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.
383 lines
13 KiB
Python
383 lines
13 KiB
Python
"""Unit tests for libs.backtest.scoring."""
|
|
from __future__ import annotations
|
|
|
|
import pytest
|
|
|
|
from libs.backtest.scoring import (
|
|
_close_strength_score,
|
|
_direction_clarity_score,
|
|
_earnings_surprise_score,
|
|
_event_quality_score,
|
|
_gap_score,
|
|
_parse_confidence_score,
|
|
_reaction_score,
|
|
_risk_penalty_score,
|
|
_text_sentiment_score,
|
|
_volume_score,
|
|
compute_entry_score,
|
|
)
|
|
|
|
|
|
class TestReactionScore:
|
|
"""Reaction day return scoring — simple directional mapping."""
|
|
|
|
def test_strong_positive(self):
|
|
"""Strong positive return (>3%)."""
|
|
assert _reaction_score({"reaction_day_return": 0.05}) == 0.7
|
|
assert _reaction_score({"reaction_day_return": 0.15}) == 0.7
|
|
|
|
def test_positive(self):
|
|
"""Positive return (0-3%)."""
|
|
assert _reaction_score({"reaction_day_return": 0.01}) == 0.6
|
|
assert _reaction_score({"reaction_day_return": 0.025}) == 0.6
|
|
assert _reaction_score({"reaction_day_return": 0.002}) == 0.6
|
|
|
|
def test_flat_to_mild_negative(self):
|
|
"""Flat to mild negative (0% to -2%)."""
|
|
assert _reaction_score({"reaction_day_return": -0.01}) == 0.45
|
|
|
|
def test_moderate_negative(self):
|
|
"""Moderate negative (-5% to -2%)."""
|
|
assert _reaction_score({"reaction_day_return": -0.03}) == 0.3
|
|
|
|
def test_strongly_bearish(self):
|
|
"""Strongly bearish (<-5%)."""
|
|
assert _reaction_score({"reaction_day_return": -0.08}) == 0.2
|
|
|
|
def test_missing_returns_default(self):
|
|
assert _reaction_score({}) == 0.5
|
|
|
|
|
|
class TestCloseStrengthScore:
|
|
"""Close location scoring."""
|
|
|
|
def test_near_high(self):
|
|
score = _close_strength_score({"close_location": 0.9})
|
|
assert score == pytest.approx(0.91, abs=0.01)
|
|
|
|
def test_near_low(self):
|
|
score = _close_strength_score({"close_location": 0.1})
|
|
assert score == pytest.approx(0.19, abs=0.01)
|
|
|
|
def test_midpoint(self):
|
|
score = _close_strength_score({"close_location": 0.5})
|
|
assert score == pytest.approx(0.55, abs=0.01)
|
|
|
|
def test_missing_returns_default(self):
|
|
assert _close_strength_score({}) == 0.5
|
|
|
|
def test_clamped_above_1(self):
|
|
"""Values > 1.0 are clamped."""
|
|
score = _close_strength_score({"close_location": 1.5})
|
|
assert score == pytest.approx(1.0, abs=0.01)
|
|
|
|
|
|
class TestVolumeScore:
|
|
"""Volume ratio scoring."""
|
|
|
|
def test_healthy_conviction(self):
|
|
assert _volume_score({"volume_ratio_20d": 1.5}) == 0.8
|
|
|
|
def test_normal(self):
|
|
assert _volume_score({"volume_ratio_20d": 1.1}) == 0.6
|
|
|
|
def test_high_possible_exhaustion(self):
|
|
assert _volume_score({"volume_ratio_20d": 2.5}) == 0.55
|
|
|
|
def test_extreme_exhaustion(self):
|
|
assert _volume_score({"volume_ratio_20d": 4.0}) == 0.4
|
|
|
|
def test_below_average(self):
|
|
assert _volume_score({"volume_ratio_20d": 0.8}) == 0.3
|
|
|
|
def test_missing_returns_default(self):
|
|
assert _volume_score({}) == 0.5
|
|
|
|
|
|
class TestGapScore:
|
|
"""Gap size scoring."""
