktestRunner
|
|
from libs.backtest.domain import (
|
|
BacktestConfig,
|
|
EventTypeProfile,
|
|
ExecutionConfig,
|
|
ExperimentManifest,
|
|
ReportingConfig,
|
|
RiskConfig,
|
|
SignalConfig,
|
|
StrategyEngineConfig,
|
|
UniverseConfig,
|
|
)
|
|
from libs.backtest.snapshot_store import SnapshotStore
|
|
import pyarrow.parquet as pq
|
|
|
|
dates = [dt.date(2026, 1, d) for d in [5, 6, 7, 8]]
|
|
store = SnapshotStore(
|
|
candidates_by_exec_date={
|
|
dt.date(2026, 1, 5): [
|
|
{
|
|
"event_id": "EVT::LONG::002",
|
|
"symbol": "NVDA",
|
|
"execution_date": dt.date(2026, 1, 5),
|
|
"entry_date": "2026-01-05",
|
|
"event_date": "2026-01-05",
|
|
"event_close": 101.0,
|
|
"entry_price": 101.0,
|
|
"score": 0.70,
|
|
"sector": "Technology",
|
|
"event_type": "earnings_release",
|
|
"event_timestamp": "2026-01-05T21:00:00+00:00",
|
|
"filing_time_bucket": "post_market",
|
|
"reaction_date": "2026-01-05",
|
|
"reaction_day_return": 0.08,
|
|
"close_location": 0.82,
|
|
"volume_ratio_20d": 1.8,
|
|
"gap_size": 0.01,
|
|
"event_direction": "bullish",
|
|
"guidance_status": "raised",
|
|
"parse_confidence_overall": 0.9,
|
|
"document_quality_score": 0.8,
|
|
"guidance_direction_score": 1.0,
|
|
"oneoff_penalty": 0.1,
|
|
"avg_dollar_volume": 120_000_000.0,
|
|
"market_cap_proxy": 10_000_000_000.0,
|
|
"atr_14": 3.0,
|
|
}
|
|
],
|
|
dt.date(2026, 1, 8): [],
|
|
},
|
|
bars_by_symbol_date={
|
|
"NVDA": {
|
|
dt.date(2026, 1, 5): {"date": dt.date(2026, 1, 5), "open": 100.0, "high": 102.0, "low": 99.0, "close": 101.0, "volume": 2_000_000},
|
|
dt.date(2026, 1, 6): {"date": dt.date(2026, 1, 6), "open": 102.0, "high": 108.0, "low": 101.0, "close": 107.0, "volume": 1_900_000},
|
|
dt.date(2026, 1, 7): {"date": dt.date(2026, 1, 7), "open": 108.0, "high": 112.0, "low": 107.0, "close": 111.0, "volume": 1_800_000},
|
|
dt.date(2026, 1, 8): {"date": dt.date(2026, 1, 8), "open": 112.0, "high": 115.0, "low": 111.0, "close": 114.0, "volume": 1_700_000},
|
|
}
|
|
},
|
|
macro_by_date={
|
|
dates[0]: {"spy_close": 500.0, "spy_sma_20": 495.0, "qqq_close": 420.0, "qqq_sma_20": 415.0},
|
|
dates[1]: {"spy_close": 503.0, "spy_sma_20": 496.0, "qqq_close": 425.0, "qqq_sma_20": 416.0},
|
|
dates[2]: {"spy_close": 505.0, "spy_sma_20": 497.0, "qqq_close": 430.0, "qqq_sma_20": 417.0},
|
|
dates[3]: {"spy_close": 507.0, "spy_sma_20": 498.0, "qqq_close": 435.0, "qqq_sma_20": 418.0},
|
|
},
|
|
)
|
|
|
|
manifest = ExperimentManifest(
|
|
experiment_name="return_max_long_v1_add_on_parent_gate",
|
|
dataset_snapshot_id="midlarge-liquid-long-v1",
|
|
base_config="configs/backtest/return_max_long_v1.json",
|
|
overrides={},
|
|
strategy_engines=[
|
|
StrategyEngineConfig(
|
|
engine_id="reaction_close_long_core",
|
|
event_types=["earnings_release"],
|
|
timing_class="same_day",
|
|
direction="long_only",
|
|
entry_timing_policy="reaction_close",
|
|
max_holding_days=20,
|
|
engine_risk_budget_pct=1.0,
|
|
reaction_day_return_min=0.04,
|
|
reaction_day_return_max=0.12,
|
|
close_location_min=0.70,
|
|
volume_ratio_min=1.5,
|
|
score_threshold_override=0.62,
|
|
),
|
|
StrategyEngineConfig(
|
|
engine_id="delayed_add_on_long",
|
|
event_types=["earnings_release"],
|
|
timing_class="any",
|
|
direction="long_only",
|
|
entry_timing_policy="next_open",
|
|
synthetic_only=True,
|
|
max_holding_days=15,
|
|
engine_risk_budget_pct=1.0,
|
|
add_on_parent_score_min=0.80,
|
|
),
|
|
],
|
|
)
|
|
config = BacktestConfig(
|
|
strategy_name="return_max_long_v1",
|
|
dataset_snapshot_id="midlarge-liquid-long-v1",
|
|
universe=UniverseConfig(
|
|
min_price=15.0,
|
|
min_avg_dollar_volume=75_000_000.0,
|
|
min_market_cap_proxy=2_000_000_000.0,
|
|
),
|
|
signal=SignalConfig(
|
|
score_threshold=0.62,
|
|
max_candidates_per_day=6,
|
|
scoring_model="return_max_long_v1",
|
|
a_tier_score_threshold=0.75,
|
|
),
|
|
risk=RiskConfig(
|
|
per_trade_risk_pct=0.004,
|
|
per_trade_risk_pct_a_tier=0.005,
|
|
max_daily_new_risk_pct=0.05,
|
|
max_positions=6,
|
|
max_positions_per_sector=2,
|
|
macro_regime_enabled=True,
|
|
macro_regime_mode="spy_qqq_scaler",
|
|
macro_regime_neutral_size_scaler=0.6,
|
|
macro_regime_risk_off_size_scaler=0.35,
|
|
macro_regime_risk_off_a_tier_only=False,
|
|
stop_atr_multiplier=2.25,
|
|
veto_oneoff_penalty=0.4,
|
|
veto_parse_confidence_min=0.6,
|
|
veto_unknown_direction=False,
|
|
veto_bearish_direction=True,
|
|
),
|
|
execution=ExecutionConfig(
|
|
entry_fill_model="next_open",
|
|
exit_fill_model="daily_bar_approximation",
|
|
slippage_bps_base=0.0,
|
|
commission_per_share=0.0,
|
|
same_bar_priority="stop_first_conservative",
|
|
target_model="fixed_r",
|
|
target_1_r=1.5,
|
|
target_1_fraction=0.5,
|
|
use_tiered_targets=True,
|
|
a_tier_target_1_r=2.0,
|
|
a_tier_target_1_fraction=0.25,
|
|
non_a_tier_target_1_r=1.5,
|
|
non_a_tier_target_1_fraction=0.5,
|
|
trailing_model="pct_6",
|
|
trailing_warmup_days=3,
|
|
max_holding_days=15,
|
|
early_failure_close_below_entry_and_reaction_close=True,
|
|
early_failure_no_progress_days=2,
|
|
early_failure_no_progress_r=0.5,
|
|
early_failure_no_progress_fraction=0.5,
|
|
),
|
|
reporting=ReportingConfig(
|
|
write_trade_blotter=True,
|
|
write_equity_curve=True,
|
|
write_metrics_summary=True,
|
|
generate_plots=False,
|
|
),
|
|
event_type_profiles={
|
|
"earnings_release": EventTypeProfile(enabled=True, max_holding_days_override=20),
|
|
"unknown": EventTypeProfile(enabled=False),
|
|
},
|
|
strategy_engines=manifest.strategy_engines,
|
|
strategy_engine_selection_mode="interleave",
|
|
)
|
|
|
|
runner = BacktestRunner(manifest=manifest, config=config, store=store, initial_equity=100_000.0)
|
|
result = runner.run(output_root=tmp_path)
|
|
|
|
run_dir = tmp_path / result.run_id
|
|
trade_blotter = pq.read_table(run_dir / "artifacts" / "trade_blotter.parquet").to_pylist()
|
|
assert not any(bool(row["is_add_on"]) for row in trade_blotter)
|
|
assert not any(row["engine_id"] == "delayed_add_on_long" for row in trade_blotter)
|
|
|
|
def test_return_max_long_add_on_parent_engine_gate_blocks_non_matching_parent(self, tmp_path):
|
|
from apps.backtester.run import BacktestRunner
|
|
from libs.backtest.domain import (
|
|
BacktestConfig,
|
|
EventTypeProfile,
|
|
ExecutionConfig,
|
|
ExperimentManifest,
|
|
ReportingConfig,
|
|
RiskConfig,
|
|
SignalConfig,
|
|
StrategyEngineConfig,
|
|
UniverseConfig,
|
|
)
|
|
from libs.backtest.snapshot_store import SnapshotStore
|
|
import pyarrow.parquet as pq
|
|
|
|
dates = [dt.date(2026, 1, d) for d in [5, 6, 7, 8]]
|
|
store = SnapshotStore(
|
|
candidates_by_exec_date={
|
|
dt.date(2026, 1, 5): [
|
|
{
|
|
"event_id": "EVT::LONG::003",
|
|
"symbol": "NVDA",
|
|
"execution_date": dt.date(2026, 1, 5),
|
|
"entry_date": "2026-01-05",
|
|
"event_date": "2026-01-05",
|
|
"event_close": 101.0,
|
|
"entry_price": 101.0,
|
|
"score": 0.72,
|
|
"sector": "Technology",
|
|
"event_type": "earnings_release",
|
|
"event_timestamp": "2026-01-05T21:00:00+00:00",
|
|
"filing_time_bucket": "post_market",
|
|
"reaction_date": "2026-01-05",
|
|
"reaction_day_return": 0.08,
|
|
"close_location": 0.82,
|
|
"volume_ratio_20d": 1.8,
|
|
"gap_size": 0.01,
|
|
"event_direction": "bullish",
|
|
"guidance_status": "raised",
|
|
"parse_confidence_overall": 0.9,
|
|
"document_quality_score": 0.8,
|
|
"guidance_direction_score": 1.0,
|
|
"oneoff_penalty": 0.1,
|
|
"avg_dollar_volume": 120_000_000.0,
|
|
"market_cap_proxy": 10_000_000_000.0,
|
|
"atr_14": 3.0,
|
|
}
|
|
],
|
|
dt.date(2026, 1, 8): [],
|
|
},
|
|
bars_by_symbol_date={
|
|
"NVDA": {
|
|
dt.date(2026, 1, 5): {"date": dt.date(2026, 1, 5), "open": 100.0, "high": 102.0, "low": 99.0, "close": 101.0, "volume": 2_000_000},
|
|
dt.date(2026, 1, 6): {"date": dt.date(2026, 1, 6), "open": 102.0, "high": 108.0, "low": 101.0, "close": 107.0, "volume": 1_900_000},
|
|
dt.date(2026, 1, 7): {"date": dt.date(2026, 1, 7), "open": 108.0, "high": 112.0, "low": 107.0, "close": 111.0, "volume": 1_800_000},
|
|
dt.date(2026, 1, 8): {"date": dt.date(2026, 1, 8), "open": 112.0, "high": 115.0, "low": 111.0, "close": 114.0, "volume": 1_700_000},
|
|
}
|
|
},
|
|
macro_by_date={
|
|
dates[0]: {"spy_close": 500.0, "spy_sma_20": 495.0, "qqq_close": 420.0, "qqq_sma_20": 415.0},
|
|
dates[1]: {"spy_close": 503.0, "spy_sma_20": 496.0, "qqq_close": 425.0, "qqq_sma_20": 416.0},
|
|
dates[2]: {"spy_close": 505.0, "spy_sma_20": 497.0, "qqq_close": 430.0, "qqq_sma_20": 417.0},
|
|
dates[3]: {"spy_close": 507.0, "spy_sma_20": 498.0, "qqq_close": 435.0, "qqq_sma_20": 418.0},
|
|
},
|
|
)
|
|
|
|
manifest = ExperimentManifest(
|
|
experiment_name="return_max_long_v1_add_on_parent_engine_gate",
|
|
dataset_snapshot_id="midlarge-liquid-long-v1",
|
|
base_config="configs/backtest/return_max_long_v1.json",
|
|
overrides={},
|
|
strategy_engines=[
|
|
StrategyEngineConfig(
|
|
engine_id="next_open_controlled_long",
|
|
event_types=["earnings_release"],
|
|
timing_class="any",
|
|
direction="long_only",
|
|
entry_timing_policy="next_open",
|
|
max_holding_days=15,
|
|
engine_risk_budget_pct=1.0,
|
|
reaction_day_return_min=0.03,
|
|
reaction_day_return_max=0.10,
|
|
close_location_min=0.65,
|
|
volume_ratio_min=1.5,
|
|
score_threshold_override=0.60,
|
|
),
|
|
StrategyEngineConfig(
|
|
engine_id="delayed_add_on_long",
|
|
event_types=["earnings_release"],
|
|
timing_class="any",
|
|
direction="long_only",
|
|
entry_timing_policy="next_open",
|
|
synthetic_only=True,
|
|
max_holding_days=15,
|
|
engine_risk_budget_pct=1.0,
|
|
add_on_parent_engine_ids=["reaction_close_long_core"],
|
|
),
|
|
],
|
|
)
|
|
config = BacktestConfig(
|
|
strategy_name="return_max_long_v1",
|
|
dataset_snapshot_id="midlarge-liquid-long-v1",
|
|
universe=UniverseConfig(
|
|
min_price=15.0,
|
|
min_avg_dollar_volume=75_000_000.0,
|
|
min_market_cap_proxy=2_000_000_000.0,
|
|
),
|
|
signal=SignalConfig(
|
|
score_threshold=0.55,
|
|
max_candidates_per_day=6,
|
|
scoring_model="return_max_long_v2",
|
|
a_tier_score_threshold=0.68,
|
|
),
|
|
risk=RiskConfig(
|
|
per_trade_risk_pct=0.0045,
|
|
per_trade_risk_pct_a_tier=0.0055,
|
|
max_daily_new_risk_pct=0.04,
|
|
max_positions=10,
|
|
max_positions_per_sector=3,
|
|
macro_regime_enabled=True,
|
|
macro_regime_mode="spy_qqq_scaler",
|
|
macro_regime_neutral_size_scaler=0.85,
|
|
macro_regime_risk_off_size_scaler=0.60,
|
|
macro_regime_risk_off_a_tier_only=False,
|
|
stop_atr_multiplier=2.25,
|
|
veto_oneoff_penalty=0.4,
|
|
veto_parse_confidence_min=0.6,
|
|
veto_unknown_direction=False,
|
|
veto_bearish_direction=True,
|
|
),
|
|
execution=ExecutionConfig(
|
|
entry_fill_model="next_open",
|
|
exit_fill_model="daily_bar_approximation",
|
|
slippage_bps_base=0.0,
|
|
commission_per_share=0.0,
|
|
same_bar_priority="stop_first_conservative",
|
|
target_model="fixed_r",
|
|
target_1_r=1.5,
|
|
target_1_fraction=0.5,
|
|
use_tiered_targets=True,
|
|
a_tier_target_1_r=2.0,
|
|
a_tier_target_1_fraction=0.25,
|
|
non_a_tier_target_1_r=1.5,
|
|
non_a_tier_target_1_fraction=0.5,
|
|
trailing_model="pct_6",
|
|
trailing_warmup_days=3,
|
|
max_holding_days=15,
|
|
early_failure_close_below_entry_and_reaction_close=True,
|
|
early_failure_no_progress_days=2,
|
|
early_failure_no_progress_r=0.5,
|
|
early_failure_no_progress_fraction=0.5,
|
|
),
|
|
reporting=ReportingConfig(
|
|
write_trade_blotter=True,
|
|
write_equity_curve=True,
|
|
write_metrics_summary=True,
|
|
generate_plots=False,
|
|
),
|
|
event_type_profiles={
|
|
"earnings_release": EventTypeProfile(enabled=True, max_holding_days_override=20),
|
|
"unknown": EventTypeProfile(enabled=False),
|
|
},
|
|
strategy_engines=manifest.strategy_engines,
|
|
strategy_engine_selection_mode="interleave",
|
|
)
|
|
|
|
runner = BacktestRunner(manifest=manifest, config=config, store=store, initial_equity=100_000.0)
|
|
result = runner.run(output_root=tmp_path)
|
|
|
|
run_dir = tmp_path / result.run_id
|
|
trade_blotter = pq.read_table(run_dir / "artifacts" / "trade_blotter.parquet").to_pylist()
|
|
assert any(row["engine_id"] == "next_open_controlled_long" for row in trade_blotter)
|
|
assert not any(bool(row["is_add_on"]) for row in trade_blotter)
|
|
assert not any(row["engine_id"] == "delayed_add_on_long" for row in trade_blotter)
|
|
|
|
def test_evaluate_pending_open_exit_can_schedule_early_pop_giveback(self):
|
|
from apps.backtester.run import BacktestRunner
|
|
from libs.backtest.domain import (
|
|
BacktestConfig,
|
|
Candidate,
|
|
ExecutionConfig,
|
|
ExperimentManifest,
|
|
OpenPosition,
|
|
PlannedOrder,
|
|
ReportingConfig,
|
|
RiskConfig,
|
|
SignalConfig,
|
|
UniverseConfig,
|
|
)
|
|
from libs.backtest.snapshot_store import SnapshotStore
|
|
|
|
store = SnapshotStore(candidates_by_exec_date={}, bars_by_symbol_date={})
|
|
manifest = ExperimentManifest(
|
|
experiment_name="giveback_eval",
|
|
dataset_snapshot_id="midlarge-liquid-long-v1",
|
|
base_config="configs/backtest/return_max_long_v1.json",
|
|
overrides={},
|
|
)
|
|
config = BacktestConfig(
|
|
strategy_name="return_max_long_v1",
|
|
dataset_snapshot_id="midlarge-liquid-long-v1",
|
|
universe=UniverseConfig(min_price=15.0, min_avg_dollar_volume=75_000_000.0),
|
|
signal=SignalConfig(score_threshold=0.45, max_candidates_per_day=10, scoring_model="return_max_long_v2"),
|
|
risk=RiskConfig(per_trade_risk_pct=0.05, max_daily_new_risk_pct=0.5, max_positions=20),
|
|
execution=ExecutionConfig(
|
|
slippage_bps_base=0.0,
|
|
commission_per_share=0.0,
|
|
max_holding_days=25,
|
|
early_pop_giveback_days_min=3,
|
|
early_pop_giveback_days_max=5,
|
|
early_pop_giveback_trigger_r=0.75,
|
|
early_pop_giveback_min_r=0.40,
|
|
early_pop_giveback_from_peak_pct=0.035,
|
|
early_pop_giveback_fraction=1.0,
|
|
),
|
|
reporting=ReportingConfig(generate_plots=False),
|
|
)
|
|
runner = BacktestRunner(manifest=manifest, config=config, store=store, initial_equity=100_000.0)
|
|
candidate = Candidate(
|
|
event_id="EVT::GIVEBACK",
|
|
symbol="AAPL",
|
|
score=0.7,
|
|
sector="Technology",
|
|
event_type="earnings_release",
|
|
event_timestamp=dt.datetime(2026, 1, 5, 21, 0, tzinfo=_UTC),
|
|
event_date=dt.date(2026, 1, 5),
|
|
filing_time_bucket="regular_hours",
|
|
reaction_date=dt.date(2026, 1, 5),
|
|
execution_date=dt.date(2026, 1, 6),
|
|
entry_price_est=100.0,
|
|
avg_dollar_volume=120_000_000.0,
|
|
atr_14=2.0,
|
|
score_bucket="high",
|
|
engine_id="reaction_close_long_core",
|
|
entry_timing_policy="reaction_close",
|
|
features={"event_direction": "bullish", "guidance_status": "raised"},
|
|
)
|
|
plan = PlannedOrder(
|
|
candidate=candidate,
|
|
shares=100,
|
|
entry_price_limit=100.0,
|
|
stop_price=95.0,
|
|
target_price=112.0,
|
|
risk_dollars=500.0,
|
|
)
|
|
position = OpenPosition(
|
|
position_id="POS::1",
|
|
plan=plan,
|
|
entry_date=dt.date(2026, 1, 6),
|
|
entry_price=100.0,
|
|
entry_fill_slippage_bps=0.0,
|
|
current_stop=95.0,
|
|
target_price=112.0,
|
|
peak_price=110.0,
|
|
shares_open=100,
|
|
shares_total=100,
|
|
days_held=4,
|
|
)
|
|
payload = runner._evaluate_pending_open_exit(
|
|
position=position,
|
|
bar={"date": dt.date(2026, 1, 10), "open": 101.0, "high": 103.0, "low": 100.0, "close": 101.0},
|
|
execution_config=config.execution,
|
|
date=dt.date(2026, 1, 10),
|
|
)
|
|
assert payload is not None
|
|
assert payload["reason"] == "GIVEBACK"
|
|
assert payload["fraction"] == pytest.approx(1.0)
|
|
|
|
def test_global_score_selection_uses_custom_ranking_fields(self, tmp_path):
|
|
from apps.backtester.run import BacktestRunner
|
|
from libs.backtest.domain import (
|
|
BacktestConfig,
|
|
ExecutionConfig,
|
|
ExperimentManifest,
|
|
ReportingConfig,
|
|
RiskConfig,
|
|
SignalConfig,
|
|
StrategyEngineConfig,
|
|
UniverseConfig,
|
|
)
|
|
from libs.backtest.snapshot_store import SnapshotStore
|
|
|
|
model_path = tmp_path / "ranker.json"
|
|
model_path.write_text(
|
|
json.dumps(
|
|
{
|
|
"model_type": "bucket_blend_v1",
|
|
"global_mean": 0.0,
|
|
"features": [
|
|
{
|
|
"name": "direction_guidance_combo",
|
|
"weight": 1.0,
|
|
"values": {
|
|
"bullish|raised": 0.25,
|
|
"unknown|not_provided": 0.05,
|
|
},
|
|
}
|
|
],
|
|
}
|
|
)
|
|
)
|
|
|
|
event_date = dt.date(2026, 1, 7)
|
|
store = SnapshotStore(
|
|
candidates_by_exec_date={
|
|
event_date: [
|
|
{
|
|
"event_id": "EVT::PRIORITY",
|
|
"symbol": "AAPL",
|
|
"execution_date": event_date,
|
|
"entry_date": event_date.isoformat(),
|
|
"event_date": event_date.isoformat(),
|
|
"event_close": 100.0,
|
|
"entry_price": 100.0,
|
|
"score": 0.60,
|
|
"sector": "Technology",
|
|
"event_type": "earnings_release",
|
|
"event_direction": "bullish",
|
|
"guidance_status": "raised",
|
|
"event_timestamp": "2026-01-07T21:00:00+00:00",
|
|
"filing_time_bucket": "post_market",
|
|
"reaction_date": event_date.isoformat(),
|
|
"reaction_day_return": 0.10,
|
|
"close_location": 0.72,
|
|
"volume_ratio_20d": 2.2,
|
|
"gap_size": 0.03,
|
|
"avg_dollar_volume": 5_000_000.0,
|
|
"atr_14": 3.0,
|
|
},
|
|
{
|
|
