Spaces:
Running
Running
File size: 15,385 Bytes
9246aa7 87523c1 9246aa7 0010624 9246aa7 0010624 50dc123 0010624 87523c1 50dc123 9246aa7 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 | from __future__ import annotations
import os
import sys
import types
import unittest
from unittest.mock import patch
import pandas as pd
sys.path.insert(0, os.path.dirname(os.path.dirname(__file__)))
from models.hr_probability_engine import build_hr_probability_result
def _sample_statcast_df() -> pd.DataFrame:
batter_rows = [
{
"player_name": "Slugger Sam",
"launch_speed": 102 + i,
"launch_angle": 24 + (i % 4),
"estimated_woba_using_speedangle": 0.390,
"stand": "L",
"game_date": f"2026-03-{10 + i:02d}",
"game_pk": 100 + i,
"events": "home_run" if i % 4 == 0 else "single",
"description": "hit_into_play",
"pitch_name": "Slider",
"plate_x": 0.1,
"plate_z": 2.6,
}
for i in range(10)
]
pitcher_rows = [
{
"player_name": "Pitcher Pete",
"launch_speed": 89 + (i % 3),
"launch_angle": 14 + (i % 5),
"release_speed": 94 + (i % 2),
"release_spin_rate": 2250 + i * 5,
"release_extension": 6.1,
"pfx_x": 0.8,
"pfx_z": 1.2,
"estimated_woba_using_speedangle": 0.305,
"p_throws": "R",
"game_date": f"2026-03-{10 + i:02d}",
"game_pk": 200 + i,
"events": "field_out",
"description": "swinging_strike" if i % 3 == 0 else "called_strike",
"pitch_name": "Four-Seam Fastball",
"pitch_type": "FF",
"plate_x": 0.2,
"plate_z": 2.8,
"release_pos_x": -1.2,
"release_pos_y": 54.0,
"release_pos_z": 5.8,
"vx0": 5.0,
"vz0": -4.0,
"ax": -10.0,
"az": -20.0,
}
for i in range(12)
]
return pd.DataFrame(batter_rows + pitcher_rows)
class TestHrProbabilityEngine(unittest.TestCase):
def setUp(self) -> None:
self.statcast_df = _sample_statcast_df()
self.batter_df = self.statcast_df[
self.statcast_df["player_name"] == "Slugger Sam"
].reset_index(drop=True)
self.pitcher_df = self.statcast_df[
self.statcast_df["player_name"] == "Pitcher Pete"
].reset_index(drop=True)
def test_baseline_parity_when_context_is_absent(self) -> None:
with patch("models.hr_probability_engine.compute_pitcher_adjustment", return_value={"hr_adj": 0.0}), \
patch("models.hr_probability_engine.compute_environment_adjustment", return_value={"env_hr_boost": 0.0, "park_hr_boost": 0.0, "weather_hr_boost": 0.0}), \
patch("models.hr_probability_engine.build_trajectory_features", return_value={}), \
patch("models.hr_probability_engine.compute_trajectory_adjustment", return_value={"hr_adj": 0.0}), \
patch("models.hr_probability_engine.compute_upcoming_rolling_adjustment", return_value={"rolling_hr_adjustment": 0.0}):
result = build_hr_probability_result(
statcast_df=self.statcast_df,
batter_name="Slugger Sam",
mode="pregame",
)
self.assertIsNotNone(result["baseline_hr_prob"])
self.assertAlmostEqual(result["baseline_hr_prob"], result["adjusted_hr_prob"], places=6)
self.assertIn("live_pitch_telemetry", result["skipped_layers"])
def test_pitcher_context_moves_probability(self) -> None:
with patch("models.hr_probability_engine.compute_pitcher_adjustment", return_value={"hr_adj": 0.02}), \
patch("models.hr_probability_engine.compute_environment_adjustment", return_value={"env_hr_boost": 0.0, "park_hr_boost": 0.0, "weather_hr_boost": 0.0}), \
patch("models.hr_probability_engine.build_trajectory_features", return_value={}), \
patch("models.hr_probability_engine.compute_trajectory_adjustment", return_value={"hr_adj": 0.0}), \
patch("models.hr_probability_engine.compute_upcoming_rolling_adjustment", return_value={"rolling_hr_adjustment": 0.0}):
result = build_hr_probability_result(
statcast_df=self.statcast_df,
batter_name="Slugger Sam",
pitcher_name="Pitcher Pete",
mode="pregame",
)
self.assertGreater(result["adjusted_hr_prob"], result["baseline_hr_prob"])
self.assertAlmostEqual(result["pregame_pitcher_context_adj"], 0.02, places=6)
self.assertIn("pitcher", result["applied_layers"])
def test_low_sample_pitcher_adjustment_is_shrunk(self) -> None:
with patch("models.hr_probability_engine.build_pitcher_feature_row", return_value={"sample_size": 10, "p_throws": "R"}), \
patch("models.hr_probability_engine.compute_pitcher_adjustment", return_value={"hr_adj": 0.02}), \
patch("models.hr_probability_engine.compute_environment_adjustment", return_value={"env_hr_boost": 0.0, "park_hr_boost": 0.0, "weather_hr_boost": 0.0}), \