|
|
|
|
def test_orderly_positive(self):
|
|
assert _gap_score({"gap_size": 0.01}) == 0.8
|
|
|
|
def test_neutral(self):
|
|
assert _gap_score({"gap_size": 0.002}) == 0.6
|
|
|
|
def test_extended(self):
|
|
assert _gap_score({"gap_size": 0.03}) == 0.5
|
|
|
|
def test_exhaustion_gap(self):
|
|
assert _gap_score({"gap_size": 0.07}) == 0.3
|
|
|
|
def test_mild_negative(self):
|
|
assert _gap_score({"gap_size": -0.01}) == 0.4
|
|
|
|
def test_bearish_gap(self):
|
|
assert _gap_score({"gap_size": -0.03}) == 0.2
|
|
|
|
def test_missing_returns_default(self):
|
|
assert _gap_score({}) == 0.5
|
|
|
|
|
|
class TestComputeEntryScore:
|
|
"""Composite score tests."""
|
|
|
|
def test_all_missing_returns_neutral(self):
|
|
"""All features missing → all defaults at 0.5 → composite 0.5."""
|
|
score = compute_entry_score({})
|
|
assert score == pytest.approx(0.5, abs=0.02)
|
|
|
|
def test_ideal_setup_scores_high(self):
|
|
"""Positive return + healthy volume (no event features → event=0.5).
|
|
|
|
event=0.5*0.65 + reaction=0.6*0.20 + volume=0.8*0.15 ≈ 0.565
|
|
"""
|
|
row = {
|
|
"reaction_day_return": 0.01, # positive → 0.6
|
|
"volume_ratio_20d": 1.5, # conviction → 0.8
|
|
}
|
|
score = compute_entry_score(row)
|
|
assert score > 0.55
|
|
|
|
def test_bearish_setup_scores_low(self):
|
|
"""Negative return + below avg volume (no event features → event=0.5).
|
|
|
|
event=0.5*0.65 + reaction=0.3*0.20 + volume=0.3*0.15 ≈ 0.43
|
|
"""
|
|
row = {
|
|
"reaction_day_return": -0.04, # moderate negative → 0.3
|
|
"volume_ratio_20d": 0.8, # no conviction → 0.3
|
|
}
|
|
score = compute_entry_score(row)
|
|
assert score < 0.45
|
|
|
|
def test_extreme_positive_penalized(self):
|
|
"""Very large positive reaction should be penalized."""
|
|
extreme = compute_entry_score({
|
|
"reaction_day_return": 0.15, # extreme → 0.2
|
|
"close_location": 0.7,
|
|
"volume_ratio_20d": 3.5, # extreme volume → 0.4
|
|
"gap_size": 0.08, # exhaustion gap → 0.3
|
|
})
|
|
moderate = compute_entry_score({
|
|
"reaction_day_return": 0.015, # sweet spot → 0.9
|
|
"close_location": 0.7,
|
|
"volume_ratio_20d": 1.5, # healthy → 0.8
|
|
"gap_size": 0.01, # orderly → 0.8
|
|
})
|
|
assert moderate > extreme
|
|
|
|
def test_score_bounded_0_to_1(self):
|
|
"""Score is always in [0, 1]."""
|
|
extremes = [
|
|
{"reaction_day_return": -0.5, "close_location": 0.0,
|
|
"volume_ratio_20d": 0.1, "gap_size": -0.1},
|
|
{"reaction_day_return": 0.5, "close_location": 1.0,
|
|
"volume_ratio_20d": 10.0, "gap_size": 0.2},
|
|
]
|
|
for row in extremes:
|
|
score = compute_entry_score(row)
|
|
assert 0.0 <= score <= 1.0
|
|
|
|
def test_aapl_like_scores_highest(self):
|
|
"""AAPL-like setup (moderate +, close near high) scores well."""
|
|
aapl = compute_entry_score({
|
|
"reaction_day_return": 0.007,
|
|
"close_location": 0.74,
|
|
"volume_ratio_20d": 1.45,
|
|
"gap_size": 0.006,
|
|
})
|
|
# DDOG-like: extreme positive, high volume
|
|
ddog = compute_entry_score({
|
|
"reaction_day_return": 0.137,
|
|
"close_location": 0.63,
|
|
"volume_ratio_20d": 2.83,
|
|
"gap_size": 0.089,
|
|
})
|
|
assert aapl > ddog, f"AAPL-like {aapl:.3f} should beat DDOG-like {ddog:.3f}"
|
|
|
|
def test_real_data_scores(self):
|
|
"""Verify scores for real data samples match expectations.