"event_id": "EVT::CORE",
|
|
"symbol": "MSFT",
|
|
"execution_date": event_date,
|
|
"entry_date": event_date.isoformat(),
|
|
"event_date": event_date.isoformat(),
|
|
"event_close": 200.0,
|
|
"entry_price": 200.0,
|
|
"score": 0.90,
|
|
"sector": "Technology",
|
|
"event_type": "earnings_release",
|
|
"event_direction": "unknown",
|
|
"guidance_status": "not_provided",
|
|
"event_timestamp": "2026-01-07T21:00:00+00:00",
|
|
"filing_time_bucket": "post_market",
|
|
"reaction_date": event_date.isoformat(),
|
|
"reaction_day_return": 0.09,
|
|
"close_location": 0.74,
|
|
"volume_ratio_20d": 2.1,
|
|
"gap_size": 0.02,
|
|
"avg_dollar_volume": 8_000_000.0,
|
|
"atr_14": 4.0,
|
|
},
|
|
]
|
|
},
|
|
bars_by_symbol_date={},
|
|
)
|
|
manifest = ExperimentManifest(
|
|
experiment_name="test_global_ranker",
|
|
dataset_snapshot_id="test_snapshot",
|
|
base_config="configs/backtest/defaults.json",
|
|
overrides={},
|
|
strategy_engines=[
|
|
StrategyEngineConfig(
|
|
engine_id="priority",
|
|
event_types=["earnings_release"],
|
|
event_directions=["bullish"],
|
|
guidance_statuses=["raised"],
|
|
timing_class="same_day",
|
|
direction="long_only",
|
|
entry_timing_policy="reaction_close",
|
|
enabled=True,
|
|
),
|
|
StrategyEngineConfig(
|
|
engine_id="core",
|
|
event_types=["earnings_release"],
|
|
timing_class="same_day",
|
|
direction="long_only",
|
|
entry_timing_policy="reaction_close",
|
|
enabled=True,
|
|
),
|
|
],
|
|
)
|
|
config = BacktestConfig(
|
|
strategy_name="test_strategy",
|
|
dataset_snapshot_id="test_snapshot",
|
|
universe=UniverseConfig(min_price=5.0, min_avg_dollar_volume=100_000),
|
|
signal=SignalConfig(
|
|
score_threshold=0.5,
|
|
max_candidates_per_day=1,
|
|
ranking_model_path=str(model_path),
|
|
ranking_fields=["-learned_rank_score", "-score"],
|
|
),
|
|
risk=RiskConfig(
|
|
per_trade_risk_pct=0.01,
|
|
max_daily_new_risk_pct=0.05,
|
|
max_positions=10,
|
|
max_positions_per_sector=5,
|
|
),
|
|
execution=ExecutionConfig(max_holding_days=5),
|
|
reporting=ReportingConfig(),
|
|
strategy_engines=manifest.strategy_engines,
|
|
strategy_engine_selection_mode="global_score",
|
|
)
|
|
|
|
runner = BacktestRunner(manifest=manifest, config=config, store=store, initial_equity=100_000.0)
|
|
candidates = runner._select_candidates_for_date(event_date)
|
|
|
|
assert [candidate.symbol for candidate in candidates] == ["AAPL"]
|
|
|
|
def test_interleave_head_score_prioritizes_stronger_core_head(self, tmp_path):
|
|
from apps.backtester.run import BacktestRunner
|
|
from libs.backtest.domain import (
|
|
BacktestConfig,
|
|
ExecutionConfig,
|
|
ExperimentManifest,
|
|
ReportingConfig,
|
|
RiskConfig,
|
|
SignalConfig,
|
|
StrategyEngineConfig,
|
|
UniverseConfig,
|
|
)
|
|
from libs.backtest.snapshot_store import SnapshotStore
|
|
|
|
event_date = dt.date(2026, 1, 7)
|
|
store = SnapshotStore(
|
|
candidates_by_exec_date={
|
|
event_date: [
|
|
{
|
|
"event_id": "EVT::PRIORITY",
|
|
"symbol": "AAPL",
|
|
"execution_date": event_date,
|
|
"entry_date": event_date.isoformat(),
|
|
"event_date": event_date.isoformat(),
|
|
"event_close": 100.0,
|
|
"entry_price": 100.0,
|
|
"score": 0.60,
|
|
"sector": "Technology",
|
|
"event_type": "earnings_release",
|
|
"event_direction": "bullish",
|
|
"guidance_status": "raised",
|
|
"event_timestamp": "2026-01-07T21:00:00+00:00",
|
|
"filing_time_bucket": "post_market",
|
|
"reaction_date": event_date.isoformat(),
|
|
"reaction_day_return": 0.10,
|
|
"close_location": 0.72,
|
|
"volume_ratio_20d": 2.2,
|
|
"gap_size": 0.03,
|
|
"avg_dollar_volume": 5_000_000.0,
|
|
"atr_14": 3.0,
|
|
},
|
|
{
|
|
"event_id": "EVT::CORE",
|
|
"symbol": "MSFT",
|
|
"execution_date": event_date,
|
|
"entry_date": event_date.isoformat(),
|
|
"event_date": event_date.isoformat(),
|
|
"event_close": 200.0,
|
|
"entry_price": 200.0,
|
|
"score": 0.90,
|
|
"sector": "Technology",
|
|
"event_type": "earnings_release",
|
|
"event_direction": "unknown",
|
|
"guidance_status": "not_provided",
|
|
"event_timestamp": "2026-01-07T21:00:00+00:00",
|
|
"filing_time_bucket": "post_market",
|
|
"reaction_date": event_date.isoformat(),
|
|
"reaction_day_return": 0.09,
|
|
"close_location": 0.74,
|
|
"volume_ratio_20d": 2.1,
|
|
"gap_size": 0.02,
|
|
"avg_dollar_volume": 8_000_000.0,
|
|
"atr_14": 4.0,
|
|
},
|
|
]
|
|
},
|
|
bars_by_symbol_date={},
|
|
)
|
|
manifest = ExperimentManifest(
|
|
experiment_name="test_interleave_head_score",
|
|
dataset_snapshot_id="test_snapshot",
|
|
base_config="configs/backtest/defaults.json",
|
|
overrides={},
|
|
strategy_engines=[
|
|
StrategyEngineConfig(
|
|
engine_id="priority",
|
|
event_types=["earnings_release"],
|
|
event_directions=["bullish"],
|
|
guidance_statuses=["raised"],
|
|
timing_class="same_day",
|
|
direction="long_only",
|
|
entry_timing_policy="reaction_close",
|
|
enabled=True,
|
|
),
|
|
StrategyEngineConfig(
|
|
engine_id="core",
|
|
event_types=["earnings_release"],
|
|
timing_class="same_day",
|
|
direction="long_only",
|
|
entry_timing_policy="reaction_close",
|
|
enabled=True,
|
|
),
|
|
],
|
|
)
|
|
config = BacktestConfig(
|
|
strategy_name="test_strategy",
|
|
dataset_snapshot_id="test_snapshot",
|
|
universe=UniverseConfig(min_price=5.0, min_avg_dollar_volume=100_000),
|
|
signal=SignalConfig(
|
|
score_threshold=0.5,
|
|
max_candidates_per_day=2,
|
|
),
|
|
risk=RiskConfig(
|
|
per_trade_risk_pct=0.01,
|
|
max_daily_new_risk_pct=0.05,
|
|
max_positions=10,
|
|
max_positions_per_sector=5,
|
|
),
|
|
execution=ExecutionConfig(max_holding_days=5),
|
|
reporting=ReportingConfig(),
|
|
strategy_engines=manifest.strategy_engines,
|
|
strategy_engine_selection_mode="interleave_head_score",
|
|
)
|
|
|
|
runner = BacktestRunner(manifest=manifest, config=config, store=store, initial_equity=100_000.0)
|
|
candidates = runner._select_candidates_for_date(event_date)
|
|
|
|
assert [candidate.symbol for candidate in candidates] == ["MSFT", "AAPL"]
|
|
|
|
|
|
@pytest.mark.integration
|
|
class TestWalkForwardRunIntegration:
|
|
def test_run_walk_forward_generates_train_and_test_folds(self, tmp_path, monkeypatch):
|
|
from apps.backtester.run import run_walk_forward
|
|
from libs.backtest.domain import ExperimentManifest
|
|
from libs.backtest.snapshot_store import SnapshotStore
|
|
|
|
trading_dates = [
|
|
dt.date(2026, 1, 5),
|
|
dt.date(2026, 1, 6),
|
|
dt.date(2026, 1, 7),
|
|
dt.date(2026, 1, 8),
|
|
dt.date(2026, 1, 9),
|
|
dt.date(2026, 1, 12),
|
|
dt.date(2026, 1, 13),
|
|
]
|
|
bars: dict[str, dict[dt.date, dict[str, float | dt.date | int]]] = {}
|
|
candidates_by_date: dict[dt.date, list[dict[str, object]]] = {}
|
|
for idx, exec_date in enumerate(trading_dates):
|
|
symbol = f"SYM{idx}"
|
|
candidates_by_date[exec_date] = [
|
|
{
|
|
"event_id": f"EVT::{idx}",
|
|
"symbol": symbol,
|
|
"execution_date": exec_date,
|
|
"entry_date": exec_date.isoformat(),
|
|
"event_date": (exec_date - dt.timedelta(days=1)).isoformat(),
|
|
"event_close": 100.0,
|
|
"entry_price": 100.0,
|
|
"score": 0.9,
|
|
"sector": "Technology",
|
|
"event_type": "earnings_release",
|
|
"event_timestamp": f"{(exec_date - dt.timedelta(days=1)).isoformat()}T21:00:00+00:00",
|
|
"filing_time_bucket": "post_market",
|
|
"reaction_date": exec_date.isoformat(),
|
|
"avg_dollar_volume": 5_000_000.0,
|
|
"atr_14": 2.0,
|
|
}
|
|
]
|
|
bars[symbol] = {
|
|
exec_date: {
|
|
"date": exec_date,
|
|
"open": 100.0,
|
|
"high": 106.0,
|
|
"low": 99.0,
|
|
"close": 105.0,
|
|
"volume": 1_000_000,
|
|
},
|
|
exec_date + dt.timedelta(days=1): {
|
|
"date": exec_date + dt.timedelta(days=1),
|
|
"open": 105.0,
|
|
"high": 110.0,
|
|
"low": 104.0,
|
|
"close": 109.0,
|
|
"volume": 900_000,
|
|
},
|
|
}
|
|
|
|
store_map = {
|
|
"train": SnapshotStore(
|
|
candidates_by_exec_date={d: candidates_by_date[d] for d in trading_dates[:3]},
|
|
bars_by_symbol_date=bars,
|
|
),
|
|
"valid": SnapshotStore(
|
|
candidates_by_exec_date={d: candidates_by_date[d] for d in trading_dates[3:5]},
|
|
bars_by_symbol_date=bars,
|
|
),
|
|
"test": SnapshotStore(
|
|
candidates_by_exec_date={d: candidates_by_date[d] for d in trading_dates[5:]},
|
|
bars_by_symbol_date=bars,
|
|
),
|
|
}
|
|
|
|
def _fake_build_store(manifest, config, split_name, snapshot_dir_override=None):
|
|
return store_map[split_name]
|
|
|
|
merged_store = SnapshotStore(
|
|
candidates_by_exec_date=candidates_by_date,
|
|
bars_by_symbol_date=bars,
|
|
)
|
|
|
|
monkeypatch.setattr("apps.backtester.run._build_store", _fake_build_store)
|
|
monkeypatch.setattr("apps.backtester.run._build_merged_snapshot_store", lambda *args, **kwargs: merged_store)
|
|
|
|
manifest = ExperimentManifest(
|
|
experiment_name="wf_test_exp",
|
|
dataset_snapshot_id="test_snapshot",
|
|
base_config="configs/backtest/defaults.json",
|
|
overrides={},
|
|
)
|
|
config = _make_config()
|
|
|
|
summary = run_walk_forward(
|
|
manifest=manifest,
|
|
config=config,
|
|
snapshot_dir_override=None,
|
|
initial_equity=100_000.0,
|
|
output_root=str(tmp_path),
|
|
train_days=3,
|
|
test_days=2,
|
|
step_days=2,
|
|
)
|
|
|
|
assert summary.fold_count == 2
|
|
assert summary.folds[0].train_end < summary.folds[0].test_start
|
|
assert summary.folds[0].train_run_id
|
|
assert summary.folds[0].test_run_id
|
|
assert summary.test_aggregate.mean_return_pct is not None
|
|
assert (tmp_path / "walk_forward" / "walk_forward_summary.json").exists()
|
|
|
|
def test_run_walk_forward_respects_explicit_date_window(self, tmp_path, monkeypatch):
|
|
from apps.backtester.run import run_walk_forward
|
|
from libs.backtest.domain import ExperimentManifest
|
|
from libs.backtest.snapshot_store import SnapshotStore
|
|
|
|
trading_dates = [
|
|
dt.date(2026, 1, 5),
|
|
dt.date(2026, 1, 6),
|
|
dt.date(2026, 1, 7),
|
|
dt.date(2026, 1, 8),
|
|
dt.date(2026, 1, 9),
|
|
dt.date(2026, 1, 12),
|
|
dt.date(2026, 1, 13),
|
|
]
|
|
bars: dict[str, dict[dt.date, dict[str, float | dt.date | int]]] = {}
|
|
candidates_by_date: dict[dt.date, list[dict[str, object]]] = {}
|
|
for idx, exec_date in enumerate(trading_dates):
|
|
symbol = f"WFS{idx}"
|
|
candidates_by_date[exec_date] = [
|
|
{
|
|
"event_id": f"WFS::{idx}",
|
|
"symbol": symbol,
|
|
"execution_date": exec_date,
|
|
"entry_date": exec_date.isoformat(),
|
|
"event_date": (exec_date - dt.timedelta(days=1)).isoformat(),
|
|
"event_close": 100.0,
|
|
"entry_price": 100.0,
|
|
"score": 0.9,
|
|
"sector": "Technology",
|
|
"event_type": "earnings_release",
|
|
"event_timestamp": f"{(exec_date - dt.timedelta(days=1)).isoformat()}T21:00:00+00:00",
|
|
"filing_time_bucket": "post_market",
|
|
"reaction_date": exec_date.isoformat(),
|
|
"avg_dollar_volume": 5_000_000.0,
|
|
"atr_14": 2.0,
|
|
}
|
|
]
|
|
bars[symbol] = {
|
|
exec_date: {
|
|
"date": exec_date,
|
|
"open": 100.0,
|
|
"high": 106.0,
|
|
"low": 99.0,
|
|
"close": 105.0,
|
|
"volume": 1_000_000,
|
|
},
|
|
exec_date + dt.timedelta(days=1): {
|
|
"date": exec_date + dt.timedelta(days=1),
|
|
"open": 105.0,
|
|
"high": 110.0,
|
|
"low": 104.0,
|
|
"close": 109.0,
|
|
"volume": 900_000,
|
|
},
|
|
}
|
|
|
|
merged_store = SnapshotStore(
|
|
candidates_by_exec_date=candidates_by_date,
|
|
bars_by_symbol_date=bars,
|
|
)
|
|
|
|
monkeypatch.setattr("apps.backtester.run._build_merged_snapshot_store", lambda *args, **kwargs: merged_store)
|
|
|
|
manifest = ExperimentManifest(
|
|
experiment_name="wf_window_test_exp",
|
|
dataset_snapshot_id="test_snapshot",
|
|
base_config="configs/backtest/defaults.json",
|
|
overrides={},
|
|
)
|
|
config = _make_config()
|
|
|
|
full_summary = run_walk_forward(
|
|
manifest=manifest,
|
|
config=config,
|
|
snapshot_dir_override=None,
|
|
initial_equity=100_000.0,
|
|
output_root=str(tmp_path / "full"),
|
|
train_days=3,
|
|
test_days=2,
|
|
step_days=2,
|
|
)
|
|
window_summary = run_walk_forward(
|
|
manifest=manifest,
|
|
config=config,
|
|
snapshot_dir_override=None,
|
|
initial_equity=100_000.0,
|
|
output_root=str(tmp_path / "window"),
|
|
train_days=3,
|
|
test_days=2,
|
|
step_days=2,
|
|
start_date=dt.date(2026, 1, 7),
|
|
end_date=dt.date(2026, 1, 13),
|
|
)
|
|
|
|
assert window_summary.fold_count < full_summary.fold_count
|
|
assert window_summary.folds[0].train_start >= dt.date(2026, 1, 7)
|
|
assert window_summary.folds[-1].test_end <= dt.date(2026, 1, 13)
|
|
|
|
|
|
@pytest.mark.integration
|
|
class TestRobustnessMatrixRunIntegration:
|
|
def test_run_robustness_matrix_generates_summary(self, tmp_path, monkeypatch):
|
|
from apps.backtester.run import run_robustness_matrix
|
|
from libs.backtest.domain import ExperimentManifest
|
|
from libs.backtest.snapshot_store import SnapshotStore
|
|
|
|
trading_dates = [
|
|
dt.date(2026, 1, 5),
|
|
dt.date(2026, 1, 6),
|
|
dt.date(2026, 1, 7),
|
|
dt.date(2026, 1, 8),
|
|
dt.date(2026, 1, 9),
|
|
dt.date(2026, 1, 12),
|
|
dt.date(2026, 1, 13),
|
|
]
|
|
bars: dict[str, dict[dt.date, dict[str, float | dt.date | int]]] = {}
|
|
candidates_by_date: dict[dt.date, list[dict[str, object]]] = {}
|
|
for idx, exec_date in enumerate(trading_dates):
|
|
symbol = f"RM{idx}"
|
|
candidates_by_date[exec_date] = [
|
|
{
|
|
"event_id": f"RM::{idx}",
|
|
"symbol": symbol,
|
|
"execution_date": exec_date,
|
|
"entry_date": exec_date.isoformat(),
|
|
"event_date": (exec_date - dt.timedelta(days=1)).isoformat(),
|
|
"event_close": 100.0,
|
|
"entry_price": 100.0,
|
|
"score": 0.9,
|
|
"sector": "Technology",
|
|
"event_type": "earnings_release",
|
|
"event_timestamp": f"{(exec_date - dt.timedelta(days=1)).isoformat()}T21:00:00+00:00",
|
|
"filing_time_bucket": "post_market",
|
|
"reaction_date": exec_date.isoformat(),
|
|
"avg_dollar_volume": 5_000_000.0,
|
|
"atr_14": 2.0,
|
|
}
|
|
]
|
|
bars[symbol] = {
|
|
exec_date: {
|
|
"date": exec_date,
|
|
"open": 100.0,
|
|
"high": 106.0,
|
|
"low": 99.0,
|
|
"close": 105.0,
|
|
"volume": 1_000_000,
|
|
},
|
|
exec_date + dt.timedelta(days=1): {
|
|
"date": exec_date + dt.timedelta(days=1),
|
|
"open": 105.0,
|
|
"high": 109.0,
|
|
"low": 104.0,
|
|
"close": 108.0,
|
|
"volume": 900_000,
|
|
},
|
|
}
|
|
|
|
store_map = {
|
|
"train": SnapshotStore(
|
|
candidates_by_exec_date={d: candidates_by_date[d] for d in trading_dates[:3]},
|
|
bars_by_symbol_date=bars,
|
|
),
|
|
"valid": SnapshotStore(
|
|
candidates_by_exec_date={d: candidates_by_date[d] for d in trading_dates[3:5]},
|
|
bars_by_symbol_date=bars,
|
|
),
|
|
"test": SnapshotStore(
|
|
candidates_by_exec_date={d: candidates_by_date[d] for d in trading_dates[5:]},
|
|
bars_by_symbol_date=bars,
|
|
),
|
|
}
|
|
|
|
def _fake_build_store(manifest, config, split_name, snapshot_dir_override=None):
|
|
return store_map[split_name]
|
|
|
|
monkeypatch.setattr("apps.backtester.run._build_store", _fake_build_store)
|
|
monkeypatch.setattr("apps.backtester.run._build_merged_snapshot_store", lambda *args, **kwargs: merged_store)
|
|
|
|
manifest = ExperimentManifest(
|
|
experiment_name="rm_test_exp",
|
|
dataset_snapshot_id="test_snapshot",
|
|
base_config="configs/backtest/defaults.json",
|
|
overrides={},
|
|
)
|
|
config = _make_config()
|
|
|
|
summary = run_robustness_matrix(
|
|
manifest=manifest,
|
|
config=config,
|
|
snapshot_dir_override=None,
|
|
initial_equity=100_000.0,
|
|
output_root=str(tmp_path),
|
|
horizons_days=[2, 4],
|
|
step_days=2,
|
|
)
|
|
|
|
assert summary.overall_window_count > 0
|
|
assert [item.horizon_days for item in summary.horizon_summaries] == [2, 4]
|
|
assert all(item.window_count > 0 for item in summary.horizon_summaries)
|
|
assert (tmp_path / "robustness_matrix" / "robustness_matrix_summary.json").exists()
|
|
|
|
def test_run_robustness_matrix_respects_explicit_date_window(self, tmp_path, monkeypatch):
|
|
from apps.backtester.run import run_robustness_matrix
|
|
from libs.backtest.domain import ExperimentManifest
|
|
from libs.backtest.snapshot_store import SnapshotStore
|
|
|
|
trading_dates = [
|
|
dt.date(2026, 1, 5),
|
|
dt.date(2026, 1, 6),
|
|
dt.date(2026, 1, 7),
|
|
dt.date(2026, 1, 8),
|
|
dt.date(2026, 1, 9),
|
|
dt.date(2026, 1, 12),
|
|
dt.date(2026, 1, 13),
|
|
]
|
|
bars: dict[str, dict[dt.date, dict[str, float | dt.date | int]]] = {}
|
|
candidates_by_date: dict[dt.date, list[dict[str, object]]] = {}
|
|
for idx, exec_date in enumerate(trading_dates):
|
|
symbol = f"RMW{idx}"
|
|
candidates_by_date[exec_date] = [
|
|
{
|
|
"event_id": f"RMW::{idx}",
|
|
"symbol": symbol,
|
|
"execution_date": exec_date,
|
|
"entry_date": exec_date.isoformat(),
|
|
"event_date": (exec_date - dt.timedelta(days=1)).isoformat(),
|
|
"event_close": 100.0,
|
|
"entry_price": 100.0,
|
|
"score": 0.9,
|
|
"sector": "Technology",
|
|
"event_type": "earnings_release",
|
|
"event_timestamp": f"{(exec_date - dt.timedelta(days=1)).isoformat()}T21:00:00+00:00",
|
|
"filing_time_bucket": "post_market",
|
|
"reaction_date": exec_date.isoformat(),
|
|
"avg_dollar_volume": 5_000_000.0,
|
|
"atr_14": 2.0,
|
|
}
|
|
]
|
|
bars[symbol] = {
|
|
exec_date: {