patch("models.hr_probability_engine.build_trajectory_features", return_value={}), \
patch("models.hr_probability_engine.compute_trajectory_adjustment", return_value={"hr_adj": 0.0}), \
patch("models.hr_probability_engine.compute_upcoming_rolling_adjustment", return_value={"rolling_hr_adjustment": 0.0, "pitcher_rolling_confidence": 0.0}):
result = build_hr_probability_result(
batter_statcast_df=self.batter_df,
batter_name="Slugger Sam",
pitcher_statcast_df=self.pitcher_df,
pitcher_name="Pitcher Pete",
mode="pregame",
)
self.assertLess(float(result["pitcher_hr_adjustment"]), 0.02)
self.assertLess(float(result["pitcher_reliability"]), 0.2)
def test_matchup_layers_are_applied_in_pregame_mode(self) -> None:
fake_batter_zone = types.SimpleNamespace(
build_batter_zone_feature_row=lambda *args, **kwargs: {}
)
fake_pitcher_zone = types.SimpleNamespace(
build_pitcher_zone_feature_row=lambda *args, **kwargs: {}
)
fake_zone_matchup = types.SimpleNamespace(
compute_zone_matchup_adjustment=lambda *args, **kwargs: {"hr_zone_boost": 0.20}
)
fake_family_store = types.SimpleNamespace(
build_batter_family_zone_feature_row=lambda *args, **kwargs: {},
build_pitcher_family_zone_feature_row=lambda *args, **kwargs: {},
)
fake_matchup_model = types.SimpleNamespace(
compute_family_zone_matchup_adjustment=lambda *args, **kwargs: {"family_zone_hr_boost": 0.10}
)
fake_batter_arsenal = types.SimpleNamespace(
build_batter_arsenal_feature_row=lambda *args, **kwargs: {}
)
fake_pitcher_arsenal = types.SimpleNamespace(
build_pitcher_arsenal_feature_row=lambda *args, **kwargs: {}
)
fake_arsenal_matchup = types.SimpleNamespace(
compute_arsenal_matchup_adjustment=lambda *args, **kwargs: {"arsenal_hr_boost": 0.10}
)
with patch("models.hr_probability_engine.compute_pitcher_adjustment", return_value={"hr_adj": 0.0}), \
patch("models.hr_probability_engine.compute_environment_adjustment", return_value={"env_hr_boost": 0.0, "park_hr_boost": 0.0, "weather_hr_boost": 0.0}), \
patch("models.hr_probability_engine.build_trajectory_features", return_value={}), \
patch("models.hr_probability_engine.compute_trajectory_adjustment", return_value={"hr_adj": 0.0}), \
patch("models.hr_probability_engine.compute_upcoming_rolling_adjustment", return_value={"rolling_hr_adjustment": 0.0}), \
patch.dict(
sys.modules,
{
"models.batter_zone_model": fake_batter_zone,
"models.pitcher_zone_model": fake_pitcher_zone,
"models.zone_matchup_model": fake_zone_matchup,
"models.family_zone_profile_store": fake_family_store,
"models.matchup_model": fake_matchup_model,
"models.batter_arsenal_model": fake_batter_arsenal,
"models.pitcher_arsenal_model": fake_pitcher_arsenal,
"models.arsenal_matchup_model": fake_arsenal_matchup,
},
):
result = build_hr_probability_result(
statcast_df=self.statcast_df,
batter_name="Slugger Sam",
pitcher_name="Pitcher Pete",
mode="pregame",
)
self.assertNotEqual(result["zone_hr_adjustment"], 0.0)
self.assertNotEqual(result["family_zone_hr_adjustment"], 0.0)
self.assertNotEqual(result["arsenal_hr_adjustment"], 0.0)
self.assertIn("zone", result["applied_layers"])
self.assertIn("family_zone", result["applied_layers"])
self.assertIn("arsenal", result["applied_layers"])
def test_environment_layers_are_recorded(self) -> None:
with patch("models.hr_probability_engine.compute_pitcher_adjustment", return_value={"hr_adj": 0.0}), \
patch("models.hr_probability_engine.compute_environment_adjustment", return_value={"env_hr_boost": 0.03, "park_hr_boost": 0.01, "weather_hr_boost": 0.02}), \
patch("models.hr_probability_engine.build_trajectory_features", return_value={}), \
patch("models.hr_probability_engine.compute_trajectory_adjustment", return_value={"hr_adj": 0.0}), \
patch("models.hr_probability_engine.compute_upcoming_rolling_adjustment", return_value={"rolling_hr_adjustment": 0.0}):
result = build_hr_probability_result(
statcast_df=self.statcast_df,
batter_name="Slugger Sam",
mode="pregame",
game_row={"venue": "Dodger Stadium"},
weather_row={"temperature_f": 88, "wind_speed_mph": 12, "wind_direction_deg": 180},
)
self.assertAlmostEqual(result["env_hr_adjustment"], 0.03, places=6)
self.assertAlmostEqual(result["park_hr_adjustment"], 0.01, places=6)