|
|
|
|
With no event features, both get event=0.5.
|
|
TSLA: 0.5*0.65 + 0.3*0.20 + 0.6*0.15 = 0.475
|
|
AAPL: 0.5*0.65 + 0.6*0.20 + 0.8*0.15 = 0.565
|
|
"""
|
|
# TSLA 1/2: bearish (return -2.6%)
|
|
tsla = compute_entry_score({
|
|
"reaction_day_return": -0.026,
|
|
"volume_ratio_20d": 1.13,
|
|
})
|
|
# AAPL 1/29: bullish (moderate +)
|
|
aapl = compute_entry_score({
|
|
"reaction_day_return": 0.007,
|
|
"volume_ratio_20d": 1.45,
|
|
})
|
|
assert aapl > 0.55, f"AAPL should be above 0.55, got {aapl:.3f}"
|
|
assert tsla < 0.49, f"TSLA should be below 0.49, got {tsla:.3f}"
|
|
assert aapl > tsla
|
|
|
|
def test_event_features_boost_score(self):
|
|
"""Strong event features should boost overall score."""
|
|
market_only = compute_entry_score({
|
|
"reaction_day_return": 0.01,
|
|
"close_location": 0.6,
|
|
"volume_ratio_20d": 1.5,
|
|
"gap_size": 0.01,
|
|
})
|
|
with_events = compute_entry_score({
|
|
"reaction_day_return": 0.01,
|
|
"close_location": 0.6,
|
|
"volume_ratio_20d": 1.5,
|
|
"gap_size": 0.01,
|
|
"signal_strength_score": 0.8,
|
|
"guidance_direction_score": 1.0,
|
|
"document_quality_score": 0.7,
|
|
"oneoff_penalty": 0.0,
|
|
})
|
|
assert with_events > market_only
|
|
|
|
def test_removed_features_dont_affect_composite(self):
|
|
"""Changing non-alpha features should not change composite score.
|
|
|
|
eps_growth_qoq, oneoff_penalty, parse_confidence_overall,
|
|
event_direction, and lm_net_sentiment are excluded from composite.
|
|
"""
|
|
base_row = {
|
|
"reaction_day_return": 0.01,
|
|
"close_location": 0.6,
|
|
"volume_ratio_20d": 1.5,
|
|
"gap_size": 0.01,
|
|
}
|
|
base_score = compute_entry_score(base_row)
|
|
|
|
# Varying removed features should not change score
|
|
for extra in [
|
|
{"eps_growth_qoq": 0.50},
|
|
{"oneoff_penalty": 1.0},
|
|
{"parse_confidence_overall": 0.9},
|
|
{"event_direction": "bullish"},
|
|
{"lm_net_sentiment": 0.01},
|
|
]:
|
|
row = {**base_row, **extra}
|
|
assert compute_entry_score(row) == pytest.approx(base_score), (
|
|
f"Feature {list(extra.keys())[0]} should not affect composite"
|
|
)
|
|
|
|
|
|
class TestEarningsSurpriseScore:
|
|
"""Earnings surprise (SUE) scoring."""
|
|
|
|
def test_strong_beat(self):
|
|
assert _earnings_surprise_score({"event_type": "earnings_release", "eps_growth_qoq": 0.30}) == 0.9
|
|
|
|
def test_moderate_beat(self):
|
|
assert _earnings_surprise_score({"event_type": "earnings_release", "eps_growth_qoq": 0.10}) == 0.75
|
|
|
|
def test_inline(self):
|
|
assert _earnings_surprise_score({"event_type": "earnings_release", "eps_growth_qoq": 0.0}) == 0.5
|
|
|
|
def test_moderate_miss(self):
|
|
assert _earnings_surprise_score({"event_type": "earnings_release", "eps_growth_qoq": -0.10}) == 0.25
|
|
|
|
def test_severe_miss(self):
|
|
assert _earnings_surprise_score({"event_type": "earnings_release", "eps_growth_qoq": -0.30}) == 0.1
|
|
|
|
def test_non_earnings_returns_neutral(self):
|
|
assert _earnings_surprise_score({"event_type": "guidance_update", "eps_growth_qoq": 0.30}) == 0.5
|
|
|
|
def test_missing_returns_neutral(self):
|
|
assert _earnings_surprise_score({"event_type": "earnings_release"}) == 0.5
|
|
|
|
def test_no_event_type_returns_neutral(self):
|
|
assert _earnings_surprise_score({}) == 0.5
|
|
|
|
|
|
class TestEventQualityScore:
|
|
"""Event quality scoring."""