|
|
"date": exec_date,
|
|
"open": 100.0,
|
|
"high": 106.0,
|
|
"low": 99.0,
|
|
"close": 105.0,
|
|
"volume": 1_000_000,
|
|
},
|
|
exec_date + dt.timedelta(days=1): {
|
|
"date": exec_date + dt.timedelta(days=1),
|
|
"open": 105.0,
|
|
"high": 109.0,
|
|
"low": 104.0,
|
|
"close": 108.0,
|
|
"volume": 900_000,
|
|
},
|
|
}
|
|
|
|
merged_store = SnapshotStore(
|
|
candidates_by_exec_date=candidates_by_date,
|
|
bars_by_symbol_date=bars,
|
|
)
|
|
|
|
monkeypatch.setattr("apps.backtester.run._build_merged_snapshot_store", lambda *args, **kwargs: merged_store)
|
|
|
|
manifest = ExperimentManifest(
|
|
experiment_name="rm_window_test_exp",
|
|
dataset_snapshot_id="test_snapshot",
|
|
base_config="configs/backtest/defaults.json",
|
|
overrides={},
|
|
)
|
|
config = _make_config()
|
|
|
|
full_summary = run_robustness_matrix(
|
|
manifest=manifest,
|
|
config=config,
|
|
snapshot_dir_override=None,
|
|
initial_equity=100_000.0,
|
|
output_root=str(tmp_path / "full"),
|
|
horizons_days=[2, 4],
|
|
step_days=2,
|
|
)
|
|
window_summary = run_robustness_matrix(
|
|
manifest=manifest,
|
|
config=config,
|
|
snapshot_dir_override=None,
|
|
initial_equity=100_000.0,
|
|
output_root=str(tmp_path / "window"),
|
|
horizons_days=[2, 4],
|
|
step_days=2,
|
|
start_date=dt.date(2026, 1, 7),
|
|
end_date=dt.date(2026, 1, 13),
|
|
)
|
|
|
|
assert window_summary.overall_window_count < full_summary.overall_window_count
|
|
assert all(item.window_count > 0 for item in window_summary.horizon_summaries)
|
|
|
|
|
|
@pytest.mark.integration
|
|
class TestRobustnessMatrixRunIntegration:
|
|
def test_run_robustness_matrix_generates_summary(self, tmp_path, monkeypatch):
|
|
from apps.backtester.run import run_robustness_matrix
|
|
from libs.backtest.domain import ExperimentManifest
|
|
from libs.backtest.snapshot_store import SnapshotStore
|
|
|
|
trading_dates = [
|
|
dt.date(2026, 1, 5),
|
|
dt.date(2026, 1, 6),
|
|
dt.date(2026, 1, 7),
|
|
dt.date(2026, 1, 8),
|
|
dt.date(2026, 1, 9),
|
|
dt.date(2026, 1, 12),
|
|
dt.date(2026, 1, 13),
|
|
]
|
|
bars: dict[str, dict[dt.date, dict[str, float | dt.date | int]]] = {}
|
|
candidates_by_date: dict[dt.date, list[dict[str, object]]] = {}
|
|
for idx, exec_date in enumerate(trading_dates):
|
|
symbol = f"RB{idx}"
|
|
candidates_by_date[exec_date] = [
|
|
{
|
|
"event_id": f"EVT::RB::{idx}",
|
|
"symbol": symbol,
|
|
"execution_date": exec_date,
|
|
"entry_date": exec_date.isoformat(),
|
|
"event_date": (exec_date - dt.timedelta(days=1)).isoformat(),
|
|
"event_close": 100.0,
|
|
"entry_price": 100.0,
|
|
"score": 0.9,
|
|
"sector": "Technology",
|
|
"event_type": "earnings_release",
|
|
"event_timestamp": f"{(exec_date - dt.timedelta(days=1)).isoformat()}T21:00:00+00:00",
|
|
"filing_time_bucket": "post_market",
|
|
"reaction_date": exec_date.isoformat(),
|
|
"avg_dollar_volume": 5_000_000.0,
|
|
"atr_14": 2.0,
|
|
}
|
|
]
|
|
bars[symbol] = {
|
|
exec_date: {
|
|
"date": exec_date,
|
|
"open": 100.0,
|
|
"high": 106.0,
|
|
"low": 99.0,
|
|
"close": 105.0,
|
|
"volume": 1_000_000,
|
|
},
|
|
exec_date + dt.timedelta(days=1): {
|
|
"date": exec_date + dt.timedelta(days=1),
|
|
"open": 105.0,
|
|
"high": 110.0,
|
|
"low": 104.0,
|
|
"close": 109.0,
|
|
"volume": 900_000,
|
|
},
|
|
}
|
|
|
|
merged_store = SnapshotStore(
|
|
candidates_by_exec_date=candidates_by_date,
|
|
bars_by_symbol_date=bars,
|
|
)
|
|
|
|
def _fake_build_store(manifest, config, split_name, snapshot_dir_override=None):
|
|
return merged_store
|
|
|
|
monkeypatch.setattr("apps.backtester.run._build_store", _fake_build_store)
|
|
monkeypatch.setattr("apps.backtester.run._build_merged_snapshot_store", lambda *args, **kwargs: merged_store)
|
|
|
|
manifest = ExperimentManifest(
|
|
experiment_name="rb_test_exp",
|
|
dataset_snapshot_id="test_snapshot",
|
|
base_config="configs/backtest/defaults.json",
|
|
overrides={},
|
|
)
|
|
config = _make_config()
|
|
|
|
summary = run_robustness_matrix(
|
|
manifest=manifest,
|
|
config=config,
|
|
snapshot_dir_override=None,
|
|
initial_equity=100_000.0,
|
|
output_root=str(tmp_path),
|
|
horizons_days=[2, 4],
|
|
step_days=2,
|
|
)
|
|
|
|
assert summary.horizons_days == [2, 4]
|
|
assert summary.step_days == 2
|
|
assert summary.overall_window_count > 0
|
|
assert len(summary.horizon_summaries) == 2
|
|
assert all(item.window_count > 0 for item in summary.horizon_summaries)
|
|
assert all(item.mean_return_pct is not None for item in summary.horizon_summaries)
|
|
assert (tmp_path / "robustness_matrix" / "robustness_matrix_summary.json").exists()
|
|
|
|
def test_run_robustness_matrix_respects_explicit_date_window(self, tmp_path, monkeypatch):
|
|
from apps.backtester.run import run_robustness_matrix
|
|
from libs.backtest.domain import ExperimentManifest
|
|
from libs.backtest.snapshot_store import SnapshotStore
|
|
|
|
trading_dates = [
|
|
dt.date(2026, 1, 5),
|
|
dt.date(2026, 1, 6),
|
|
dt.date(2026, 1, 7),
|
|
dt.date(2026, 1, 8),
|
|
dt.date(2026, 1, 9),
|
|
dt.date(2026, 1, 12),
|
|
dt.date(2026, 1, 13),
|
|
]
|
|
bars: dict[str, dict[dt.date, dict[str, float | dt.date | int]]] = {}
|
|
candidates_by_date: dict[dt.date, list[dict[str, object]]] = {}
|
|
for idx, exec_date in enumerate(trading_dates):
|
|
symbol = f"RMWX{idx}"
|
|
candidates_by_date[exec_date] = [
|
|
{
|
|
"event_id": f"RMWX::{idx}",
|
|
"symbol": symbol,
|
|
"execution_date": exec_date,
|
|
"entry_date": exec_date.isoformat(),
|
|
"event_date": (exec_date - dt.timedelta(days=1)).isoformat(),
|
|
"event_close": 100.0,
|
|
"entry_price": 100.0,
|
|
"score": 0.9,
|
|
"sector": "Technology",
|
|
"event_type": "earnings_release",
|
|
"event_timestamp": f"{(exec_date - dt.timedelta(days=1)).isoformat()}T21:00:00+00:00",
|
|
"filing_time_bucket": "post_market",
|
|
"reaction_date": exec_date.isoformat(),
|
|
"avg_dollar_volume": 5_000_000.0,
|
|
"atr_14": 2.0,
|
|
}
|
|
]
|
|
bars[symbol] = {
|
|
exec_date: {
|
|
"date": exec_date,
|
|
"open": 100.0,
|
|
"high": 106.0,
|
|
"low": 99.0,
|
|
"close": 105.0,
|
|
"volume": 1_000_000,
|
|
},
|
|
exec_date + dt.timedelta(days=1): {
|
|
"date": exec_date + dt.timedelta(days=1),
|
|
"open": 105.0,
|
|
"high": 109.0,
|
|
"low": 104.0,
|
|
"close": 108.0,
|
|
"volume": 900_000,
|
|
},
|
|
}
|
|
|
|
merged_store = SnapshotStore(
|
|
candidates_by_exec_date=candidates_by_date,
|
|
bars_by_symbol_date=bars,
|
|
)
|
|
|
|
monkeypatch.setattr("apps.backtester.run._build_merged_snapshot_store", lambda *args, **kwargs: merged_store)
|
|
|
|
manifest = ExperimentManifest(
|
|
experiment_name="rm_window_test_exp",
|
|
dataset_snapshot_id="test_snapshot",
|
|
base_config="configs/backtest/defaults.json",
|
|
overrides={},
|
|
)
|
|
config = _make_config()
|
|
|
|
full_summary = run_robustness_matrix(
|
|
manifest=manifest,
|
|
config=config,
|
|
snapshot_dir_override=None,
|
|
initial_equity=100_000.0,
|
|
output_root=str(tmp_path / "full"),
|
|
horizons_days=[2, 4],
|
|
step_days=2,
|
|
)
|
|
window_summary = run_robustness_matrix(
|
|
manifest=manifest,
|
|
config=config,
|
|
snapshot_dir_override=None,
|
|
initial_equity=100_000.0,
|
|
output_root=str(tmp_path / "window"),
|
|
horizons_days=[2, 4],
|
|
step_days=2,
|
|
start_date=dt.date(2026, 1, 7),
|
|
end_date=dt.date(2026, 1, 13),
|
|
)
|
|
|
|
assert window_summary.overall_window_count < full_summary.overall_window_count
|
|
assert all(item.window_count > 0 for item in window_summary.horizon_summaries)
|
|
|
|
def test_delayed_entry_can_use_shadow_only_source_candidates(self, tmp_path):
|
|
from apps.backtester.run import BacktestRunner
|
|
import pyarrow.parquet as pq
|
|
from libs.backtest.domain import (
|
|
BacktestConfig,
|
|
EventTypeProfile,
|
|
ExperimentManifest,
|
|
ExecutionConfig,
|
|
ReportingConfig,
|
|
RiskConfig,
|
|
SignalConfig,
|
|
StrategyEngineConfig,
|
|
UniverseConfig,
|
|
)
|
|
from libs.backtest.snapshot_store import SnapshotStore
|
|
|
|
dates = [dt.date(2026, 1, d) for d in [5, 6, 7, 8, 9]]
|
|
store = SnapshotStore(
|
|
candidates_by_exec_date={
|
|
dates[0]: [
|
|
{
|
|
"event_id": "EVT::DRIFT::001",
|
|
"symbol": "AAPL",
|
|
"execution_date": dates[0],
|
|
"entry_date": dates[0].isoformat(),
|
|
"event_date": dates[0].isoformat(),
|
|
"event_close": 100.0,
|
|
"entry_price": 100.0,
|
|
"score": 0.82,
|
|
"sector": "Technology",
|
|
"event_type": "earnings_release",
|
|
"event_timestamp": "2026-01-05T21:00:00+00:00",
|
|
"filing_time_bucket": "post_market",
|
|
"reaction_date": dates[0].isoformat(),
|
|
"reaction_day_return": 0.08,
|
|
"close_location": 0.82,
|
|
"volume_ratio_20d": 2.2,
|
|
"gap_size": 0.01,
|
|
"avg_dollar_volume": 9_000_000.0,
|
|
"atr_14": 3.0,
|
|
}
|
|
],
|
|
dates[1]: [],
|
|
dates[2]: [],
|
|
dates[3]: [],
|
|
dates[4]: [],
|
|
},
|
|
bars_by_symbol_date={
|
|
"AAPL": {
|
|
dates[0]: {"date": dates[0], "open": 98.0, "high": 101.0, "low": 97.0, "close": 100.0, "volume": 1_000_000},
|
|
dates[1]: {"date": dates[1], "open": 101.0, "high": 104.0, "low": 100.0, "close": 103.0, "volume": 900_000},
|
|
dates[2]: {"date": dates[2], "open": 103.0, "high": 106.0, "low": 102.0, "close": 105.0, "volume": 850_000},
|
|
dates[3]: {"date": dates[3], "open": 105.0, "high": 109.0, "low": 104.0, "close": 108.0, "volume": 800_000},
|
|
dates[4]: {"date": dates[4], "open": 109.0, "high": 111.0, "low": 108.0, "close": 110.0, "volume": 780_000},
|
|
}
|
|
},
|
|
)
|
|
|
|
manifest = ExperimentManifest(
|
|
experiment_name="return_max_long_v1_shadow_delayed",
|
|
dataset_snapshot_id="test_snapshot",
|
|
base_config="configs/backtest/return_max_long_v1.json",
|
|
overrides={},
|
|
strategy_engines=[
|
|
StrategyEngineConfig(
|
|
engine_id="reaction_close_long_core",
|
|
event_types=["earnings_release"],
|
|
timing_class="same_day",
|
|
direction="long_only",
|
|
entry_timing_policy="reaction_close",
|
|
max_holding_days=10,
|
|
engine_risk_budget_pct=1.0,
|
|
reaction_day_return_min=0.04,
|
|
reaction_day_return_max=0.20,
|
|
close_location_min=0.70,
|
|
volume_ratio_min=1.5,
|
|
score_threshold_override=0.60,
|
|
shadow_only=True,
|
|
),
|
|
StrategyEngineConfig(
|
|
engine_id="delayed_primary_long_drift",
|
|
event_types=["earnings_release"],
|
|
timing_class="any",
|
|
direction="long_only",
|
|
entry_timing_policy="next_open",
|
|
max_holding_days=10,
|
|
engine_risk_budget_pct=1.0,
|
|
delayed_entry_lookback_days=3,
|
|
delayed_entry_source_engine_ids=["reaction_close_long_core"],
|
|
delayed_entry_min_drift_pct=0.05,
|
|
delayed_entry_close_location_min=0.50,
|
|
score_threshold_override=0.0,
|
|
per_trade_risk_pct_override=0.01,
|
|
synthetic_only=True,
|
|
),
|
|
],
|
|
)
|
|
config = BacktestConfig(
|
|
strategy_name="return_max_long_v1",
|
|
dataset_snapshot_id="test_snapshot",
|
|
universe=UniverseConfig(min_price=15.0, min_avg_dollar_volume=100_000.0),
|
|
signal=SignalConfig(score_threshold=0.60, max_candidates_per_day=6, scoring_model="patient_drift"),
|
|
risk=RiskConfig(
|
|
per_trade_risk_pct=0.01,
|
|
per_trade_risk_pct_a_tier=0.01,
|
|
max_daily_new_risk_pct=0.05,
|
|
max_positions=6,
|
|
max_positions_per_sector=3,
|
|
),
|
|
execution=ExecutionConfig(
|
|
entry_fill_model="next_open",
|
|
exit_fill_model="daily_bar_approximation",
|
|
slippage_bps_base=0.0,
|
|
commission_per_share=0.0,
|
|
same_bar_priority="stop_first_conservative",
|
|
target_model="fixed_r",
|
|
target_1_r=99.0,
|
|
target_1_fraction=0.0,
|
|
trailing_model=None,
|
|
trailing_warmup_days=10,
|
|
max_holding_days=10,
|
|
),
|
|
reporting=ReportingConfig(
|
|
write_trade_blotter=True,
|
|
write_equity_curve=True,
|
|
write_metrics_summary=True,
|
|
generate_plots=False,
|
|
),
|
|
event_type_profiles={"earnings_release": EventTypeProfile(enabled=True, max_holding_days_override=10)},
|
|
strategy_engines=manifest.strategy_engines,
|
|
strategy_engine_selection_mode="interleave",
|
|
)
|
|
|
|
runner = BacktestRunner(manifest=manifest, config=config, store=store, initial_equity=100_000.0)
|
|
result = runner.run(output_root=tmp_path)
|
|
run_dir = tmp_path / result.run_id
|
|
trade_blotter = pq.read_table(run_dir / "artifacts" / "trade_blotter.parquet").to_pylist()
|
|
|
|
engine_ids = [row["engine_id"] for row in trade_blotter]
|
|
assert "reaction_close_long_core" not in engine_ids
|
|
assert "delayed_primary_long_drift" in engine_ids
|
|
|
|
def test_delayed_entry_preserves_parse_confidence_override(self, tmp_path):
|
|
from apps.backtester.run import BacktestRunner
|
|
import pyarrow.parquet as pq
|
|
from libs.backtest.domain import (
|
|
BacktestConfig,
|
|
EventTypeProfile,
|
|
ExperimentManifest,
|
|
ExecutionConfig,
|
|
ReportingConfig,
|
|
RiskConfig,
|
|
SignalConfig,
|
|
StrategyEngineConfig,
|
|
UniverseConfig,
|
|
)
|
|
from libs.backtest.snapshot_store import SnapshotStore
|
|
|
|
dates = [dt.date(2026, 1, d) for d in [5, 6, 7, 8, 9]]
|
|
store = SnapshotStore(
|
|
candidates_by_exec_date={
|
|
dates[0]: [
|
|
{
|
|
"event_id": "EVT::PARSE::001",
|
|
"symbol": "AAPL",
|
|
"execution_date": dates[0],
|
|
"entry_date": dates[0].isoformat(),
|
|
"event_date": dates[0].isoformat(),
|
|
"event_close": 100.0,
|
|
"entry_price": 100.0,
|
|
"score": 0.82,
|
|
"sector": "Technology",
|
|
"event_type": "earnings_release",
|
|
"event_timestamp": "2026-01-05T21:00:00+00:00",
|
|
"filing_time_bucket": "post_market",
|
|
"reaction_date": dates[0].isoformat(),
|
|
"reaction_day_return": 0.06,
|
|
"close_location": 0.70,
|
|
"volume_ratio_20d": 2.0,
|
|
"gap_size": 0.01,
|
|
"parse_confidence_overall": 0.30,
|
|
"avg_dollar_volume": 9_000_000.0,
|
|
"atr_14": 3.0,
|
|
}
|
|
],
|
|
dates[1]: [],
|
|
dates[2]: [],
|
|
dates[3]: [],
|
|
dates[4]: [],
|
|
},
|
|
bars_by_symbol_date={
|
|
"AAPL": {
|
|
dates[0]: {"date": dates[0], "open": 98.0, "high": 101.0, "low": 97.0, "close": 100.0, "volume": 1_000_000},
|
|
dates[1]: {"date": dates[1], "open": 101.0, "high": 102.0, "low": 100.0, "close": 101.5, "volume": 900_000},
|
|
dates[2]: {"date": dates[2], "open": 101.5, "high": 104.0, "low": 101.0, "close": 103.2, "volume": 850_000},
|
|
dates[3]: {"date": dates[3], "open": 103.0, "high": 106.0, "low": 102.0, "close": 105.0, "volume": 800_000},
|
|
dates[4]: {"date": dates[4], "open": 105.0, "high": 107.0, "low": 104.0, "close": 106.0, "volume": 780_000},
|
|
}
|
|
},
|
|
)
|
|
|
|
manifest = ExperimentManifest(
|
|
experiment_name="return_max_long_v1_shadow_delayed_parse",
|
|
dataset_snapshot_id="test_snapshot",
|
|
base_config="configs/backtest/return_max_long_v1.json",
|
|
overrides={},
|
|
strategy_engines=[
|
|
StrategyEngineConfig(
|
|
engine_id="reaction_close_long_core",
|
|
event_types=["earnings_release"],
|
|
timing_class="same_day",
|
|
direction="long_only",
|
|
entry_timing_policy="reaction_close",
|
|
max_holding_days=10,
|
|
engine_risk_budget_pct=1.0,
|
|
reaction_day_return_min=0.04,
|
|
reaction_day_return_max=0.20,
|
|
close_location_min=0.60,
|
|
volume_ratio_min=1.5,
|
|
score_threshold_override=0.60,
|
|
shadow_only=True,
|
|
),
|
|
StrategyEngineConfig(
|
|
engine_id="delayed_primary_long_parse",
|
|
event_types=["earnings_release"],
|
|
timing_class="any",
|
|
direction="long_only",
|
|
entry_timing_policy="next_open",
|
|
max_holding_days=10,
|
|
engine_risk_budget_pct=1.0,
|
|
delayed_entry_lookback_days=2,
|
|
delayed_entry_source_engine_ids=["reaction_close_long_core"],
|
|
delayed_entry_min_drift_pct=0.02,
|
|
delayed_entry_close_location_min=0.60,
|
|
score_threshold_override=0.0,
|
|
per_trade_risk_pct_override=0.01,
|
|
veto_parse_confidence_min_override=0.20,
|
|
synthetic_only=True,
|
|
),
|
|
],
|
|
)
|
|
config = BacktestConfig(
|
|
strategy_name="return_max_long_v1",
|
|
dataset_snapshot_id="test_snapshot",
|
|
universe=UniverseConfig(min_price=15.0, min_avg_dollar_volume=100_000.0),
|
|
signal=SignalConfig(score_threshold=0.60, max_candidates_per_day=6, scoring_model="patient_drift"),
|
|
risk=RiskConfig(
|
|
per_trade_risk_pct=0.01,
|
|
per_trade_risk_pct_a_tier=0.01,
|
|
max_daily_new_risk_pct=0.05,
|
|
max_positions=6,
|
|
max_positions_per_sector=3,
|
|
veto_parse_confidence_min=0.40,
|
|
),
|
|
execution=ExecutionConfig(
|
|
entry_fill_model="next_open",
|
|
exit_fill_model="daily_bar_approximation",
|
|
slippage_bps_base=0.0,
|
|
commission_per_share=0.0,
|
|
same_bar_priority="stop_first_conservative",
|
|
target_model="fixed_r",
|
|
target_1_r=99.0,
|
|
target_1_fraction=0.0,
|
|
trailing_model=None,
|
|
trailing_warmup_days=10,
|
|
max_holding_days=10,
|
|
),
|
|
reporting=ReportingConfig(
|
|
write_trade_blotter=True,
|
|
write_equity_curve=True,
|
|
write_metrics_summary=True,
|
|
generate_plots=False,
|
|
),
|
|
event_type_profiles={"earnings_release": EventTypeProfile(enabled=True, max_holding_days_override=10)},
|
|
strategy_engines=manifest.strategy_engines,
|
|
strategy_engine_selection_mode="interleave",
|
|
)
|
|
|
|
runner = BacktestRunner(manifest=manifest, config=config, store=store, initial_equity=100_000.0)