self.assertAlmostEqual(result["weather_hr_adjustment"], 0.02, places=6)
self.assertIn("environment", result["applied_layers"])
def test_engine_returns_raw_calibrated_confidence_and_opportunity_fields(self) -> None:
with patch("models.hr_probability_engine.compute_pitcher_adjustment", return_value={"hr_adj": 0.01}), \
patch("models.hr_probability_engine.compute_environment_adjustment", return_value={"env_hr_boost": 0.01, "park_hr_boost": 0.01, "weather_hr_boost": 0.0}), \
patch("models.hr_probability_engine.build_trajectory_features", return_value={}), \
patch("models.hr_probability_engine.compute_trajectory_adjustment", return_value={"hr_adj": 0.0}), \
patch("models.hr_probability_engine.compute_upcoming_rolling_adjustment", return_value={"rolling_hr_adjustment": 0.01, "pitcher_rolling_confidence": 0.9}):
result = build_hr_probability_result(
batter_statcast_df=self.batter_df,
batter_name="Slugger Sam",
pitcher_statcast_df=self.pitcher_df,
pitcher_name="Pitcher Pete",
game_row={
"venue": "Yankee Stadium",
"lineup_slot": 3,
"lineup_slot_source": "projected",
"team_total": 4.8,
"team_total_source": "projected",
},
mode="pregame",
)
self.assertIsNotNone(result["raw_hr_prob"])
self.assertIsNotNone(result["calibrated_hr_prob"])
self.assertAlmostEqual(result["pregame_hr_prob"], result["calibrated_hr_prob"], places=6)
self.assertTrue(1.0 <= float(result["confidence_score"]) <= 100.0)
self.assertEqual(result["lineup_slot_used"], 3)
self.assertEqual(result["team_total_used"], 4.8)
self.assertIn("opportunity", result["applied_layers"])
self.assertNotEqual(result["raw_hr_prob"], result["calibrated_hr_prob"])
self.assertEqual(result["projected_home_pitcher"], "")
self.assertEqual(result["hr_model_tier"], "partial_telemetry")
self.assertTrue(bool(result["modeled_row_available"]))
def test_separate_batter_and_pitcher_dataframes_are_supported(self) -> None:
with patch("models.hr_probability_engine.compute_pitcher_adjustment", return_value={"hr_adj": 0.01}), \
patch("models.hr_probability_engine.compute_environment_adjustment", return_value={"env_hr_boost": 0.0, "park_hr_boost": 0.0, "weather_hr_boost": 0.0}), \
patch("models.hr_probability_engine.build_trajectory_features", return_value={}), \
patch("models.hr_probability_engine.compute_trajectory_adjustment", return_value={"hr_adj": 0.0}), \
patch("models.hr_probability_engine.compute_upcoming_rolling_adjustment", return_value={"rolling_hr_adjustment": 0.0}):
result = build_hr_probability_result(
batter_statcast_df=self.batter_df,
batter_name="Slugger Sam",
pitcher_statcast_df=self.pitcher_df,
pitcher_name="Pitcher Pete",
mode="pregame",
)
self.assertIsNotNone(result["baseline_hr_prob"])
self.assertGreater(result["adjusted_hr_prob"], result["baseline_hr_prob"])
self.assertEqual(result["pitcher_name"], "Pitcher Pete")
def test_engine_tracks_projected_starter_fields_when_supplied(self) -> None:
with patch("models.hr_probability_engine.compute_pitcher_adjustment", return_value={"hr_adj": 0.01}), \
patch("models.hr_probability_engine.compute_environment_adjustment", return_value={"env_hr_boost": 0.0, "park_hr_boost": 0.0, "weather_hr_boost": 0.0}), \
patch("models.hr_probability_engine.build_trajectory_features", return_value={}), \
patch("models.hr_probability_engine.compute_trajectory_adjustment", return_value={"hr_adj": 0.0}), \
patch("models.hr_probability_engine.compute_upcoming_rolling_adjustment", return_value={"rolling_hr_adjustment": 0.0}):
result = build_hr_probability_result(
batter_statcast_df=self.batter_df,
batter_name="Slugger Sam",
pitcher_statcast_df=self.pitcher_df,
pitcher_name="Pitcher Pete",
game_row={
"projected_home_pitcher": "Pitcher Pete",
"projected_away_pitcher": "Other Arm",
"projected_starter_available": True,
"projected_starter_match_status": "matched_projected_home",
},
mode="pregame",
)
self.assertEqual(result["projected_home_pitcher"], "Pitcher Pete")
self.assertEqual(result["projected_away_pitcher"], "Other Arm")
self.assertTrue(bool(result["projected_starter_available"]))
self.assertEqual(result["projected_starter_match_status"], "matched_projected_home")
self.assertTrue(bool(result["modeled_row_available"]))
if __name__ == "__main__":
unittest.main()
|