|
|
|
|
def test_strong_signals(self):
|
|
score = _event_quality_score({
|
|
"signal_strength_score": 0.8,
|
|
"guidance_direction_score": 1.0,
|
|
"document_quality_score": 0.7,
|
|
})
|
|
assert score > 0.7
|
|
|
|
def test_weak_signals(self):
|
|
score = _event_quality_score({
|
|
"signal_strength_score": 0.0,
|
|
"guidance_direction_score": 0.0,
|
|
"document_quality_score": 0.3,
|
|
})
|
|
assert score < 0.15
|
|
|
|
def test_missing_returns_neutral(self):
|
|
assert _event_quality_score({}) == 0.5
|
|
|
|
def test_partial_features(self):
|
|
"""Works with only some event features present."""
|
|
score = _event_quality_score({"signal_strength_score": 0.8})
|
|
assert score == pytest.approx(0.8, abs=0.01)
|
|
|
|
|
|
class TestRiskPenaltyScore:
|
|
"""Risk penalty scoring."""
|
|
|
|
def test_no_risk(self):
|
|
assert _risk_penalty_score({"oneoff_penalty": 0.0}) == pytest.approx(0.9)
|
|
|
|
def test_max_risk(self):
|
|
assert _risk_penalty_score({"oneoff_penalty": 1.0}) == pytest.approx(0.2)
|
|
|
|
def test_moderate_risk(self):
|
|
score = _risk_penalty_score({"oneoff_penalty": 0.5})
|
|
assert 0.4 < score < 0.7
|
|
|
|
def test_missing_returns_neutral(self):
|
|
assert _risk_penalty_score({}) == 0.5
|
|
|
|
|
|
class TestParseConfidenceScore:
|
|
"""Parse confidence scoring."""
|
|
|
|
def test_high_confidence(self):
|
|
assert _parse_confidence_score({"parse_confidence_overall": 0.9}) == 0.9
|
|
|
|
def test_moderate_confidence(self):
|
|
assert _parse_confidence_score({"parse_confidence_overall": 0.7}) == 0.7
|
|
|
|
def test_borderline_confidence(self):
|
|
assert _parse_confidence_score({"parse_confidence_overall": 0.55}) == 0.5
|
|
|
|
def test_low_confidence(self):
|
|
assert _parse_confidence_score({"parse_confidence_overall": 0.45}) == 0.3
|
|
|
|
def test_very_low_confidence(self):
|
|
assert _parse_confidence_score({"parse_confidence_overall": 0.3}) == 0.2
|
|
|
|
def test_missing_returns_neutral(self):
|
|
assert _parse_confidence_score({}) == 0.5
|
|
|
|
|
|
class TestDirectionClarityScore:
|
|
"""Direction clarity scoring."""
|
|
|
|
def test_bullish(self):
|
|
assert _direction_clarity_score({"event_direction": "bullish"}) == 0.9
|
|
|
|
def test_mixed(self):
|
|
assert _direction_clarity_score({"event_direction": "mixed"}) == 0.4
|
|
|
|
def test_neutral(self):
|
|
assert _direction_clarity_score({"event_direction": "neutral"}) == 0.3
|
|
|
|
def test_bearish(self):
|
|
assert _direction_clarity_score({"event_direction": "bearish"}) == 0.1
|
|
|
|
def test_unknown(self):
|
|
assert _direction_clarity_score({"event_direction": "unknown"}) == 0.2
|
|
|
|
def test_missing_returns_neutral(self):
|
|
assert _direction_clarity_score({}) == 0.5
|