|
|
result = runner.run(output_root=tmp_path)
|
|
run_dir = tmp_path / result.run_id
|
|
trade_blotter = pq.read_table(run_dir / "artifacts" / "trade_blotter.parquet").to_pylist()
|
|
|
|
assert len(trade_blotter) == 1
|
|
assert trade_blotter[0]["engine_id"] == "delayed_primary_long_parse"
|
|
|
|
def test_delayed_entry_preserves_no_progress_override(self, tmp_path):
|
|
from apps.backtester.run import BacktestRunner
|
|
import pyarrow.parquet as pq
|
|
from libs.backtest.domain import (
|
|
BacktestConfig,
|
|
EventTypeProfile,
|
|
ExperimentManifest,
|
|
ExecutionConfig,
|
|
ReportingConfig,
|
|
RiskConfig,
|
|
SignalConfig,
|
|
StrategyEngineConfig,
|
|
UniverseConfig,
|
|
)
|
|
from libs.backtest.snapshot_store import SnapshotStore
|
|
|
|
dates = [dt.date(2026, 2, d) for d in [2, 3, 4, 5, 6, 9, 10]]
|
|
store = SnapshotStore(
|
|
candidates_by_exec_date={
|
|
dates[0]: [
|
|
{
|
|
"event_id": "EVT::NP::001",
|
|
"symbol": "AAPL",
|
|
"execution_date": dates[0],
|
|
"entry_date": dates[0].isoformat(),
|
|
"event_date": dates[0].isoformat(),
|
|
"event_close": 100.0,
|
|
"entry_price": 100.0,
|
|
"score": 0.82,
|
|
"sector": "Technology",
|
|
"event_type": "earnings_release",
|
|
"event_timestamp": "2026-02-02T21:00:00+00:00",
|
|
"filing_time_bucket": "post_market",
|
|
"reaction_date": dates[0].isoformat(),
|
|
"reaction_day_return": 0.06,
|
|
"close_location": 0.70,
|
|
"volume_ratio_20d": 2.0,
|
|
"gap_size": 0.01,
|
|
"parse_confidence_overall": 0.80,
|
|
"avg_dollar_volume": 9_000_000.0,
|
|
"atr_14": 3.0,
|
|
}
|
|
],
|
|
dates[1]: [],
|
|
dates[2]: [],
|
|
dates[3]: [],
|
|
dates[4]: [],
|
|
dates[5]: [],
|
|
dates[6]: [],
|
|
},
|
|
bars_by_symbol_date={
|
|
"AAPL": {
|
|
dates[0]: {"date": dates[0], "open": 98.0, "high": 101.0, "low": 97.0, "close": 100.0, "volume": 1_000_000},
|
|
dates[1]: {"date": dates[1], "open": 100.5, "high": 101.0, "low": 99.8, "close": 100.8, "volume": 900_000},
|
|
dates[2]: {"date": dates[2], "open": 100.7, "high": 101.2, "low": 100.1, "close": 100.9, "volume": 850_000},
|
|
dates[3]: {"date": dates[3], "open": 101.0, "high": 101.8, "low": 100.7, "close": 101.3, "volume": 820_000},
|
|
dates[4]: {"date": dates[4], "open": 101.2, "high": 102.0, "low": 100.9, "close": 101.6, "volume": 810_000},
|
|
dates[5]: {"date": dates[5], "open": 101.7, "high": 102.4, "low": 101.3, "close": 102.0, "volume": 800_000},
|
|
dates[6]: {"date": dates[6], "open": 102.0, "high": 102.5, "low": 101.6, "close": 102.2, "volume": 790_000},
|
|
}
|
|
},
|
|
)
|
|
|
|
manifest = ExperimentManifest(
|
|
experiment_name="return_max_long_v1_shadow_delayed_np",
|
|
dataset_snapshot_id="test_snapshot",
|
|
base_config="configs/backtest/return_max_long_v1.json",
|
|
overrides={},
|
|
strategy_engines=[
|
|
StrategyEngineConfig(
|
|
engine_id="reaction_close_long_core",
|
|
event_types=["earnings_release"],
|
|
timing_class="same_day",
|
|
direction="long_only",
|
|
entry_timing_policy="reaction_close",
|
|
max_holding_days=10,
|
|
engine_risk_budget_pct=1.0,
|
|
reaction_day_return_min=0.04,
|
|
reaction_day_return_max=0.20,
|
|
close_location_min=0.60,
|
|
volume_ratio_min=1.5,
|
|
score_threshold_override=0.60,
|
|
shadow_only=True,
|
|
),
|
|
StrategyEngineConfig(
|
|
engine_id="delayed_primary_long_np",
|
|
event_types=["earnings_release"],
|
|
timing_class="any",
|
|
direction="long_only",
|
|
entry_timing_policy="next_open",
|
|
max_holding_days=10,
|
|
engine_risk_budget_pct=1.0,
|
|
delayed_entry_lookback_days=2,
|
|
delayed_entry_source_engine_ids=["reaction_close_long_core"],
|
|
delayed_entry_min_drift_pct=0.005,
|
|
delayed_entry_close_location_min=0.55,
|
|
score_threshold_override=0.0,
|
|
per_trade_risk_pct_override=0.01,
|
|
early_failure_no_progress_days_override=5,
|
|
early_failure_no_progress_r_override=0.1,
|
|
early_failure_no_progress_fraction_override=1.0,
|
|
synthetic_only=True,
|
|
),
|
|
],
|
|
)
|
|
config = BacktestConfig(
|
|
strategy_name="return_max_long_v1",
|
|
dataset_snapshot_id="test_snapshot",
|
|
universe=UniverseConfig(min_price=15.0, min_avg_dollar_volume=100_000.0),
|
|
signal=SignalConfig(score_threshold=0.60, max_candidates_per_day=6, scoring_model="patient_drift"),
|
|
risk=RiskConfig(
|
|
per_trade_risk_pct=0.01,
|
|
per_trade_risk_pct_a_tier=0.01,
|
|
max_daily_new_risk_pct=0.05,
|
|
max_positions=6,
|
|
max_positions_per_sector=3,
|
|
),
|
|
execution=ExecutionConfig(
|
|
entry_fill_model="next_open",
|
|
exit_fill_model="daily_bar_approximation",
|
|
slippage_bps_base=0.0,
|
|
commission_per_share=0.0,
|
|
same_bar_priority="stop_first_conservative",
|
|
target_model="fixed_r",
|
|
target_1_r=99.0,
|
|
target_1_fraction=0.0,
|
|
trailing_model=None,
|
|
trailing_warmup_days=10,
|
|
max_holding_days=10,
|
|
early_failure_no_progress_days=2,
|
|
early_failure_no_progress_r=0.1,
|
|
early_failure_no_progress_fraction=1.0,
|
|
),
|
|
reporting=ReportingConfig(
|
|
write_trade_blotter=True,
|
|
write_equity_curve=True,
|
|
write_metrics_summary=True,
|
|
generate_plots=False,
|
|
),
|
|
event_type_profiles={"earnings_release": EventTypeProfile(enabled=True, max_holding_days_override=10)},
|
|
strategy_engines=manifest.strategy_engines,
|
|
strategy_engine_selection_mode="interleave",
|
|
)
|
|
|
|
runner = BacktestRunner(manifest=manifest, config=config, store=store, initial_equity=100_000.0)
|
|
result = runner.run(output_root=tmp_path)
|
|
run_dir = tmp_path / result.run_id
|
|
trade_blotter = pq.read_table(run_dir / "artifacts" / "trade_blotter.parquet").to_pylist()
|
|
|
|
assert len(trade_blotter) == 1
|
|
assert trade_blotter[0]["engine_id"] == "delayed_primary_long_np"
|
|
assert trade_blotter[0]["holding_days"] > 2
|
|
|
|
def test_selected_candidates_are_annotated_with_daily_breadth_features(self, tmp_path):
|
|
from apps.backtester.run import BacktestRunner
|
|
from libs.backtest.domain import (
|
|
BacktestConfig,
|
|
Candidate,
|
|
ExecutionConfig,
|
|
ExperimentManifest,
|
|
ReportingConfig,
|
|
RiskConfig,
|
|
SignalConfig,
|
|
StrategyEngineConfig,
|
|
UniverseConfig,
|
|
)
|
|
|
|
store = _build_multi_engine_store()
|
|
manifest = ExperimentManifest(
|
|
experiment_name="breadth_annotation_smoke",
|
|
dataset_snapshot_id="test_snapshot",
|
|
base_config="configs/backtest/defaults.json",
|
|
overrides={},
|
|
strategy_engines=[
|
|
StrategyEngineConfig(
|
|
engine_id="same_day_short",
|
|
event_types=["earnings_release"],
|
|
timing_class="same_day",
|
|
direction="short_only",
|
|
entry_timing_policy="next_open",
|
|
reaction_day_return_max=-0.05,
|
|
score_threshold_override=0.50,
|
|
),
|
|
StrategyEngineConfig(
|
|
engine_id="after_close_long",
|
|
event_types=["earnings_release"],
|
|
timing_class="after_close",
|
|
direction="long_only",
|
|
entry_timing_policy="next_open",
|
|
reaction_day_return_min=0.05,
|
|
score_threshold_override=0.50,
|
|
),
|
|
],
|
|
)
|
|
config = BacktestConfig(
|
|
strategy_name="breadth_annotation_smoke",
|
|
dataset_snapshot_id="test_snapshot",
|
|
universe=UniverseConfig(min_price=5.0, min_avg_dollar_volume=100_000.0),
|
|
signal=SignalConfig(score_threshold=0.50, max_candidates_per_day=5),
|
|
risk=RiskConfig(
|
|
per_trade_risk_pct=0.01,
|
|
max_daily_new_risk_pct=0.05,
|
|
max_positions=10,
|
|
max_positions_per_sector=5,
|
|
),
|
|
execution=ExecutionConfig(
|
|
entry_fill_model="next_open",
|
|
exit_fill_model="daily_bar_approximation",
|
|
slippage_bps_base=0.0,
|
|
commission_per_share=0.0,
|
|
same_bar_priority="stop_first_conservative",
|
|
max_holding_days=5,
|
|
),
|
|
reporting=ReportingConfig(
|
|
write_trade_blotter=True,
|
|
write_equity_curve=True,
|
|
write_metrics_summary=True,
|
|
generate_plots=False,
|
|
),
|
|
strategy_engines=manifest.strategy_engines,
|
|
)
|
|
|
|
runner = BacktestRunner(manifest=manifest, config=config, store=store, initial_equity=100_000.0)
|
|
selected = runner._select_candidates_for_date(dt.date(2026, 1, 7))
|
|
|
|
assert len(selected) == 2
|
|
for candidate in selected:
|
|
assert candidate.features["daily_candidate_count_selected"] == 2
|
|
assert candidate.features["daily_unique_sector_count_selected"] == 2
|
|
assert candidate.features["daily_sector_candidate_count_selected"] == 1
|
|
assert candidate.features["daily_engine_candidate_count_selected"] == 1
|
|
|
|
def test_macro_bullish_engine_schedules_synthetic_qqq_candidate(self, tmp_path):
|
|
from apps.backtester.run import BacktestRunner
|
|
from libs.backtest.domain import (
|
|
BacktestConfig,
|
|
Candidate,
|
|
ExecutionConfig,
|
|
ExperimentManifest,
|
|
ReportingConfig,
|
|
RiskConfig,
|
|
SignalConfig,
|
|
StrategyEngineConfig,
|
|
UniverseConfig,
|
|
)
|
|
from libs.backtest.snapshot_store import SnapshotStore
|
|
|
|
signal_date = dt.date(2026, 1, 6)
|
|
next_date = dt.date(2026, 1, 7)
|
|
store = SnapshotStore(
|
|
candidates_by_exec_date={next_date: []},
|
|
bars_by_symbol_date={
|
|
"QQQ": {
|
|
dt.date(2026, 1, 5): {"date": dt.date(2026, 1, 5), "open": 100.0, "high": 101.0, "low": 99.0, "close": 100.0, "volume": 100},
|
|
signal_date: {"date": signal_date, "open": 102.0, "high": 107.0, "low": 101.5, "close": 106.0, "volume": 400},
|
|
next_date: {"date": next_date, "open": 106.5, "high": 108.0, "low": 105.0, "close": 107.5, "volume": 350},
|
|
}
|
|
},
|
|
macro_by_date={signal_date: {"VIXCLS": 22.0}},
|
|
)
|
|
manifest = ExperimentManifest(
|
|
experiment_name="macro_bullish_qqq_smoke",
|
|
dataset_snapshot_id="test_snapshot",
|
|
base_config="configs/backtest/defaults.json",
|
|
overrides={},
|
|
strategy_engines=[
|
|
StrategyEngineConfig(
|
|
engine_id="macro_bullish_qqq",
|
|
synthetic_only=True,
|
|
entry_timing_policy="next_open",
|
|
max_holding_days=8,
|
|
engine_risk_budget_pct=0.25,
|
|
per_trade_risk_pct_override=0.01,
|
|
stop_atr_multiplier_override=2.0,
|
|
trailing_warmup_days_override=4,
|
|
macro_long_symbol="QQQ",
|
|
macro_long_reaction_day_return_min=0.015,
|
|
macro_long_volume_ratio_min=2.0,
|
|
macro_long_gap_size_min=0.01,
|
|
macro_long_close_location_min=0.75,
|
|
macro_long_min_daily_candidate_count=2,
|
|
macro_long_min_unique_sector_count=2,
|
|
macro_vix_max=30.0,
|
|
)
|
|
],
|
|
)
|
|
config = BacktestConfig(
|
|
strategy_name="macro_bullish_qqq_smoke",
|
|
dataset_snapshot_id="test_snapshot",
|
|
universe=UniverseConfig(min_price=5.0, min_avg_dollar_volume=100_000.0),
|
|
signal=SignalConfig(score_threshold=0.50, max_candidates_per_day=5),
|
|
risk=RiskConfig(
|
|
per_trade_risk_pct=0.01,
|
|
max_daily_new_risk_pct=0.05,
|
|
max_positions=10,
|
|
max_positions_per_sector=5,
|
|
),
|
|
execution=ExecutionConfig(
|
|
entry_fill_model="next_open",
|
|
exit_fill_model="daily_bar_approximation",
|
|
slippage_bps_base=0.0,
|
|
commission_per_share=0.0,
|
|
same_bar_priority="stop_first_conservative",
|
|
max_holding_days=10,
|
|
),
|
|
reporting=ReportingConfig(
|
|
write_trade_blotter=True,
|
|
write_equity_curve=True,
|
|
write_metrics_summary=True,
|
|
generate_plots=False,
|
|
),
|
|
strategy_engines=manifest.strategy_engines,
|
|
)
|
|
|
|
runner = BacktestRunner(manifest=manifest, config=config, store=store, initial_equity=100_000.0)
|
|
runner._simulation_dates = [signal_date, next_date]
|
|
runner._next_trading_day = {signal_date: next_date}
|
|
runner._recent_scored_candidates[signal_date] = [
|
|
Candidate(
|
|
event_id="BREADTH::1",
|
|
symbol="AAPL",
|
|
score=0.8,
|
|
sector="Technology",
|
|
event_type="earnings_release",
|
|
event_timestamp=dt.datetime(2026, 1, 6, 21, 0, tzinfo=_UTC),
|
|
filing_time_bucket="post_market",
|
|
reaction_date=signal_date,
|
|
execution_date=next_date,
|
|
entry_price_est=100.0,
|
|
avg_dollar_volume=1_000_000.0,
|
|
atr_14=2.0,
|
|
score_bucket="high",
|
|
engine_id="core",
|
|
),
|
|
Candidate(
|
|
event_id="BREADTH::2",
|
|
symbol="LLY",
|
|
score=0.78,
|
|
sector="Health Care",
|
|
event_type="earnings_release",
|
|
event_timestamp=dt.datetime(2026, 1, 6, 21, 0, tzinfo=_UTC),
|
|
filing_time_bucket="post_market",
|
|
reaction_date=signal_date,
|
|
execution_date=next_date,
|
|
entry_price_est=100.0,
|
|
avg_dollar_volume=1_000_000.0,
|
|
atr_14=2.0,
|
|
score_bucket="high",
|
|
engine_id="core",
|
|
),
|
|
]
|
|
|
|
runner._schedule_macro_long_candidates(signal_date)
|
|
|
|
scheduled = runner._scheduled_delayed_entries[next_date]
|
|
assert len(scheduled) == 1
|
|
candidate = scheduled[0]
|
|
assert candidate.symbol == "QQQ"
|
|
assert candidate.event_type == "macro_bullish_event"
|
|
assert candidate.features["macro_long_daily_candidate_count"] == 2
|
|
assert candidate.features["macro_long_daily_unique_sector_count"] == 2
|
|
assert candidate.features["macro_long_reaction_day_return"] == pytest.approx(0.06)
|
|
|
|
def test_leader_follower_engine_schedules_preentry_peer_candidate(self, tmp_path):
|
|
from apps.backtester.run import BacktestRunner
|
|
from libs.backtest.domain import (
|
|
BacktestConfig,
|
|
ExecutionConfig,
|
|
ExperimentManifest,
|
|
ReportingConfig,
|
|
RiskConfig,
|
|
SignalConfig,
|
|
StrategyEngineConfig,
|
|
UniverseConfig,
|
|
)
|
|
from libs.backtest.snapshot_store import SnapshotStore
|
|
|
|
signal_date = dt.date(2026, 1, 6)
|
|
entry_date = dt.date(2026, 1, 7)
|
|
pre_event_date = dt.date(2026, 1, 8)
|
|
follower_event_date = dt.date(2026, 1, 9)
|
|
follower_exec_date = dt.date(2026, 1, 12)
|
|
|
|
store = SnapshotStore(
|
|
candidates_by_exec_date={
|
|
entry_date: [
|
|
{
|
|
"event_id": "LEADER::NVDA",
|
|
"symbol": "NVDA",
|
|
"execution_date": entry_date,
|
|
"entry_date": str(entry_date),
|
|
"event_date": str(signal_date),
|
|
"event_close": 120.0,
|
|
"entry_price": 121.0,
|
|
"score": 0.86,
|
|
"sector": "Technology",
|
|
"event_type": "earnings_release",
|
|
"event_direction": "bullish",
|
|
"guidance_status": "raised",
|
|
"event_timestamp": "2026-01-06T21:00:00+00:00",
|
|
"filing_time_bucket": "post_market",
|
|
"reaction_date": str(signal_date),
|
|
"reaction_day_return": 0.14,
|
|
"close_location": 0.88,
|
|
"volume_ratio": 3.4,
|
|
"gap_size": 0.07,
|
|
"avg_dollar_volume": 50_000_000.0,
|
|
"market_cap_proxy": 100_000_000_000.0,
|
|
"document_quality_score": 0.80,
|
|
"parse_confidence_overall": 0.82,
|
|
"atr_14": 4.0,
|
|
}
|
|
],
|
|
follower_exec_date: [
|
|
{
|
|
"event_id": "FOLLOWER::AMD",
|
|
"symbol": "AMD",
|
|
"execution_date": follower_exec_date,
|
|
"entry_date": str(follower_exec_date),
|
|
"event_date": str(follower_event_date),
|
|
"event_close": 51.5,
|
|
"entry_price": 52.0,
|
|
"score": 0.40,
|
|
"sector": "Technology",
|
|
"event_type": "earnings_release",
|
|
"event_direction": "unknown",
|
|
"guidance_status": "not_provided",
|
|
"event_timestamp": "2026-01-09T21:00:00+00:00",
|
|
"filing_time_bucket": "post_market",
|
|
"reaction_date": str(follower_event_date),
|
|
"reaction_day_return": 0.03,
|
|
"close_location": 0.55,
|
|
"volume_ratio": 1.4,
|
|
"gap_size": 0.01,
|
|
"avg_dollar_volume": 30_000_000.0,
|
|
"market_cap_proxy": 60_000_000_000.0,
|
|
"document_quality_score": 0.60,
|
|
"parse_confidence_overall": 0.60,
|
|
"atr_14": 2.5,
|
|
}
|
|
],
|
|
},
|
|
bars_by_symbol_date={
|
|
"NVDA": {
|
|
signal_date: {"date": signal_date, "open": 108.0, "high": 121.0, "low": 107.0, "close": 120.0, "volume": 4_000_000},
|
|
entry_date: {"date": entry_date, "open": 121.0, "high": 123.0, "low": 118.0, "close": 122.0, "volume": 3_000_000},
|
|
},
|
|
"AMD": {
|
|
dt.date(2026, 1, 5): {"date": dt.date(2026, 1, 5), "open": 49.5, "high": 50.5, "low": 49.0, "close": 50.0, "volume": 1_000_000},
|
|
signal_date: {"date": signal_date, "open": 50.2, "high": 51.6, "low": 49.8, "close": 51.0, "volume": 1_100_000},
|
|
entry_date: {"date": entry_date, "open": 51.1, "high": 52.0, "low": 50.8, "close": 51.7, "volume": 1_050_000},
|
|
pre_event_date: {"date": pre_event_date, "open": 51.9, "high": 52.6, "low": 51.4, "close": 52.3, "volume": 1_030_000},
|
|
follower_event_date: {"date": follower_event_date, "open": 52.4, "high": 53.4, "low": 51.8, "close": 53.0, "volume": 1_200_000},
|
|
follower_exec_date: {"date": follower_exec_date, "open": 54.5, "high": 55.0, "low": 53.5, "close": 54.0, "volume": 1_250_000},
|
|
},
|
|
},
|
|
)
|
|
manifest = ExperimentManifest(
|
|
experiment_name="leader_follower_preentry_smoke",
|
|
dataset_snapshot_id="test_snapshot",
|
|
base_config="configs/backtest/defaults.json",
|
|
overrides={},
|
|
strategy_engines=[
|
|
StrategyEngineConfig(
|
|
engine_id="leader_follower_preentry",
|
|
synthetic_only=True,
|
|
entry_timing_policy="next_open",
|
|
event_types=["earnings_release"],
|
|
event_directions=["bullish"],
|
|
guidance_statuses=["raised"],
|
|
filing_time_buckets=["post_market"],
|
|
allowed_sectors=["Technology"],
|
|
direction="long_only",
|
|
max_holding_days=5,
|
|
engine_risk_budget_pct=0.05,
|
|
per_trade_risk_pct_override=0.01,
|
|
reaction_day_return_min=0.10,
|
|
close_location_min=0.75,
|
|
volume_ratio_min=2.0,
|
|
gap_size_min=0.04,
|
|
min_market_cap_proxy=40_000_000_000.0,
|
|
document_quality_score_min=0.5,
|
|
parse_confidence_overall_min=0.5,
|
|
leader_follower_lookahead_days=3,
|
|
leader_follower_min_days_to_event=2,
|
|
leader_follower_hold_buffer_days=1,
|
|
proxy_reaction_day_return_max=0.04,
|
|
proxy_gap_size_max=0.03,
|
|
proxy_close_location_max=0.80,
|
|
proxy_avg_dollar_volume_min=10_000_000.0,
|
|
next_open_gap_cap_pct=0.08,
|
|
stop_atr_multiplier_override=2.5,
|
|
trailing_warmup_days_override=3,
|
|
early_failure_no_progress_days_override=1,
|
|
early_failure_no_progress_r_override=0.0,
|
|
early_failure_no_progress_fraction_override=1.0,
|
|
)
|
|
],
|
|
)
|
|
config = BacktestConfig(
|
|
strategy_name="leader_follower_preentry_smoke",
|
|
dataset_snapshot_id="test_snapshot",
|
|
universe=UniverseConfig(min_price=5.0, min_avg_dollar_volume=100_000.0),
|
|
signal=SignalConfig(score_threshold=0.50, max_candidates_per_day=5),
|
|
risk=RiskConfig(
|
|
per_trade_risk_pct=0.01,
|
|
max_daily_new_risk_pct=0.05,
|
|
max_positions=10,
|
|
max_positions_per_sector=5,
|
|
),
|
|
execution=ExecutionConfig(
|
|
entry_fill_model="next_open",
|
|
exit_fill_model="daily_bar_approximation",
|
|
slippage_bps_base=0.0,
|
|
commission_per_share=0.0,
|
|
same_bar_priority="stop_first_conservative",
|
|
max_holding_days=10,
|
|
),
|
|
reporting=ReportingConfig(
|
|
write_trade_blotter=True,
|
|
write_equity_curve=True,
|
|
write_metrics_summary=True,
|
|
generate_plots=False,
|
|
),
|
|
strategy_engines=manifest.strategy_engines,
|
|
)
|
|
|
|
runner = BacktestRunner(manifest=manifest, config=config, store=store, initial_equity=100_000.0)
|
|
runner._simulation_dates = [signal_date, entry_date, pre_event_date, follower_event_date, follower_exec_date]
|
|
runner._next_trading_day = {
|
|
signal_date: entry_date,
|
|
entry_date: pre_event_date,
|
|
pre_event_date: follower_event_date,
|
|
follower_event_date: follower_exec_date,
|
|
}
|
|
|
|
runner._schedule_leader_follower_candidates(signal_date)
|
|
|
|
scheduled = runner._scheduled_delayed_entries[entry_date]
|
|
assert len(scheduled) == 1
|
|
candidate = scheduled[0]
|
|
assert candidate.symbol == "AMD"
|
|
assert candidate.source_symbol == "NVDA"
|
|
assert candidate.event_type == "leader_follower_preearnings"
|
|
assert candidate.engine_max_holding_days == 1
|
|
assert candidate.features["leader_follower_days_to_event"] == 2
|
|
assert candidate.features["leader_symbol"] == "NVDA"
|
|
assert candidate.features["follower_symbol"] == "AMD"
|
|
|
|
def test_leader_follower_prefers_calm_orderly_follower(self, tmp_path):
|
|
from apps.backtester.run import BacktestRunner
|
|
from libs.backtest.domain import (
|
|
BacktestConfig,
|
|
ExecutionConfig,
|
|
ExperimentManifest,
|
|
ReportingConfig,
|
|
RiskConfig,
|
|
SignalConfig,
|
|
StrategyEngineConfig,
|
|
UniverseConfig,
|
|
)
|
|
from libs.backtest.snapshot_store import SnapshotStore
|
|
|
|
signal_date = dt.date(2026, 1, 6)
|
|
entry_date = dt.date(2026, 1, 7)
|
|
mid_date = dt.date(2026, 1, 8)
|
|
crm_reaction_date = dt.date(2026, 1, 9)
|
|
nvda_reaction_date = crm_reaction_date
|
|
crm_exec_date = dt.date(2026, 1, 12)
|
|
nvda_exec_date = dt.date(2026, 1, 13)
|
|
store = SnapshotStore(
|
|
candidates_by_exec_date={
|
|
entry_date: [
|
|
{
|
|
"event_id": "LEADER::SNOW",
|
|
"symbol": "SNOW",
|
|
"execution_date": entry_date,
|
|
"entry_date": str(entry_date),
|
|
"event_date": str(signal_date),
|
|
"event_close": 116.0,
|
|
"entry_price": 116.5,
|
|
"score": 0.86,
|
|
"sector": "Technology",
|
|
"event_type": "earnings_release",
|
|
"event_direction": "bullish",
|
|
"guidance_status": "raised",
|
|
"event_timestamp": "2026-01-06T21:00:00+00:00",
|
|
"filing_time_bucket": "post_market",
|
|
"reaction_date": str(signal_date),
|
|
"reaction_day_return": 0.16,
|
|
"close_location": 0.87,
|
|
"volume_ratio": 3.2,
|
|
"gap_size": 0.06,
|
|
"avg_dollar_volume": 90_000_000.0,
|
|
"market_cap_proxy": 80_000_000_000.0,
|
|
"document_quality_score": 0.8,
|
|
"parse_confidence_overall": 0.8,
|
|
"atr_14": 4.0,
|
|
}
|
|
],
|
|
crm_exec_date: [
|
|
{
|
|
"event_id": "FOLLOWER::CRM",
|
|
"symbol": "CRM",
|
|
"execution_date": crm_exec_date,
|
|
"entry_date": str(crm_exec_date),
|
|
"event_date": str(crm_reaction_date),
|
|
"event_close": 96.4,
|
|
"entry_price": 96.6,
|
|
"score": 0.4,
|
|
"sector": "Technology",
|
|
"event_type": "earnings_release",
|
|
"event_direction": "unknown",
|
|
"guidance_status": "not_provided",
|
|
"event_timestamp": "2026-01-09T21:00:00+00:00",
|
|
"filing_time_bucket": "post_market",
|
|
"reaction_date": str(crm_reaction_date),
|
|
"reaction_day_return": 0.03,
|
|
"close_location": 0.55,
|
|
"volume_ratio": 1.3,
|
|
"gap_size": 0.01,
|
|
"avg_dollar_volume": 50_000_000.0,
|
|
"market_cap_proxy": 150_000_000_000.0,
|
|
"document_quality_score": 0.7,
|
|
"parse_confidence_overall": 0.7,
|
|
"atr_14": 3.0,
|
|
}
|
|
],
|
|
nvda_exec_date: [
|
|
{
|
|
"event_id": "FOLLOWER::NVDA",
|
|
"symbol": "NVDA",
|
|
"execution_date": nvda_exec_date,
|
|
"entry_date": str(nvda_exec_date),
|
|
"event_date": str(nvda_reaction_date),
|
|
"event_close": 98.8,
|
|
"entry_price": 99.0,
|
|
"score": 0.4,
|
|
"sector": "Technology",
|
|
"event_type": "earnings_release",
|
|
"event_direction": "unknown",
|
|
"guidance_status": "not_provided",
|
|
"event_timestamp": "2026-01-09T21:00:00+00:00",
|
|
"filing_time_bucket": "post_market",
|
|
"reaction_date": str(nvda_reaction_date),
|
|
"reaction_day_return": 0.03,
|
|
"close_location": 0.55,
|
|
"volume_ratio": 1.3,
|
|
"gap_size": 0.01,
|
|
"avg_dollar_volume": 50_000_000.0,
|
|
"market_cap_proxy": 300_000_000_000.0,
|
|
"document_quality_score": 0.7,
|
|
"parse_confidence_overall": 0.7,
|
|
"atr_14": 3.0,
|
|
}
|
|
],
|
|
},
|
|
bars_by_symbol_date={
|
|
"SNOW": {
|
|
dt.date(2026, 1, 5): {"date": dt.date(2026, 1, 5), "open": 100.0, "high": 101.0, "low": 99.0, "close": 100.0, "volume": 1_000_000},
|
|
signal_date: {"date": signal_date, "open": 106.0, "high": 117.0, "low": 105.5, "close": 116.0, "volume": 3_200_000},
|
|
entry_date: {"date": entry_date, "open": 116.5, "high": 118.0, "low": 114.0, "close": 117.0, "volume": 2_500_000},
|
|
},
|
|
"CRM": {
|
|
dt.date(2026, 1, 5): {"date": dt.date(2026, 1, 5), "open": 100.0, "high": 101.0, "low": 99.0, "close": 100.0, "volume": 1_000_000},
|
|
signal_date: {"date": signal_date, "open": 98.4, "high": 101.0, "low": 95.5, "close": 96.4, "volume": 1_800_000},
|
|
entry_date: {"date": entry_date, "open": 96.0, "high": 97.0, "low": 95.0, "close": 96.5, "volume": 1_200_000},
|
|
},
|
|
"NVDA": {
|
|
dt.date(2026, 1, 5): {"date": dt.date(2026, 1, 5), "open": 100.0, "high": 101.0, "low": 99.0, "close": 100.0, "volume": 1_000_000},
|
|
signal_date: {"date": signal_date, "open": 97.9, "high": 99.65, "low": 97.5, "close": 98.8, "volume": 950_000},
|
|
entry_date: {"date": entry_date, "open": 98.9, "high": 100.2, "low": 98.0, "close": 99.8, "volume": 1_100_000},
|
|
},
|
|
},
|
|
)
|
|
manifest = ExperimentManifest(
|
|
experiment_name="leader_follower_prefers_orderly",
|
|
dataset_snapshot_id="test_snapshot",
|
|
base_config="configs/backtest/defaults.json",
|
|
overrides={},
|
|
strategy_engines=[
|
|
StrategyEngineConfig(
|
|
engine_id="leader_follower_preentry",
|
|
synthetic_only=True,
|
|
entry_timing_policy="next_open",
|
|
event_types=["earnings_release"],
|
|
event_directions=["bullish"],
|
|
guidance_statuses=["raised"],
|
|
filing_time_buckets=["post_market"],
|
|
allowed_sectors=["Technology"],
|
|
direction="long_only",
|
|
max_holding_days=5,
|
|
engine_risk_budget_pct=0.05,
|
|
per_trade_risk_pct_override=0.01,
|
|
reaction_day_return_min=0.10,
|
|
close_location_min=0.75,
|
|
volume_ratio_min=2.0,
|
|
gap_size_min=0.04,
|
|
min_market_cap_proxy=40_000_000_000.0,
|
|
document_quality_score_min=0.5,
|
|
parse_confidence_overall_min=0.5,
|
|
leader_follower_lookahead_days=3,
|
|
leader_follower_min_days_to_event=2,
|
|
leader_follower_hold_buffer_days=1,
|
|
proxy_reaction_day_return_max=0.05,
|
|
proxy_gap_size_max=0.03,
|
|
proxy_close_location_max=0.85,
|
|
proxy_avg_dollar_volume_min=10_000_000.0,
|
|
)
|
|
],
|
|
)
|
|
config = BacktestConfig(
|
|
strategy_name="leader_follower_prefers_orderly",
|
|
dataset_snapshot_id="test_snapshot",
|
|
universe=UniverseConfig(min_price=5.0, min_avg_dollar_volume=100_000.0),
|
|
signal=SignalConfig(score_threshold=0.50, max_candidates_per_day=5),
|
|
risk=RiskConfig(
|
|
per_trade_risk_pct=0.01,
|
|
max_daily_new_risk_pct=0.05,
|
|
max_positions=10,
|
|
max_positions_per_sector=5,
|
|
),
|
|
execution=ExecutionConfig(
|
|
entry_fill_model="next_open",
|
|
exit_fill_model="daily_bar_approximation",
|
|
slippage_bps_base=0.0,
|
|
commission_per_share=0.0,
|
|
same_bar_priority="stop_first_conservative",
|
|
max_holding_days=10,
|
|
),
|
|
reporting=ReportingConfig(
|
|
write_trade_blotter=True,
|
|
write_equity_curve=True,
|
|
write_metrics_summary=True,
|
|
generate_plots=False,
|
|
),
|
|
strategy_engines=manifest.strategy_engines,
|
|
)
|
|
|
|
runner = BacktestRunner(manifest=manifest, config=config, store=store, initial_equity=100_000.0)
|
|
runner._simulation_dates = [signal_date, entry_date, mid_date, crm_reaction_date, nvda_exec_date]
|
|
runner._next_trading_day = {
|
|
signal_date: entry_date,
|
|
entry_date: mid_date,
|
|
mid_date: crm_reaction_date,
|
|
crm_reaction_date: nvda_exec_date,
|
|
}
|
|
|
|
runner._schedule_leader_follower_candidates(signal_date)
|
|
|
|
scheduled = runner._scheduled_delayed_entries[entry_date]
|
|
assert len(scheduled) == 2
|
|
assert scheduled[0].symbol == "NVDA"
|
|
assert scheduled[1].symbol == "CRM"
|
|
assert scheduled[0].score > scheduled[1].score
|
|
|
|
def test_leader_follower_engine_uses_pit_calendar_when_available(self, tmp_path):
|
|
from apps.backtester.run import BacktestRunner
|
|
from libs.backtest.domain import (
|
|
BacktestConfig,
|
|
ExecutionConfig,
|
|
ExperimentManifest,
|
|
ReportingConfig,
|
|
RiskConfig,
|
|
SignalConfig,
|
|
StrategyEngineConfig,
|
|
UniverseConfig,
|
|
)
|
|
from libs.backtest.earnings_calendar import load_pit_earnings_calendar
|
|
from libs.backtest.snapshot_store import SnapshotStore
|
|
|
|
signal_date = dt.date(2026, 1, 6)
|
|
entry_date = dt.date(2026, 1, 7)
|
|
pre_event_date = dt.date(2026, 1, 8)
|
|
follower_reaction_date = dt.date(2026, 1, 9)
|
|
follower_exec_date = dt.date(2026, 1, 12)
|
|
|
|
calendar_path = tmp_path / "earnings_calendar_pit.parquet"
|
|
pq.write_table(
|
|
pa.Table.from_pylist(
|
|
[
|
|
{
|
|
"symbol": "AMD",
|
|
"as_of_date": signal_date.isoformat(),
|
|
"expected_reaction_date": follower_reaction_date.isoformat(),
|
|
"expected_event_date": pre_event_date.isoformat(),
|
|
"filing_time_bucket": "post_market",
|
|
"source": "unit_test",
|
|
}
|
|
]
|
|
),
|
|
calendar_path,
|
|
)
|
|
load_pit_earnings_calendar.cache_clear()
|
|
|
|
store = SnapshotStore(
|
|
candidates_by_exec_date={
|
|
entry_date: [
|
|
{
|
|
"event_id": "LEADER::NVDA",
|
|
"symbol": "NVDA",
|
|
"execution_date": entry_date,
|
|
"entry_date": str(entry_date),
|
|
"event_date": str(signal_date),
|
|
"event_close": 120.0,
|
|
"entry_price": 121.0,
|
|
"score": 0.86,
|
|
"sector": "Technology",
|
|
"event_type": "earnings_release",
|
|
"event_direction": "bullish",
|
|
"guidance_status": "raised",
|
|
"event_timestamp": "2026-01-06T21:00:00+00:00",
|
|
"filing_time_bucket": "post_market",
|
|
"reaction_date": str(signal_date),
|
|
"reaction_day_return": 0.14,
|
|
"close_location": 0.88,
|
|
"volume_ratio": 3.4,
|
|
"gap_size": 0.07,
|
|
"avg_dollar_volume": 50_000_000.0,
|
|
"market_cap_proxy": 100_000_000_000.0,
|
|
"document_quality_score": 0.80,
|
|
"parse_confidence_overall": 0.82,
|
|
"atr_14": 4.0,
|
|
}
|
|
],
|
|
},
|
|
bars_by_symbol_date={
|
|
"NVDA": {
|
|
signal_date: {"date": signal_date, "open": 108.0, "high": 121.0, "low": 107.0, "close": 120.0, "volume": 4_000_000},
|
|
entry_date: {"date": entry_date, "open": 121.0, "high": 123.0, "low": 118.0, "close": 122.0, "volume": 3_000_000},
|
|
},
|
|
"AMD": {
|
|
dt.date(2026, 1, 5): {"date": dt.date(2026, 1, 5), "open": 49.5, "high": 50.5, "low": 49.0, "close": 50.0, "volume": 1_000_000},
|
|
signal_date: {"date": signal_date, "open": 50.2, "high": 51.6, "low": 49.8, "close": 51.0, "volume": 1_100_000},
|
|
entry_date: {"date": entry_date, "open": 51.1, "high": 52.0, "low": 50.8, "close": 51.7, "volume": 1_050_000},
|
|
pre_event_date: {"date": pre_event_date, "open": 51.9, "high": 52.6, "low": 51.4, "close": 52.3, "volume": 1_030_000},
|
|
follower_reaction_date: {"date": follower_reaction_date, "open": 52.4, "high": 53.4, "low": 51.8, "close": 53.0, "volume": 1_200_000},
|
|
follower_exec_date: {"date": follower_exec_date, "open": 54.5, "high": 55.0, "low": 53.5, "close": 54.0, "volume": 1_250_000},
|
|
},
|
|
},
|
|
)
|
|
manifest = ExperimentManifest(
|
|
experiment_name="leader_follower_preentry_pit_smoke",
|
|
dataset_snapshot_id="test_snapshot",
|
|
base_config="configs/backtest/defaults.json",
|
|
overrides={},
|
|
strategy_engines=[
|
|
StrategyEngineConfig(
|
|
engine_id="leader_follower_preentry",
|
|
synthetic_only=True,
|
|
entry_timing_policy="next_open",
|
|
event_types=["earnings_release"],
|
|
event_directions=["bullish"],
|
|
guidance_statuses=["raised"],
|
|
filing_time_buckets=["post_market"],
|
|
allowed_sectors=["Technology"],
|
|
direction="long_only",
|
|
max_holding_days=5,
|
|
engine_risk_budget_pct=0.05,
|
|
per_trade_risk_pct_override=0.01,
|
|
reaction_day_return_min=0.10,
|
|
close_location_min=0.75,
|
|
volume_ratio_min=2.0,
|
|
gap_size_min=0.04,
|
|
min_market_cap_proxy=40_000_000_000.0,
|
|
document_quality_score_min=0.5,
|
|
parse_confidence_overall_min=0.5,
|
|
leader_follower_lookahead_days=3,
|
|
leader_follower_min_days_to_event=2,
|
|
leader_follower_hold_buffer_days=1,
|
|
leader_follower_calendar_mode="pit_calendar",
|
|
proxy_reaction_day_return_max=0.04,
|
|
proxy_gap_size_max=0.03,
|
|
proxy_close_location_max=0.80,
|
|
proxy_avg_dollar_volume_min=10_000_000.0,
|
|
)
|
|
],
|
|
)
|
|
config = BacktestConfig(
|
|
strategy_name="leader_follower_preentry_pit_smoke",
|
|
dataset_snapshot_id="test_snapshot",
|
|
earnings_calendar_pit_path=str(calendar_path),
|
|
universe=UniverseConfig(min_price=5.0, min_avg_dollar_volume=100_000.0),
|
|
signal=SignalConfig(score_threshold=0.50, max_candidates_per_day=5),
|
|
risk=RiskConfig(
|
|
per_trade_risk_pct=0.01,
|
|
max_daily_new_risk_pct=0.05,
|
|
max_positions=10,
|
|
max_positions_per_sector=5,
|
|
),
|
|
execution=ExecutionConfig(
|
|
entry_fill_model="next_open",
|
|
exit_fill_model="daily_bar_approximation",
|
|
slippage_bps_base=0.0,
|
|
commission_per_share=0.0,
|
|
same_bar_priority="stop_first_conservative",
|
|
max_holding_days=10,
|
|
),
|
|
reporting=ReportingConfig(
|
|
write_trade_blotter=True,
|
|
write_equity_curve=True,
|
|
write_metrics_summary=True,
|
|
generate_plots=False,
|
|
),
|
|
strategy_engines=manifest.strategy_engines,
|
|
)
|
|
|
|
runner = BacktestRunner(manifest=manifest, config=config, store=store, initial_equity=100_000.0)
|
|
runner._simulation_dates = [signal_date, entry_date, pre_event_date, follower_reaction_date, follower_exec_date]
|
|
runner._next_trading_day = {
|
|
signal_date: entry_date,
|
|
entry_date: pre_event_date,
|
|
pre_event_date: follower_reaction_date,
|
|
follower_reaction_date: follower_exec_date,
|
|
}
|
|
|
|
runner._schedule_leader_follower_candidates(signal_date)
|
|
|
|
scheduled = runner._scheduled_delayed_entries[entry_date]
|
|
assert len(scheduled) == 1
|
|
candidate = scheduled[0]
|
|
assert candidate.symbol == "AMD"
|
|
assert candidate.source_symbol == "NVDA"
|
|
assert candidate.features["leader_follower_upcoming_reaction_date"] == follower_reaction_date.isoformat()
|
|
|
|
def test_leader_follower_engine_uses_oracle_pit_calendar_without_local_file(self):
|
|
from apps.backtester.run import BacktestRunner
|
|
from libs.backtest.domain import (
|
|
BacktestConfig,
|
|
ExecutionConfig,
|
|
ExperimentManifest,
|
|
ReportingConfig,
|
|
RiskConfig,
|
|
SignalConfig,
|
|
StrategyEngineConfig,
|
|
UniverseConfig,
|
|
)
|
|
from libs.backtest.snapshot_store import SnapshotStore
|
|
|
|
class _FakeOracleCalendar:
|
|
def get_known_upcoming_reaction_dates(self, as_of_date, allowed_reaction_dates, symbols=None):
|
|
assert as_of_date == signal_date
|
|
assert follower_reaction_date in allowed_reaction_dates
|
|
assert "AMD" in (symbols or [])
|
|
return {"AMD": follower_reaction_date}
|
|
|
|
signal_date = dt.date(2026, 1, 6)
|
|
entry_date = dt.date(2026, 1, 7)
|
|
pre_event_date = dt.date(2026, 1, 8)
|
|
follower_reaction_date = dt.date(2026, 1, 9)
|
|
follower_exec_date = dt.date(2026, 1, 12)
|
|
|
|
store = SnapshotStore(
|
|
candidates_by_exec_date={
|
|
entry_date: [
|
|
{
|
|
"event_id": "LEADER::NVDA",
|
|
"symbol": "NVDA",
|
|
"execution_date": entry_date,
|
|
"entry_date": str(entry_date),
|
|
"event_date": str(signal_date),
|
|
"event_close": 120.0,
|
|
"entry_price": 121.0,
|
|
"score": 0.86,
|
|
"sector": "Technology",
|
|
"event_type": "earnings_release",
|
|
"event_direction": "bullish",
|
|
"guidance_status": "raised",
|
|
"event_timestamp": "2026-01-06T21:00:00+00:00",
|
|
"filing_time_bucket": "post_market",
|
|
"reaction_date": str(signal_date),
|
|
"reaction_day_return": 0.14,
|
|
"close_location": 0.88,
|
|
"volume_ratio": 3.4,
|
|
"gap_size": 0.07,
|
|
"avg_dollar_volume": 50_000_000.0,
|
|
"market_cap_proxy": 100_000_000_000.0,
|
|
"document_quality_score": 0.80,
|
|
"parse_confidence_overall": 0.82,
|
|
"atr_14": 4.0,
|
|
}
|
|
],
|
|
},
|
|
bars_by_symbol_date={
|
|
"NVDA": {
|
|
signal_date: {"date": signal_date, "open": 108.0, "high": 121.0, "low": 107.0, "close": 120.0, "volume": 4_000_000},
|
|
entry_date: {"date": entry_date, "open": 121.0, "high": 123.0, "low": 118.0, "close": 122.0, "volume": 3_000_000},
|
|
},
|
|
"AMD": {
|
|
dt.date(2026, 1, 5): {"date": dt.date(2026, 1, 5), "open": 49.5, "high": 50.5, "low": 49.0, "close": 50.0, "volume": 1_000_000},
|
|
signal_date: {"date": signal_date, "open": 50.2, "high": 51.6, "low": 49.8, "close": 51.0, "volume": 1_100_000},
|
|
entry_date: {"date": entry_date, "open": 51.1, "high": 52.0, "low": 50.8, "close": 51.7, "volume": 1_050_000},
|
|
pre_event_date: {"date": pre_event_date, "open": 51.9, "high": 52.6, "low": 51.4, "close": 52.3, "volume": 1_030_000},
|
|
follower_reaction_date: {"date": follower_reaction_date, "open": 52.4, "high": 53.4, "low": 51.8, "close": 53.0, "volume": 1_200_000},
|
|
follower_exec_date: {"date": follower_exec_date, "open": 54.5, "high": 55.0, "low": 53.5, "close": 54.0, "volume": 1_250_000},
|
|
},
|
|
},
|
|
)
|
|
manifest = ExperimentManifest(
|
|
experiment_name="leader_follower_preentry_oracle_pit_smoke",
|
|
dataset_snapshot_id="test_snapshot",
|
|
base_config="configs/backtest/defaults.json",
|
|
overrides={},
|
|
strategy_engines=[
|
|
StrategyEngineConfig(
|
|
engine_id="leader_follower_preentry",
|
|
synthetic_only=True,
|
|
entry_timing_policy="next_open",
|
|
event_types=["earnings_release"],
|
|
event_directions=["bullish"],
|
|
guidance_statuses=["raised"],
|
|
filing_time_buckets=["post_market"],
|
|
allowed_sectors=["Technology"],
|
|
direction="long_only",
|
|
max_holding_days=5,
|
|
engine_risk_budget_pct=0.05,
|
|
per_trade_risk_pct_override=0.01,
|
|
reaction_day_return_min=0.10,
|
|
close_location_min=0.75,
|
|
volume_ratio_min=2.0,
|
|
gap_size_min=0.04,
|
|
min_market_cap_proxy=40_000_000_000.0,
|
|
document_quality_score_min=0.5,
|
|
parse_confidence_overall_min=0.5,
|
|
leader_follower_lookahead_days=3,
|
|
leader_follower_min_days_to_event=2,
|
|
leader_follower_hold_buffer_days=1,
|
|
leader_follower_calendar_mode="pit_calendar",
|
|
proxy_reaction_day_return_max=0.04,
|
|
proxy_gap_size_max=0.03,
|
|
proxy_close_location_max=0.80,
|
|
proxy_avg_dollar_volume_min=10_000_000.0,
|
|
)
|
|
],
|
|
)
|
|
config = BacktestConfig(
|
|
strategy_name="leader_follower_preentry_oracle_pit_smoke",
|
|
dataset_snapshot_id="test_snapshot",
|
|
universe=UniverseConfig(min_price=5.0, min_avg_dollar_volume=100_000.0),
|
|
signal=SignalConfig(score_threshold=0.50, max_candidates_per_day=5),
|
|
risk=RiskConfig(
|
|
per_trade_risk_pct=0.01,
|
|
max_daily_new_risk_pct=0.05,
|
|
max_positions=10,
|
|
max_positions_per_sector=5,
|
|
),
|
|
execution=ExecutionConfig(
|
|
entry_fill_model="next_open",
|
|
exit_fill_model="daily_bar_approximation",
|
|
slippage_bps_base=0.0,
|
|
commission_per_share=0.0,
|
|
same_bar_priority="stop_first_conservative",
|
|
max_holding_days=10,
|
|
),
|
|
reporting=ReportingConfig(
|
|
write_trade_blotter=True,
|
|
write_equity_curve=True,
|
|
write_metrics_summary=True,
|
|
generate_plots=False,
|
|
),
|
|
strategy_engines=manifest.strategy_engines,
|
|
)
|
|
|
|
runner = BacktestRunner(manifest=manifest, config=config, store=store, initial_equity=100_000.0)
|
|
runner._pit_earnings_calendar = None
|
|
runner._oracle_pit_earnings_calendar = _FakeOracleCalendar()
|
|
runner._simulation_dates = [signal_date, entry_date, pre_event_date, follower_reaction_date, follower_exec_date]
|
|
runner._next_trading_day = {
|
|
signal_date: entry_date,
|
|
entry_date: pre_event_date,
|
|
pre_event_date: follower_reaction_date,
|
|
follower_reaction_date: follower_exec_date,
|
|
}
|
|
|
|
runner._schedule_leader_follower_candidates(signal_date)
|
|
|
|
scheduled = runner._scheduled_delayed_entries[entry_date]
|
|
assert len(scheduled) == 1
|
|
candidate = scheduled[0]
|
|
assert candidate.symbol == "AMD"
|
|
assert candidate.source_symbol == "NVDA"
|
|
assert candidate.features["leader_follower_upcoming_reaction_date"] == follower_reaction_date.isoformat()
|
|
|
|
def test_leader_follower_legacy_mode_ignores_pit_calendar(self, tmp_path):
|
|
from apps.backtester.run import BacktestRunner
|
|
from libs.backtest.domain import (
|
|
BacktestConfig,
|
|
ExecutionConfig,
|
|
ExperimentManifest,
|
|
ReportingConfig,
|
|
RiskConfig,
|
|
SignalConfig,
|
|
StrategyEngineConfig,
|
|
UniverseConfig,
|
|
)
|
|
from libs.backtest.earnings_calendar import load_pit_earnings_calendar
|
|
from libs.backtest.snapshot_store import SnapshotStore
|
|
|
|
signal_date = dt.date(2026, 1, 6)
|
|
entry_date = dt.date(2026, 1, 7)
|
|
pre_event_date = dt.date(2026, 1, 8)
|
|
follower_reaction_date = dt.date(2026, 1, 9)
|
|
follower_exec_date = dt.date(2026, 1, 12)
|
|
|
|
calendar_path = tmp_path / "earnings_calendar_pit.parquet"
|
|
pq.write_table(
|
|
pa.Table.from_pylist(
|
|
[
|
|
{
|
|
"symbol": "BABA",
|
|
"as_of_date": signal_date.isoformat(),
|
|
"expected_reaction_date": follower_reaction_date.isoformat(),
|
|
"expected_event_date": pre_event_date.isoformat(),
|
|
"filing_time_bucket": "post_market",
|
|
"source": "unit_test",
|
|
}
|
|
]
|
|
),
|
|
calendar_path,
|
|
)
|
|
load_pit_earnings_calendar.cache_clear()
|
|
|
|
store = SnapshotStore(
|
|
candidates_by_exec_date={
|
|
entry_date: [
|
|
{
|
|
"event_id": "LEADER::NVDA",
|
|
"symbol": "NVDA",
|
|
"execution_date": entry_date,
|
|
"entry_date": str(entry_date),
|
|
"event_date": str(signal_date),
|
|
"event_close": 120.0,
|
|
"entry_price": 121.0,
|
|
"score": 0.86,
|
|
"sector": "Technology",
|
|
"event_type": "earnings_release",
|
|
"event_direction": "bullish",
|
|
"guidance_status": "raised",
|
|
"event_timestamp": "2026-01-06T21:00:00+00:00",
|
|
"filing_time_bucket": "post_market",
|
|
"reaction_date": str(signal_date),
|
|
"reaction_day_return": 0.14,
|
|
"close_location": 0.88,
|
|
"volume_ratio": 3.4,
|
|
"gap_size": 0.07,
|
|
"avg_dollar_volume": 50_000_000.0,
|
|
"market_cap_proxy": 100_000_000_000.0,
|
|
"document_quality_score": 0.80,
|
|
"parse_confidence_overall": 0.82,
|
|
"atr_14": 4.0,
|
|
}
|
|
],
|
|
follower_exec_date: [
|
|
{
|
|
"event_id": "FOLLOWER::AMD",
|
|
"symbol": "AMD",
|
|
"execution_date": follower_exec_date,
|
|
"entry_date": str(follower_exec_date),
|
|
"event_date": str(follower_reaction_date),
|
|
"event_close": 51.5,
|
|
"entry_price": 52.0,
|
|
"score": 0.40,
|
|
"sector": "Technology",
|
|
"event_type": "earnings_release",
|
|
"event_direction": "unknown",
|
|
"guidance_status": "not_provided",
|
|
"event_timestamp": "2026-01-09T21:00:00+00:00",
|
|
"filing_time_bucket": "post_market",
|
|
"reaction_date": str(follower_reaction_date),
|
|
"reaction_day_return": 0.03,
|
|
"close_location": 0.55,
|
|
"volume_ratio": 1.4,
|
|
"gap_size": 0.01,
|
|
"avg_dollar_volume": 30_000_000.0,
|
|
"market_cap_proxy": 60_000_000_000.0,
|
|
"document_quality_score": 0.60,
|
|
"parse_confidence_overall": 0.60,
|
|
"atr_14": 2.5,
|
|
}
|
|
],
|
|
},
|
|
bars_by_symbol_date={
|
|
"NVDA": {
|
|
signal_date: {"date": signal_date, "open": 108.0, "high": 121.0, "low": 107.0, "close": 120.0, "volume": 4_000_000},
|
|
entry_date: {"date": entry_date, "open": 121.0, "high": 123.0, "low": 118.0, "close": 122.0, "volume": 3_000_000},
|
|
},
|
|
"AMD": {
|
|
dt.date(2026, 1, 5): {"date": dt.date(2026, 1, 5), "open": 49.5, "high": 50.5, "low": 49.0, "close": 50.0, "volume": 1_000_000},
|
|
signal_date: {"date": signal_date, "open": 50.2, "high": 51.6, "low": 49.8, "close": 51.0, "volume": 1_100_000},
|
|
entry_date: {"date": entry_date, "open": 51.1, "high": 52.0, "low": 50.8, "close": 51.7, "volume": 1_050_000},
|
|
pre_event_date: {"date": pre_event_date, "open": 51.9, "high": 52.6, "low": 51.4, "close": 52.3, "volume": 1_030_000},
|
|
follower_reaction_date: {"date": follower_reaction_date, "open": 52.4, "high": 53.4, "low": 51.8, "close": 53.0, "volume": 1_200_000},
|
|
follower_exec_date: {"date": follower_exec_date, "open": 54.5, "high": 55.0, "low": 53.5, "close": 54.0, "volume": 1_250_000},
|
|
},
|
|
},
|
|
)
|
|
manifest = ExperimentManifest(
|
|
experiment_name="leader_follower_legacy_mode",
|
|
dataset_snapshot_id="test_snapshot",
|
|
base_config="configs/backtest/defaults.json",
|
|
overrides={},
|
|
strategy_engines=[
|
|
StrategyEngineConfig(
|
|
engine_id="leader_follower_preentry",
|
|
synthetic_only=True,
|
|
entry_timing_policy="next_open",
|
|
event_types=["earnings_release"],
|
|
event_directions=["bullish"],
|
|
guidance_statuses=["raised"],
|
|
filing_time_buckets=["post_market"],
|
|
allowed_sectors=["Technology"],
|
|
direction="long_only",
|
|
reaction_day_return_min=0.10,
|
|
close_location_min=0.75,
|
|
volume_ratio_min=2.0,
|
|
gap_size_min=0.04,
|
|
min_market_cap_proxy=40_000_000_000.0,
|
|
document_quality_score_min=0.5,
|
|
parse_confidence_overall_min=0.5,
|
|
leader_follower_lookahead_days=3,
|
|
leader_follower_min_days_to_event=2,
|
|
leader_follower_hold_buffer_days=1,
|
|
proxy_reaction_day_return_max=0.04,
|
|
proxy_gap_size_max=0.03,
|
|
proxy_close_location_max=0.80,
|
|
proxy_avg_dollar_volume_min=10_000_000.0,
|
|
)
|
|
],
|
|
)
|
|
config = BacktestConfig(
|
|
strategy_name="leader_follower_legacy_mode",
|
|
dataset_snapshot_id="test_snapshot",
|
|
earnings_calendar_pit_path=str(calendar_path),
|
|
universe=UniverseConfig(min_price=5.0, min_avg_dollar_volume=100_000.0),
|
|
signal=SignalConfig(score_threshold=0.50, max_candidates_per_day=5),
|
|
risk=RiskConfig(
|
|
per_trade_risk_pct=0.01,
|
|
max_daily_new_risk_pct=0.05,
|
|
max_positions=10,
|
|
max_positions_per_sector=5,
|
|
),
|
|
execution=ExecutionConfig(
|
|
entry_fill_model="next_open",
|
|
exit_fill_model="daily_bar_approximation",
|
|
slippage_bps_base=0.0,
|
|
commission_per_share=0.0,
|
|
same_bar_priority="stop_first_conservative",
|
|
max_holding_days=10,
|
|
),
|
|
reporting=ReportingConfig(
|
|
write_trade_blotter=True,
|
|
write_equity_curve=True,
|
|
write_metrics_summary=True,
|
|
generate_plots=False,
|
|
),
|
|
strategy_engines=manifest.strategy_engines,
|
|
)
|
|
|
|
runner = BacktestRunner(manifest=manifest, config=config, store=store, initial_equity=100_000.0)
|
|
runner._simulation_dates = [signal_date, entry_date, pre_event_date, follower_reaction_date, follower_exec_date]
|
|
runner._next_trading_day = {
|
|
signal_date: entry_date,
|
|
entry_date: pre_event_date,
|
|
pre_event_date: follower_reaction_date,
|
|
follower_reaction_date: follower_exec_date,
|
|
}
|
|
|
|
runner._schedule_leader_follower_candidates(signal_date)
|
|
scheduled = runner._scheduled_delayed_entries[entry_date]
|
|
assert len(scheduled) == 1
|
|
assert scheduled[0].symbol == "AMD"
|
|
|
|
def test_leader_follower_engine_supports_engine_local_peer_overrides(self):
|
|
from apps.backtester.run import BacktestRunner
|
|
from libs.backtest.domain import (
|
|
BacktestConfig,
|
|
ExecutionConfig,
|
|
ExperimentManifest,
|
|
ReportingConfig,
|
|
RiskConfig,
|
|
SignalConfig,
|
|
StrategyEngineConfig,
|
|
UniverseConfig,
|
|
)
|
|
from libs.backtest.snapshot_store import SnapshotStore
|
|
|
|
signal_date = dt.date(2025, 11, 5)
|
|
entry_date = dt.date(2025, 11, 6)
|
|
pre_event_date = dt.date(2025, 11, 7)
|
|
follower_reaction_date = dt.date(2025, 11, 10)
|
|
follower_exec_date = dt.date(2025, 11, 11)
|
|
|
|
store = SnapshotStore(
|
|
candidates_by_exec_date={
|
|
entry_date: [
|
|
{
|
|
"event_id": "LEADER::NVDA",
|
|
"symbol": "NVDA",
|
|
"execution_date": entry_date,
|
|
"entry_date": str(entry_date),
|
|
"event_date": str(signal_date),
|
|
"event_close": 130.0,
|
|
"entry_price": 131.0,
|
|
"score": 0.90,
|
|
"sector": "Technology",
|
|
"event_type": "earnings_release",
|
|
"event_direction": "bullish",
|
|
"guidance_status": "raised",
|
|
"event_timestamp": "2025-11-05T21:00:00+00:00",
|
|
"filing_time_bucket": "post_market",
|
|
"reaction_date": str(signal_date),
|
|
"reaction_day_return": 0.15,
|
|
"close_location": 0.86,
|
|
"volume_ratio": 3.0,
|
|
"gap_size": 0.06,
|
|
"avg_dollar_volume": 200_000_000.0,
|
|
"market_cap_proxy": 2_000_000_000_000.0,
|
|
"document_quality_score": 0.85,
|
|
"parse_confidence_overall": 0.90,
|
|
"atr_14": 4.0,
|
|
}
|
|
],
|
|
},
|
|
bars_by_symbol_date={
|
|
"NVDA": {
|
|
signal_date: {"date": signal_date, "open": 120.0, "high": 132.0, "low": 118.0, "close": 130.0, "volume": 3_000_000},
|
|
entry_date: {"date": entry_date, "open": 131.0, "high": 133.0, "low": 128.0, "close": 132.0, "volume": 2_500_000},
|
|
},
|
|
"COHR": {
|
|
signal_date: {"date": signal_date, "open": 74.0, "high": 76.0, "low": 73.0, "close": 75.5, "volume": 500_000},
|
|
entry_date: {"date": entry_date, "open": 75.6, "high": 76.4, "low": 75.0, "close": 76.0, "volume": 480_000},
|
|
pre_event_date: {"date": pre_event_date, "open": 76.1, "high": 76.9, "low": 75.4, "close": 76.3, "volume": 470_000},
|
|
follower_reaction_date: {"date": follower_reaction_date, "open": 76.4, "high": 77.1, "low": 75.8, "close": 76.8, "volume": 510_000},
|
|
follower_exec_date: {"date": follower_exec_date, "open": 76.9, "high": 77.3, "low": 76.2, "close": 76.7, "volume": 450_000},
|
|
},
|
|
},
|
|
)
|
|
store._market_feature_cache[("COHR", signal_date)] = {
|
|
"reaction_day_return": 0.01,
|
|
"gap_size": 0.005,
|
|
"close_location": 0.55,
|
|
"volume_ratio_20d": 1.2,
|
|
"avg_dollar_volume_20d": 120_000_000.0,
|
|
"event_close": 75.5,
|
|
"atr_14": 2.1,
|
|
}
|
|
|
|
class _FakeOracleCalendar:
|
|
def get_known_upcoming_reaction_dates(self, *, as_of_date, allowed_reaction_dates, symbols):
|
|
assert as_of_date == signal_date
|
|
assert "COHR" in symbols
|
|
return {"COHR": follower_reaction_date}
|
|
|
|
manifest = ExperimentManifest(
|
|
experiment_name="leader_follower_peer_override",
|
|
dataset_snapshot_id="test_snapshot",
|
|
base_config="configs/backtest/defaults.json",
|
|
overrides={},
|
|
strategy_engines=[
|
|
StrategyEngineConfig(
|
|
engine_id="leader_follower_preentry",
|
|
synthetic_only=True,
|
|
entry_timing_policy="next_open",
|
|
event_types=["earnings_release"],
|
|
event_directions=["bullish"],
|
|
guidance_statuses=["raised"],
|
|
filing_time_buckets=["post_market"],
|
|
allowed_sectors=["Technology"],
|
|
direction="long_only",
|
|
reaction_day_return_min=0.10,
|
|
close_location_min=0.75,
|
|
volume_ratio_min=2.0,
|
|
gap_size_min=0.04,
|
|
min_market_cap_proxy=40_000_000_000.0,
|
|
document_quality_score_min=0.5,
|
|
parse_confidence_overall_min=0.5,
|
|
leader_follower_lookahead_days=3,
|
|
leader_follower_min_days_to_event=2,
|
|
leader_follower_hold_buffer_days=1,
|
|
leader_follower_calendar_mode="pit_calendar",
|
|
leader_follower_extra_peer_symbols_by_leader={"NVDA": ["COHR"]},
|
|
leader_follower_allowed_peer_symbols=["COHR"],
|
|
proxy_reaction_day_return_max=0.04,
|
|
proxy_gap_size_max=0.03,
|
|
proxy_close_location_max=0.80,
|
|
proxy_avg_dollar_volume_min=10_000_000.0,
|
|
)
|
|
],
|
|
)
|
|
config = BacktestConfig(
|
|
strategy_name="leader_follower_peer_override",
|
|
dataset_snapshot_id="test_snapshot",
|
|
universe=UniverseConfig(min_price=5.0, min_avg_dollar_volume=100_000.0),
|
|
signal=SignalConfig(score_threshold=0.50, max_candidates_per_day=5),
|
|
risk=RiskConfig(
|
|
per_trade_risk_pct=0.01,
|
|
max_daily_new_risk_pct=0.05,
|
|
max_positions=10,
|
|
max_positions_per_sector=5,
|
|
),
|
|
execution=ExecutionConfig(
|
|
entry_fill_model="next_open",
|
|
exit_fill_model="daily_bar_approximation",
|
|
slippage_bps_base=0.0,
|
|
commission_per_share=0.0,
|
|
same_bar_priority="stop_first_conservative",
|
|
max_holding_days=10,
|
|
),
|
|
reporting=ReportingConfig(
|
|
write_trade_blotter=True,
|
|
write_equity_curve=True,
|
|
write_metrics_summary=True,
|
|
generate_plots=False,
|
|
),
|
|
strategy_engines=manifest.strategy_engines,
|
|
)
|
|
|
|
runner = BacktestRunner(manifest=manifest, config=config, store=store, initial_equity=100_000.0)
|
|
runner._pit_earnings_calendar = None
|
|
runner._oracle_pit_earnings_calendar = _FakeOracleCalendar()
|
|
runner._simulation_dates = [signal_date, entry_date, pre_event_date, follower_reaction_date, follower_exec_date]
|
|
runner._next_trading_day = {
|
|
signal_date: entry_date,
|
|
entry_date: pre_event_date,
|
|
pre_event_date: follower_reaction_date,
|
|
follower_reaction_date: follower_exec_date,
|
|
}
|
|
|
|
runner._schedule_leader_follower_candidates(signal_date)
|
|
scheduled = runner._scheduled_delayed_entries[entry_date]
|
|
assert len(scheduled) == 1
|
|
assert scheduled[0].symbol == "COHR"
|
|
assert scheduled[0].source_symbol == "NVDA"
|
|
|
|
def test_leader_follower_engine_fetches_missing_peer_market_data(self, monkeypatch):
|
|
from apps.backtester.run import BacktestRunner
|
|
from libs.backtest.domain import (
|
|
BacktestConfig,
|
|
ExecutionConfig,
|
|
ExperimentManifest,
|
|
ReportingConfig,
|
|
RiskConfig,
|
|
SignalConfig,
|
|
StrategyEngineConfig,
|
|
UniverseConfig,
|
|
)
|
|
from libs.backtest.snapshot_store import SnapshotStore
|
|
|
|
signal_date = dt.date(2025, 11, 5)
|
|
entry_date = dt.date(2025, 11, 6)
|
|
pre_event_date = dt.date(2025, 11, 7)
|
|
follower_reaction_date = dt.date(2025, 11, 10)
|
|
follower_exec_date = dt.date(2025, 11, 11)
|
|
|
|
store = SnapshotStore(
|
|
candidates_by_exec_date={
|
|
entry_date: [
|
|
{
|
|
"event_id": "LEADER::NVDA",
|
|
"symbol": "NVDA",
|
|
"execution_date": entry_date,
|
|
"entry_date": str(entry_date),
|
|
"event_date": str(signal_date),
|
|
"event_close": 130.0,
|
|
"entry_price": 131.0,
|
|
"score": 0.90,
|
|
"sector": "Technology",
|
|
"event_type": "earnings_release",
|
|
"event_direction": "bullish",
|
|
"guidance_status": "raised",
|
|
"event_timestamp": "2025-11-05T21:00:00+00:00",
|
|
"filing_time_bucket": "post_market",
|
|
"reaction_date": str(signal_date),
|
|
"reaction_day_return": 0.15,
|
|
"close_location": 0.86,
|
|
"volume_ratio": 3.0,
|
|
"gap_size": 0.06,
|
|
"avg_dollar_volume": 200_000_000.0,
|
|
"market_cap_proxy": 2_000_000_000_000.0,
|
|
"document_quality_score": 0.85,
|
|
"parse_confidence_overall": 0.90,
|
|
"atr_14": 4.0,
|
|
}
|
|
],
|
|
},
|
|
bars_by_symbol_date={
|
|
"NVDA": {
|
|
signal_date: {"date": signal_date, "open": 120.0, "high": 132.0, "low": 118.0, "close": 130.0, "volume": 3_000_000},
|
|
entry_date: {"date": entry_date, "open": 131.0, "high": 133.0, "low": 128.0, "close": 132.0, "volume": 2_500_000},
|
|
},
|
|
},
|
|
)
|
|
|
|
class _FakeOracleCalendar:
|
|
def get_known_upcoming_reaction_dates(self, *, as_of_date, allowed_reaction_dates, symbols):
|
|
assert as_of_date == signal_date
|
|
assert "COHR" in symbols
|
|
return {"COHR": follower_reaction_date}
|
|
|
|
async def _fake_fetch_price_data(symbols, date_range, oracle_url, concurrency=8):
|
|
assert "COHR" in symbols
|
|
start_date, end_date = date_range
|
|
current = start_date
|
|
price = 60.0
|
|
bars: dict[dt.date, dict[str, float | int | dt.date]] = {}
|
|
while current <= end_date:
|
|
if current.weekday() < 5:
|
|
bars[current] = {
|
|
"date": current,
|
|
"open": price,
|
|
"high": price * 1.01,
|
|
"low": price * 0.99,
|
|
"close": price * 1.002,
|
|
"volume": 2_000_000,
|
|
}
|
|
price += 0.2
|
|
current += dt.timedelta(days=1)
|
|
return {"COHR": bars}, {"COHR": 120_000_000.0}
|
|
|
|
monkeypatch.setattr(SnapshotStore, "_fetch_price_data", staticmethod(_fake_fetch_price_data))
|
|
|
|
manifest = ExperimentManifest(
|
|
experiment_name="leader_follower_peer_fetch",
|
|
dataset_snapshot_id="test_snapshot",
|
|
base_config="configs/backtest/defaults.json",
|
|
overrides={},
|
|
strategy_engines=[
|
|
StrategyEngineConfig(
|
|
engine_id="leader_follower_preentry",
|
|
synthetic_only=True,
|
|
entry_timing_policy="next_open",
|
|
event_types=["earnings_release"],
|
|
event_directions=["bullish"],
|
|
guidance_statuses=["raised"],
|
|
filing_time_buckets=["post_market"],
|
|
allowed_sectors=["Technology"],
|
|
direction="long_only",
|
|
reaction_day_return_min=0.10,
|
|
close_location_min=0.75,
|
|
volume_ratio_min=2.0,
|
|
gap_size_min=0.04,
|
|
min_market_cap_proxy=40_000_000_000.0,
|
|
document_quality_score_min=0.5,
|
|
parse_confidence_overall_min=0.5,
|
|
leader_follower_lookahead_days=3,
|
|
leader_follower_min_days_to_event=2,
|
|
leader_follower_hold_buffer_days=1,
|
|
leader_follower_calendar_mode="pit_calendar",
|
|
leader_follower_extra_peer_symbols_by_leader={"NVDA": ["COHR"]},
|
|
leader_follower_allowed_peer_symbols=["COHR"],
|
|
proxy_reaction_day_return_max=0.05,
|
|
proxy_gap_size_max=0.04,
|
|
proxy_close_location_min=0.3,
|
|
proxy_close_location_max=0.85,
|
|
proxy_avg_dollar_volume_min=10_000_000.0,
|
|
)
|
|
],
|
|
)
|
|
config = BacktestConfig(
|
|
strategy_name="leader_follower_peer_fetch",
|
|
dataset_snapshot_id="test_snapshot",
|
|
universe=UniverseConfig(min_price=5.0, min_avg_dollar_volume=100_000.0),
|
|
signal=SignalConfig(score_threshold=0.50, max_candidates_per_day=5),
|
|
risk=RiskConfig(
|
|
per_trade_risk_pct=0.01,
|
|
max_daily_new_risk_pct=0.05,
|
|
max_positions=10,
|
|
max_positions_per_sector=5,
|
|
),
|
|
execution=ExecutionConfig(
|
|
entry_fill_model="next_open",
|
|
exit_fill_model="daily_bar_approximation",
|
|
slippage_bps_base=0.0,
|
|
commission_per_share=0.0,
|
|
same_bar_priority="stop_first_conservative",
|
|
max_holding_days=10,
|
|
),
|
|
reporting=ReportingConfig(
|
|
write_trade_blotter=True,
|
|
write_equity_curve=True,
|
|
write_metrics_summary=True,
|
|
generate_plots=False,
|
|
),
|
|
strategy_engines=manifest.strategy_engines,
|
|
)
|
|
|
|
runner = BacktestRunner(manifest=manifest, config=config, store=store, initial_equity=100_000.0)
|
|
runner._pit_earnings_calendar = None
|
|
runner._oracle_pit_earnings_calendar = _FakeOracleCalendar()
|
|
runner._simulation_dates = [signal_date, entry_date, pre_event_date, follower_reaction_date, follower_exec_date]
|
|
runner._next_trading_day = {
|
|
signal_date: entry_date,
|
|
entry_date: pre_event_date,
|
|
pre_event_date: follower_reaction_date,
|
|
follower_reaction_date: follower_exec_date,
|
|
}
|
|
|
|
runner._schedule_leader_follower_candidates(signal_date)
|
|
scheduled = runner._scheduled_delayed_entries[entry_date]
|
|
assert len(scheduled) == 1
|
|
assert scheduled[0].symbol == "COHR"
|
|
assert store.get_latest_bar_on_or_before("COHR", signal_date) is not None
|
|
|
|
def test_macro_bullish_engine_respects_breadth_gate(self, tmp_path):
|
|
from apps.backtester.run import BacktestRunner
|
|
from libs.backtest.domain import (
|
|
BacktestConfig,
|
|
Candidate,
|
|
ExecutionConfig,
|
|
ExperimentManifest,
|
|
ReportingConfig,
|
|
RiskConfig,
|
|
SignalConfig,
|
|
StrategyEngineConfig,
|
|
UniverseConfig,
|
|
)
|
|
from libs.backtest.snapshot_store import SnapshotStore
|
|
|
|
signal_date = dt.date(2026, 1, 6)
|
|
next_date = dt.date(2026, 1, 7)
|
|
store = SnapshotStore(
|
|
candidates_by_exec_date={next_date: []},
|
|
bars_by_symbol_date={
|
|
"QQQ": {
|
|
dt.date(2026, 1, 5): {"date": dt.date(2026, 1, 5), "open": 100.0, "high": 101.0, "low": 99.0, "close": 100.0, "volume": 100},
|
|
signal_date: {"date": signal_date, "open": 102.0, "high": 107.0, "low": 101.5, "close": 106.0, "volume": 400},
|
|
next_date: {"date": next_date, "open": 106.5, "high": 108.0, "low": 105.0, "close": 107.5, "volume": 350},
|
|
}
|
|
},
|
|
macro_by_date={signal_date: {"VIXCLS": 22.0}},
|
|
)
|
|
manifest = ExperimentManifest(
|
|
experiment_name="macro_bullish_qqq_breadth_block",
|
|
dataset_snapshot_id="test_snapshot",
|
|
base_config="configs/backtest/defaults.json",
|
|
overrides={},
|
|
strategy_engines=[
|
|
StrategyEngineConfig(
|
|
engine_id="macro_bullish_qqq",
|
|
synthetic_only=True,
|
|
entry_timing_policy="next_open",
|
|
macro_long_symbol="QQQ",
|
|
macro_long_reaction_day_return_min=0.015,
|
|
macro_long_volume_ratio_min=2.0,
|
|
macro_long_gap_size_min=0.01,
|
|
macro_long_close_location_min=0.75,
|
|
macro_long_min_daily_candidate_count=3,
|
|
macro_long_min_unique_sector_count=2,
|
|
)
|
|
],
|
|
)
|
|
config = BacktestConfig(
|
|
strategy_name="macro_bullish_qqq_breadth_block",
|
|
dataset_snapshot_id="test_snapshot",
|
|
universe=UniverseConfig(min_price=5.0, min_avg_dollar_volume=100_000.0),
|
|
signal=SignalConfig(score_threshold=0.50, max_candidates_per_day=5),
|
|
risk=RiskConfig(
|
|
per_trade_risk_pct=0.01,
|
|
max_daily_new_risk_pct=0.05,
|
|
max_positions=10,
|
|
max_positions_per_sector=5,
|
|
),
|
|
execution=ExecutionConfig(
|
|
entry_fill_model="next_open",
|
|
exit_fill_model="daily_bar_approximation",
|
|
slippage_bps_base=0.0,
|
|
commission_per_share=0.0,
|
|
same_bar_priority="stop_first_conservative",
|
|
max_holding_days=10,
|
|
),
|
|
reporting=ReportingConfig(
|
|
write_trade_blotter=True,
|
|
write_equity_curve=True,
|
|
write_metrics_summary=True,
|
|
generate_plots=False,
|
|
),
|
|
strategy_engines=manifest.strategy_engines,
|
|
)
|
|
|
|
runner = BacktestRunner(manifest=manifest, config=config, store=store, initial_equity=100_000.0)
|
|
runner._simulation_dates = [signal_date, next_date]
|
|
runner._next_trading_day = {signal_date: next_date}
|
|
runner._recent_scored_candidates[signal_date] = [
|
|
Candidate(
|
|
event_id="BREADTH::1",
|
|
symbol="AAPL",
|
|
score=0.8,
|
|
sector="Technology",
|
|
event_type="earnings_release",
|
|
event_timestamp=dt.datetime(2026, 1, 6, 21, 0, tzinfo=_UTC),
|
|
filing_time_bucket="post_market",
|
|
reaction_date=signal_date,
|
|
execution_date=next_date,
|
|
entry_price_est=100.0,
|
|
avg_dollar_volume=1_000_000.0,
|
|
atr_14=2.0,
|
|
score_bucket="high",
|
|
engine_id="core",
|
|
),
|
|
Candidate(
|
|
event_id="BREADTH::2",
|
|
symbol="LLY",
|
|
score=0.78,
|
|
sector="Health Care",
|
|
event_type="earnings_release",
|
|
event_timestamp=dt.datetime(2026, 1, 6, 21, 0, tzinfo=_UTC),
|
|
filing_time_bucket="post_market",
|
|
reaction_date=signal_date,
|
|
execution_date=next_date,
|
|
entry_price_est=100.0,
|
|
avg_dollar_volume=1_000_000.0,
|
|
atr_14=2.0,
|
|
score_bucket="high",
|
|
engine_id="core",
|
|
),
|
|
]
|
|
|
|
runner._schedule_macro_long_candidates(signal_date)
|
|
|
|
assert runner._scheduled_delayed_entries[next_date] == []
|
|
|
|
def test_macro_bullish_engine_uses_etf_basket_breadth(self, tmp_path):
|
|
from apps.backtester.run import BacktestRunner
|
|
from libs.backtest.domain import (
|
|
BacktestConfig,
|
|
ExecutionConfig,
|
|
ExperimentManifest,
|
|
ReportingConfig,
|
|
RiskConfig,
|
|
SignalConfig,
|
|
StrategyEngineConfig,
|
|
UniverseConfig,
|
|
)
|
|
from libs.backtest.snapshot_store import SnapshotStore
|
|
|
|
signal_date = dt.date(2026, 1, 6)
|
|
next_date = dt.date(2026, 1, 7)
|
|
store = SnapshotStore(
|
|
candidates_by_exec_date={next_date: []},
|
|
bars_by_symbol_date={
|
|
"SPY": {
|
|
dt.date(2026, 1, 5): {"date": dt.date(2026, 1, 5), "open": 100.0, "high": 101.0, "low": 99.0, "close": 100.0, "volume": 100},
|
|
signal_date: {"date": signal_date, "open": 101.0, "high": 103.0, "low": 100.5, "close": 102.0, "volume": 130},
|
|
},
|
|
"QQQ": {
|
|
dt.date(2026, 1, 5): {"date": dt.date(2026, 1, 5), "open": 100.0, "high": 101.0, "low": 99.0, "close": 100.0, "volume": 100},
|
|
signal_date: {"date": signal_date, "open": 102.0, "high": 107.0, "low": 101.5, "close": 106.0, "volume": 180},
|
|
next_date: {"date": next_date, "open": 106.5, "high": 108.0, "low": 105.0, "close": 107.5, "volume": 170},
|
|
},
|
|
"XLK": {
|
|
dt.date(2026, 1, 5): {"date": dt.date(2026, 1, 5), "open": 200.0, "high": 201.0, "low": 199.0, "close": 200.0, "volume": 100},
|
|
signal_date: {"date": signal_date, "open": 202.0, "high": 209.0, "low": 201.5, "close": 208.0, "volume": 160},
|
|
},
|
|
"SMH": {
|
|
dt.date(2026, 1, 5): {"date": dt.date(2026, 1, 5), "open": 300.0, "high": 301.0, "low": 299.0, "close": 300.0, "volume": 100},
|
|
signal_date: {"date": signal_date, "open": 304.0, "high": 314.0, "low": 303.0, "close": 312.0, "volume": 175},
|
|
},
|
|
"IWM": {
|
|
dt.date(2026, 1, 5): {"date": dt.date(2026, 1, 5), "open": 150.0, "high": 151.0, "low": 149.0, "close": 150.0, "volume": 100},
|
|
signal_date: {"date": signal_date, "open": 149.5, "high": 151.0, "low": 148.0, "close": 149.0, "volume": 95},
|
|
},
|
|
},
|
|
macro_by_date={signal_date: {"VIXCLS": 22.0}},
|
|
)
|
|
manifest = ExperimentManifest(
|
|
experiment_name="macro_bullish_etf_breadth",
|
|
dataset_snapshot_id="test_snapshot",
|
|
base_config="configs/backtest/defaults.json",
|
|
overrides={},
|
|
strategy_engines=[
|
|
StrategyEngineConfig(
|
|
engine_id="macro_bullish_qqq_etf_breadth",
|
|
synthetic_only=True,
|
|
entry_timing_policy="next_open",
|
|
max_holding_days=8,
|
|
engine_risk_budget_pct=0.25,
|
|
per_trade_risk_pct_override=0.01,
|
|
stop_atr_multiplier_override=2.0,
|
|
trailing_warmup_days_override=4,
|
|
macro_long_symbol="QQQ",
|
|
macro_long_reaction_day_return_min=0.015,
|
|
macro_long_gap_size_min=0.01,
|
|
macro_long_close_location_min=0.75,
|
|
macro_long_breadth_symbols=["QQQ", "XLK", "SMH", "IWM"],
|
|
macro_long_min_breadth_count=2,
|
|
macro_long_breadth_reaction_day_return_min=0.015,
|
|
macro_long_breadth_close_location_min=0.70,
|
|
macro_long_leadership_vs_spy_min=0.02,
|
|
)
|
|
],
|
|
)
|
|
config = BacktestConfig(
|
|
strategy_name="macro_bullish_etf_breadth",
|
|
dataset_snapshot_id="test_snapshot",
|
|
universe=UniverseConfig(min_price=5.0, min_avg_dollar_volume=100_000.0),
|
|
signal=SignalConfig(score_threshold=0.50, max_candidates_per_day=5),
|
|
risk=RiskConfig(
|
|
per_trade_risk_pct=0.01,
|
|
max_daily_new_risk_pct=0.05,
|
|
max_positions=10,
|
|
max_positions_per_sector=5,
|
|
),
|
|
execution=ExecutionConfig(
|
|
entry_fill_model="next_open",
|
|
exit_fill_model="daily_bar_approximation",
|
|
slippage_bps_base=0.0,
|
|
commission_per_share=0.0,
|
|
same_bar_priority="stop_first_conservative",
|
|
max_holding_days=10,
|
|
),
|
|
reporting=ReportingConfig(
|
|
write_trade_blotter=True,
|
|
write_equity_curve=True,
|
|
write_metrics_summary=True,
|
|
generate_plots=False,
|
|
),
|
|
strategy_engines=manifest.strategy_engines,
|
|
)
|
|
|
|
runner = BacktestRunner(manifest=manifest, config=config, store=store, initial_equity=100_000.0)
|
|
runner._simulation_dates = [signal_date, next_date]
|
|
runner._next_trading_day = {signal_date: next_date}
|
|
|
|
runner._schedule_macro_long_candidates(signal_date)
|
|
|
|
scheduled = runner._scheduled_delayed_entries[next_date]
|
|
assert len(scheduled) == 1
|
|
candidate = scheduled[0]
|
|
assert candidate.symbol == "QQQ"
|
|
assert candidate.features["macro_long_breadth_count"] == 3
|
|
assert candidate.features["macro_long_breadth_symbols"] == ["QQQ", "XLK", "SMH"]
|
|
assert candidate.features["macro_long_leadership_vs_spy"] == pytest.approx(0.04)
|
|
assert candidate.features["macro_long_event_candidate_count"] == 0
|
|
|
|
def test_macro_bullish_engine_can_trade_strongest_breadth_etf(self, tmp_path):
|
|
from apps.backtester.run import BacktestRunner
|
|
from libs.backtest.domain import (
|
|
BacktestConfig,
|
|
Candidate,
|
|
DailyPortfolioState,
|
|
ExecutionConfig,
|
|
ExperimentManifest,
|
|
OpenPosition,
|
|
PlannedOrder,
|
|
ReportingConfig,
|
|
RiskConfig,
|
|
SignalConfig,
|
|
StrategyEngineConfig,
|
|
UniverseConfig,
|
|
)
|
|
from libs.backtest.snapshot_store import SnapshotStore
|
|
|
|
signal_date = dt.date(2026, 1, 6)
|
|
next_date = dt.date(2026, 1, 7)
|
|
store = SnapshotStore(
|
|
candidates_by_exec_date={next_date: []},
|
|
bars_by_symbol_date={
|
|
"SPY": {
|
|
dt.date(2026, 1, 5): {"date": dt.date(2026, 1, 5), "open": 100.0, "high": 101.0, "low": 99.0, "close": 100.0, "volume": 100},
|
|
signal_date: {"date": signal_date, "open": 101.0, "high": 103.0, "low": 100.5, "close": 102.0, "volume": 130},
|
|
},
|
|
"QQQ": {
|
|
dt.date(2026, 1, 5): {"date": dt.date(2026, 1, 5), "open": 100.0, "high": 101.0, "low": 99.0, "close": 100.0, "volume": 100},
|
|
signal_date: {"date": signal_date, "open": 101.0, "high": 105.0, "low": 100.5, "close": 104.0, "volume": 150},
|
|
next_date: {"date": next_date, "open": 104.5, "high": 106.0, "low": 103.0, "close": 105.0, "volume": 140},
|
|
},
|
|
"XLK": {
|
|
dt.date(2026, 1, 5): {"date": dt.date(2026, 1, 5), "open": 200.0, "high": 201.0, "low": 199.0, "close": 200.0, "volume": 100},
|
|
signal_date: {"date": signal_date, "open": 205.0, "high": 219.0, "low": 204.0, "close": 217.0, "volume": 190},
|
|
next_date: {"date": next_date, "open": 217.5, "high": 221.0, "low": 216.0, "close": 220.0, "volume": 180},
|
|
},
|
|
"SMH": {
|
|
dt.date(2026, 1, 5): {"date": dt.date(2026, 1, 5), "open": 300.0, "high": 301.0, "low": 299.0, "close": 300.0, "volume": 100},
|
|
signal_date: {"date": signal_date, "open": 303.0, "high": 309.0, "low": 302.0, "close": 307.0, "volume": 160},
|
|
next_date: {"date": next_date, "open": 307.5, "high": 310.0, "low": 306.0, "close": 309.0, "volume": 150},
|
|
},
|
|
},
|
|
macro_by_date={signal_date: {"VIXCLS": 22.0}},
|
|
)
|
|
manifest = ExperimentManifest(
|
|
experiment_name="macro_bullish_etf_breadth_leader",
|
|
dataset_snapshot_id="test_snapshot",
|
|
base_config="configs/backtest/defaults.json",
|
|
overrides={},
|
|
strategy_engines=[
|
|
StrategyEngineConfig(
|
|
engine_id="macro_bullish_qqq_etf_breadth_leader",
|
|
synthetic_only=True,
|
|
entry_timing_policy="next_open",
|
|
macro_long_symbol="QQQ",
|
|
macro_long_trade_symbol_mode="leader",
|
|
macro_long_reaction_day_return_min=0.015,
|
|
macro_long_close_location_min=0.60,
|
|
macro_long_breadth_symbols=["QQQ", "XLK", "SMH"],
|
|
macro_long_min_breadth_count=2,
|
|
macro_long_breadth_reaction_day_return_min=0.015,
|
|
macro_long_breadth_close_location_min=0.60,
|
|
macro_long_leadership_vs_spy_min=0.01,
|
|
)
|
|
],
|
|
)
|
|
config = BacktestConfig(
|
|
strategy_name="macro_bullish_etf_breadth_leader",
|
|
dataset_snapshot_id="test_snapshot",
|
|
universe=UniverseConfig(min_price=5.0, min_avg_dollar_volume=100_000.0),
|
|
signal=SignalConfig(score_threshold=0.50, max_candidates_per_day=5),
|
|
risk=RiskConfig(
|
|
per_trade_risk_pct=0.01,
|
|
max_daily_new_risk_pct=0.05,
|
|
max_positions=10,
|
|
max_positions_per_sector=5,
|
|
),
|
|
execution=ExecutionConfig(
|
|
entry_fill_model="next_open",
|
|
exit_fill_model="daily_bar_approximation",
|
|
slippage_bps_base=0.0,
|
|
commission_per_share=0.0,
|
|
same_bar_priority="stop_first_conservative",
|
|
max_holding_days=10,
|
|
),
|
|
reporting=ReportingConfig(
|
|
write_trade_blotter=True,
|
|
write_equity_curve=True,
|
|
write_metrics_summary=True,
|
|
generate_plots=False,
|
|
),
|
|
strategy_engines=manifest.strategy_engines,
|
|
)
|
|
|
|
runner = BacktestRunner(manifest=manifest, config=config, store=store, initial_equity=100_000.0)
|
|
runner._simulation_dates = [signal_date, next_date]
|
|
runner._next_trading_day = {signal_date: next_date}
|
|
|
|
runner._schedule_macro_long_candidates(signal_date)
|
|
|
|
scheduled = runner._scheduled_delayed_entries[next_date]
|
|
assert len(scheduled) == 1
|
|
candidate = scheduled[0]
|
|
assert candidate.symbol == "XLK"
|
|
assert candidate.source_symbol == "QQQ"
|
|
assert candidate.features["macro_long_trade_symbol"] == "XLK"
|
|
assert candidate.features["macro_long_trade_symbol_mode"] == "leader"
|
|
|
|
def test_capital_bucket_reserves_cash_for_macro_sleeve(self, tmp_path):
|
|
from apps.backtester.run import BacktestRunner
|
|
from libs.backtest.domain import (
|
|
BacktestConfig,
|
|
Candidate,
|
|
DailyPortfolioState,
|
|
ExecutionConfig,
|
|
ExperimentManifest,
|
|
OpenPosition,
|
|
PlannedOrder,
|
|
ReportingConfig,
|
|
RiskConfig,
|
|
SignalConfig,
|
|
StrategyEngineConfig,
|
|
UniverseConfig,
|
|
)
|
|
from libs.backtest.snapshot_store import SnapshotStore
|
|
|
|
date = dt.date(2026, 1, 7)
|
|
store = SnapshotStore(
|
|
candidates_by_exec_date={},
|
|
bars_by_symbol_date={
|
|
"XLK": {
|
|
date: {"date": date, "open": 210.0, "high": 212.0, "low": 209.0, "close": 211.0, "volume": 100},
|
|
}
|
|
},
|
|
)
|
|
manifest = ExperimentManifest(
|
|
experiment_name="macro_capital_bucket_reserve",
|
|
dataset_snapshot_id="test_snapshot",
|
|
base_config="configs/backtest/defaults.json",
|
|
overrides={},
|
|
strategy_engines=[
|
|
StrategyEngineConfig(engine_id="core"),
|
|
StrategyEngineConfig(
|
|
engine_id="macro_bullish_qqq_etf_breadth",
|
|
synthetic_only=True,
|
|
capital_bucket_id="macro_etf",
|
|
capital_bucket_allocation_pct=0.05,
|
|
),
|
|
],
|
|
)
|
|
config = BacktestConfig(
|
|
strategy_name="macro_capital_bucket_reserve",
|
|
dataset_snapshot_id="test_snapshot",
|
|
universe=UniverseConfig(min_price=5.0, min_avg_dollar_volume=100_000.0),
|
|
signal=SignalConfig(score_threshold=0.50, max_candidates_per_day=5),
|
|
risk=RiskConfig(
|
|
per_trade_risk_pct=0.01,
|
|
max_daily_new_risk_pct=0.05,
|
|
max_positions=10,
|
|
max_positions_per_sector=5,
|
|
),
|
|
execution=ExecutionConfig(
|
|
entry_fill_model="next_open",
|
|
exit_fill_model="daily_bar_approximation",
|
|
slippage_bps_base=0.0,
|
|
commission_per_share=0.0,
|
|
same_bar_priority="stop_first_conservative",
|
|
max_holding_days=10,
|
|
),
|
|
reporting=ReportingConfig(
|
|
write_trade_blotter=True,
|
|
write_equity_curve=True,
|
|
write_metrics_summary=True,
|
|
generate_plots=False,
|
|
),
|
|
strategy_engines=manifest.strategy_engines,
|
|
)
|
|
|
|
runner = BacktestRunner(manifest=manifest, config=config, store=store, initial_equity=100_000.0)
|
|
portfolio_state = DailyPortfolioState(
|
|
date=date,
|
|
equity=100_000.0,
|
|
sizing_equity=100_000.0,
|
|
cash_available=100_000.0,
|
|
gross_exposure=0.0,
|
|
net_exposure=0.0,
|
|
reserved_risk_budget=0.0,
|
|
unrealized_pnl=0.0,
|
|
realized_pnl=0.0,
|
|
open_positions=[],
|
|
daily_new_risk_used=0.0,
|
|
peak_equity=100_000.0,
|
|
current_drawdown_pct=0.0,
|
|
)
|
|
core_candidate = Candidate(
|
|
event_id="core::1",
|
|
symbol="AAPL",
|
|
score=0.7,
|
|
sector="Technology",
|
|
event_type="earnings_release",
|
|
event_timestamp=dt.datetime(2026, 1, 6, 21, 0, tzinfo=_UTC),
|
|
filing_time_bucket="post_market",
|
|
reaction_date=date,
|
|
execution_date=date,
|
|
entry_price_est=100.0,
|
|
avg_dollar_volume=1_000_000.0,
|
|
atr_14=2.0,
|
|
score_bucket="high",
|
|
engine_id="core",
|
|
)
|
|
macro_candidate = Candidate(
|
|
event_id="macro::1",
|
|
symbol="XLK",
|
|
source_symbol="QQQ",
|
|
score=0.7,
|
|
sector="Technology",
|
|
event_type="macro_bullish_event",
|
|
event_timestamp=dt.datetime(2026, 1, 6, 21, 0, tzinfo=_UTC),
|
|
filing_time_bucket="post_market",
|
|
reaction_date=date,
|
|
execution_date=date,
|
|
entry_price_est=210.0,
|
|
avg_dollar_volume=1_000_000.0,
|
|
atr_14=4.0,
|
|
score_bucket="high",
|
|
engine_id="macro_bullish_qqq_etf_breadth",
|
|
engine_capital_bucket_id="macro_etf",
|
|
engine_capital_bucket_allocation_pct=0.05,
|
|
)
|
|
active_bucket_ids = runner._active_capital_bucket_ids_for_candidates([core_candidate, macro_candidate])
|
|
|
|
core_state = runner._adjust_portfolio_state_for_candidate(
|
|
date=date,
|
|
candidate=core_candidate,
|
|
portfolio_state=portfolio_state,
|
|
active_bucket_ids=active_bucket_ids,
|
|
)
|
|
macro_state = runner._adjust_portfolio_state_for_candidate(
|
|
date=date,
|
|
candidate=macro_candidate,
|
|
portfolio_state=portfolio_state,
|
|
active_bucket_ids=active_bucket_ids,
|
|
)
|
|
|
|
assert core_state.cash_available == pytest.approx(95_000.0)
|
|
assert macro_state.cash_available == pytest.approx(5_000.0)
|
|
assert core_state.sizing_equity == pytest.approx(95_000.0)
|
|
assert macro_state.sizing_equity == pytest.approx(5_000.0)
|
|
|
|
macro_plan = PlannedOrder(
|
|
candidate=macro_candidate,
|
|
shares=10,
|
|
entry_price_limit=210.0,
|
|
stop_price=202.0,
|
|
target_price=230.0,
|
|
risk_dollars=80.0,
|
|
event_date=date,
|
|
timing_class="after_close",
|
|
engine_id="macro_bullish_qqq_etf_breadth",
|
|
entry_timing_policy="next_open",
|
|
)
|
|
runner._open_positions = [
|
|
OpenPosition(
|
|
position_id="macro-pos-1",
|
|
plan=macro_plan,
|
|
entry_date=date,
|
|
entry_price=210.0,
|
|
entry_fill_slippage_bps=0.0,
|
|
current_stop=202.0,
|
|
target_price=230.0,
|
|
peak_price=211.0,
|
|
shares_open=10,
|
|
shares_total=10,
|
|
)
|
|
]
|
|
|
|
core_state_after_fill = runner._adjust_portfolio_state_for_candidate(
|
|
date=date,
|
|
candidate=core_candidate,
|
|
portfolio_state=portfolio_state,
|
|
active_bucket_ids=active_bucket_ids,
|
|
)
|
|
macro_state_after_fill = runner._adjust_portfolio_state_for_candidate(
|
|
date=date,
|
|
candidate=macro_candidate,
|
|
portfolio_state=portfolio_state,
|
|
active_bucket_ids=active_bucket_ids,
|
|
)
|
|
|
|
assert core_state_after_fill.cash_available == pytest.approx(97_100.0)
|
|
assert macro_state_after_fill.cash_available == pytest.approx(2_900.0)
|
|
assert core_state_after_fill.sizing_equity == pytest.approx(94_990.0)
|
|
assert macro_state_after_fill.sizing_equity == pytest.approx(5_010.0)
|
|
|
|
def test_capital_bucket_does_not_reserve_when_bucket_inactive(self, tmp_path):
|
|
from apps.backtester.run import BacktestRunner
|
|
from libs.backtest.domain import (
|
|
BacktestConfig,
|
|
Candidate,
|
|
DailyPortfolioState,
|
|
ExecutionConfig,
|
|
ExperimentManifest,
|
|
ReportingConfig,
|
|
RiskConfig,
|
|
SignalConfig,
|
|
StrategyEngineConfig,
|
|
UniverseConfig,
|
|
)
|
|
from libs.backtest.snapshot_store import SnapshotStore
|
|
|
|
date = dt.date(2026, 1, 7)
|
|
store = SnapshotStore(candidates_by_exec_date={}, bars_by_symbol_date={})
|
|
manifest = ExperimentManifest(
|
|
experiment_name="macro_capital_bucket_inactive",
|
|
dataset_snapshot_id="test_snapshot",
|
|
base_config="configs/backtest/defaults.json",
|
|
overrides={},
|
|
strategy_engines=[
|
|
StrategyEngineConfig(engine_id="core"),
|
|
StrategyEngineConfig(
|
|
engine_id="macro_bullish_qqq_etf_breadth",
|
|
synthetic_only=True,
|
|
capital_bucket_id="macro_etf",
|
|
capital_bucket_allocation_pct=0.05,
|
|
),
|
|
],
|
|
)
|
|
config = BacktestConfig(
|
|
strategy_name="macro_capital_bucket_inactive",
|
|
dataset_snapshot_id="test_snapshot",
|
|
universe=UniverseConfig(min_price=5.0, min_avg_dollar_volume=100_000.0),
|
|
signal=SignalConfig(score_threshold=0.50, max_candidates_per_day=5),
|
|
risk=RiskConfig(
|
|
per_trade_risk_pct=0.01,
|
|
max_daily_new_risk_pct=0.05,
|
|
max_positions=10,
|
|
max_positions_per_sector=5,
|
|
),
|
|
execution=ExecutionConfig(
|
|
entry_fill_model="next_open",
|
|
exit_fill_model="daily_bar_approximation",
|
|
slippage_bps_base=0.0,
|
|
commission_per_share=0.0,
|
|
same_bar_priority="stop_first_conservative",
|
|
max_holding_days=10,
|
|
),
|
|
reporting=ReportingConfig(
|
|
write_trade_blotter=True,
|
|
write_equity_curve=True,
|
|
write_metrics_summary=True,
|
|
generate_plots=False,
|
|
),
|
|
strategy_engines=manifest.strategy_engines,
|
|
)
|
|
runner = BacktestRunner(manifest=manifest, config=config, store=store, initial_equity=100_000.0)
|
|
portfolio_state = DailyPortfolioState(
|
|
date=date,
|
|
equity=100_000.0,
|
|
sizing_equity=100_000.0,
|
|
cash_available=100_000.0,
|
|
gross_exposure=0.0,
|
|
net_exposure=0.0,
|
|
reserved_risk_budget=0.0,
|
|
unrealized_pnl=0.0,
|
|
realized_pnl=0.0,
|
|
open_positions=[],
|
|
daily_new_risk_used=0.0,
|
|
peak_equity=100_000.0,
|
|
current_drawdown_pct=0.0,
|
|
)
|
|
core_candidate = Candidate(
|
|
event_id="core::inactive",
|
|
symbol="AAPL",
|
|
score=0.7,
|
|
sector="Technology",
|
|
event_type="earnings_release",
|
|
event_timestamp=dt.datetime(2026, 1, 6, 21, 0, tzinfo=_UTC),
|
|
filing_time_bucket="post_market",
|
|
reaction_date=date,
|
|
execution_date=date,
|
|
entry_price_est=100.0,
|
|
avg_dollar_volume=1_000_000.0,
|
|
atr_14=2.0,
|
|
score_bucket="high",
|
|
engine_id="core",
|
|
)
|
|
|
|
active_bucket_ids = runner._active_capital_bucket_ids_for_candidates([core_candidate])
|
|
adjusted_state = runner._adjust_portfolio_state_for_candidate(
|
|
date=date,
|
|
candidate=core_candidate,
|
|
portfolio_state=portfolio_state,
|
|
active_bucket_ids=active_bucket_ids,
|
|
)
|
|
|
|
assert adjusted_state.cash_available == pytest.approx(100_000.0)
|
|
assert adjusted_state.sizing_equity == pytest.approx(100_000.0)
|
|
|
|
def test_parallel_sgov_marks_to_market_and_realizes_proportional_pnl(self, tmp_path):
|
|
from apps.backtester.run import BacktestRunner
|
|
from libs.backtest.domain import (
|
|
BacktestConfig,
|
|
ExecutionConfig,
|
|
ExperimentManifest,
|
|
ReportingConfig,
|
|
RiskConfig,
|
|
SignalConfig,
|
|
UniverseConfig,
|
|
)
|
|
from libs.backtest.snapshot_store import SnapshotStore
|
|
|
|
start = dt.date(2026, 1, 6)
|
|
mark_date = dt.date(2026, 1, 7)
|
|
store = SnapshotStore(
|
|
candidates_by_exec_date={},
|
|
bars_by_symbol_date={},
|
|
macro_by_date={
|
|
start: {"sgov_close": 100.0},
|
|
mark_date: {"sgov_close": 101.0},
|
|
},
|
|
)
|
|
manifest = ExperimentManifest(
|
|
experiment_name="parallel_sgov_mtm",
|
|
dataset_snapshot_id="test_snapshot",
|
|
base_config="configs/backtest/defaults.json",
|
|
overrides={},
|
|
)
|
|
config = BacktestConfig(
|
|
strategy_name="parallel_sgov_mtm",
|
|
dataset_snapshot_id="test_snapshot",
|
|
universe=UniverseConfig(min_price=5.0, min_avg_dollar_volume=100_000.0),
|
|
signal=SignalConfig(score_threshold=0.50, max_candidates_per_day=5),
|
|
risk=RiskConfig(
|
|
per_trade_risk_pct=0.01,
|
|
max_daily_new_risk_pct=0.05,
|
|
max_positions=10,
|
|
max_positions_per_sector=5,
|
|
),
|
|
execution=ExecutionConfig(
|
|
entry_fill_model="next_open",
|
|
exit_fill_model="daily_bar_approximation",
|
|
slippage_bps_base=0.0,
|
|
commission_per_share=0.0,
|
|
same_bar_priority="stop_first_conservative",
|
|
max_holding_days=10,
|
|
),
|
|
reporting=ReportingConfig(
|
|
write_trade_blotter=True,
|
|
write_equity_curve=True,
|
|
write_metrics_summary=True,
|
|
generate_plots=False,
|
|
),
|
|
)
|
|
|
|
runner = BacktestRunner(manifest=manifest, config=config, store=store, initial_equity=100_000.0)
|
|
runner._cash = 95_000.0
|
|
runner._parking_entry_date = start
|
|
runner._allocate_parallel_sgov(date=start, amount=5_000.0, macro={"sgov_close": 100.0})
|
|
|
|
assert runner._get_parking_value(mark_date) == pytest.approx(5_050.0)
|
|
|
|
runner._liquidate_parking_for_cash(mark_date, 1_000.0)
|
|
|
|
assert runner._cash == pytest.approx(96_000.0)
|
|
assert runner._parking_sgov_value == pytest.approx(4_050.0)
|
|
assert runner._parking_sgov_entry_value == pytest.approx(4_009.9009901)
|
|
assert runner._realized_pnl == pytest.approx(9.9009901)
|
|
|
|
runner._liquidate_parking(mark_date)
|
|
|
|
assert runner._cash == pytest.approx(100_050.0)
|
|
assert runner._realized_pnl == pytest.approx(50.0)
|
|
|
|
def test_macro_bullish_engine_respects_leadership_vs_spy_gate(self, tmp_path):
|
|
from apps.backtester.run import BacktestRunner
|
|
from libs.backtest.domain import (
|
|
BacktestConfig,
|
|
ExecutionConfig,
|
|
ExperimentManifest,
|
|
ReportingConfig,
|
|
RiskConfig,
|
|
SignalConfig,
|
|
StrategyEngineConfig,
|
|
UniverseConfig,
|
|
)
|
|
from libs.backtest.snapshot_store import SnapshotStore
|
|
|
|
signal_date = dt.date(2026, 1, 6)
|
|
next_date = dt.date(2026, 1, 7)
|
|
store = SnapshotStore(
|
|
candidates_by_exec_date={next_date: []},
|
|
bars_by_symbol_date={
|
|
"SPY": {
|
|
dt.date(2026, 1, 5): {"date": dt.date(2026, 1, 5), "open": 100.0, "high": 101.0, "low": 99.0, "close": 100.0, "volume": 100},
|
|
signal_date: {"date": signal_date, "open": 102.0, "high": 105.0, "low": 101.5, "close": 104.0, "volume": 140},
|
|
},
|
|
"QQQ": {
|
|
dt.date(2026, 1, 5): {"date": dt.date(2026, 1, 5), "open": 100.0, "high": 101.0, "low": 99.0, "close": 100.0, "volume": 100},
|
|
signal_date: {"date": signal_date, "open": 101.5, "high": 104.0, "low": 101.0, "close": 103.0, "volume": 150},
|
|
next_date: {"date": next_date, "open": 103.5, "high": 105.0, "low": 102.0, "close": 104.0, "volume": 140},
|
|
},
|
|
"XLK": {
|
|
dt.date(2026, 1, 5): {"date": dt.date(2026, 1, 5), "open": 200.0, "high": 201.0, "low": 199.0, "close": 200.0, "volume": 100},
|
|
signal_date: {"date": signal_date, "open": 202.0, "high": 208.0, "low": 201.0, "close": 206.0, "volume": 150},
|
|
},
|
|
},
|
|
macro_by_date={signal_date: {"VIXCLS": 22.0}},
|
|
)
|
|
manifest = ExperimentManifest(
|
|
experiment_name="macro_bullish_etf_breadth_leadership_block",
|
|
dataset_snapshot_id="test_snapshot",
|
|
base_config="configs/backtest/defaults.json",
|
|
overrides={},
|
|
strategy_engines=[
|
|
StrategyEngineConfig(
|
|
engine_id="macro_bullish_qqq_etf_breadth",
|
|
synthetic_only=True,
|
|
entry_timing_policy="next_open",
|
|
macro_long_symbol="QQQ",
|
|
macro_long_reaction_day_return_min=0.015,
|
|
macro_long_close_location_min=0.60,
|
|
macro_long_breadth_symbols=["QQQ", "XLK"],
|
|
macro_long_min_breadth_count=2,
|
|
macro_long_breadth_reaction_day_return_min=0.015,
|
|
macro_long_breadth_close_location_min=0.60,
|
|
macro_long_leadership_vs_spy_min=0.02,
|
|
)
|
|
],
|
|
)
|
|
config = BacktestConfig(
|
|
strategy_name="macro_bullish_etf_breadth_leadership_block",
|
|
dataset_snapshot_id="test_snapshot",
|
|
universe=UniverseConfig(min_price=5.0, min_avg_dollar_volume=100_000.0),
|
|
signal=SignalConfig(score_threshold=0.50, max_candidates_per_day=5),
|
|
risk=RiskConfig(
|
|
per_trade_risk_pct=0.01,
|
|
max_daily_new_risk_pct=0.05,
|
|
max_positions=10,
|
|
max_positions_per_sector=5,
|
|
),
|
|
execution=ExecutionConfig(
|
|
entry_fill_model="next_open",
|
|
exit_fill_model="daily_bar_approximation",
|
|
slippage_bps_base=0.0,
|
|
commission_per_share=0.0,
|
|
same_bar_priority="stop_first_conservative",
|
|
max_holding_days=10,
|
|
),
|
|
reporting=ReportingConfig(
|
|
write_trade_blotter=True,
|
|
write_equity_curve=True,
|
|
write_metrics_summary=True,
|
|
generate_plots=False,
|
|
),
|
|
strategy_engines=manifest.strategy_engines,
|
|
)
|
|
|
|
runner = BacktestRunner(manifest=manifest, config=config, store=store, initial_equity=100_000.0)
|
|
runner._simulation_dates = [signal_date, next_date]
|
|
runner._next_trading_day = {signal_date: next_date}
|
|
|
|
runner._schedule_macro_long_candidates(signal_date)
|
|
assert runner._scheduled_delayed_entries[next_date] == []
|