Spaces:
Sleeping
Sleeping
File size: 17,494 Bytes
c9d1b27 | 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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 | import sys
import os
sys.path.insert(0, os.path.abspath(os.path.join(os.path.dirname(__file__), '..')))
from env.mahoraga_env import MahoragaEnv
from env.mechanics import (
new_resistances, apply_resistance_change, compute_enemy_damage,
compute_judgment_damage, check_correct_adaptation
)
from env.enemy import CurriculumEnemy
from env.rewards import compute_rewards
from utils.constants import SUBTYPES, ATTACK_TYPES, MAX_HP, ENEMY_HP
PASS = 0
FAIL = 0
def check(name, condition):
global PASS, FAIL
if condition:
PASS += 1
print(f" [PASS] {name}")
else:
FAIL += 1
print(f" [FAIL] {name}")
# ========== PHASE 1 CORE TESTS ==========
def test_resistance_update():
print("\n--- Test: Resistance Update ---")
res = new_resistances()
res = apply_resistance_change(res, "PHYSICAL")
check("PHYSICAL +40", res["PHYSICAL"] == 40)
check("CE -20 clamped to 0", res["CE"] == 0)
check("TECHNIQUE -20 clamped to 0", res["TECHNIQUE"] == 0)
res = apply_resistance_change(res, "PHYSICAL")
check("PHYSICAL +40 again = 80", res["PHYSICAL"] == 80)
check("CE stays 0", res["CE"] == 0)
check("TECHNIQUE stays 0", res["TECHNIQUE"] == 0)
def test_clamp_logic():
print("\n--- Test: Clamp Logic ---")
res = {"PHYSICAL": 70, "CE": 10, "TECHNIQUE": 10}
res = apply_resistance_change(res, "PHYSICAL")
check("PHYSICAL clamped to 80 (70+40)", res["PHYSICAL"] == 80)
check("CE clamped to 0 (10-20)", res["CE"] == 0)
check("TECHNIQUE clamped to 0 (10-20)", res["TECHNIQUE"] == 0)
def test_damage_formula():
print("\n--- Test: Damage Formula ---")
res = new_resistances()
dmg = compute_enemy_damage("PHYSICAL", res)
check("PHYSICAL full damage = 120", dmg == 120)
dmg = compute_enemy_damage("CE", res)
check("CE full damage = 150", dmg == 150)
dmg = compute_enemy_damage("TECHNIQUE", res)
check("TECHNIQUE full damage = 220", dmg == 220)
res["PHYSICAL"] = 50
dmg = compute_enemy_damage("PHYSICAL", res)
check("PHYSICAL with 50 res = 60", dmg == 60)
def test_judgment_damage():
print("\n--- Test: Judgment Damage (Adaptation-Match) ---")
# No adaptation — base damage
dmg = compute_judgment_damage(None, "PHYSICAL")
check("Base judgment (no adapt) = 100", dmg == 100)
# Matching adaptation — burst
dmg = compute_judgment_damage("PHYSICAL", "PHYSICAL")
check("Burst judgment (matching adapt) = 350", dmg == 350)
# Non-matching adaptation — no burst
dmg = compute_judgment_damage("CE", "PHYSICAL")
check("No burst when non-matching adapt = 100", dmg == 100)
def test_episode_termination():
print("\n--- Test: Episode Termination ---")
env = MahoragaEnv()
env.reset()
# Ensure agent survives all 25 turns by giving it massive HP
env.agent_hp = 99999
# Also prevent accidental enemy death by giving enemy massive HP
env.enemy_hp = 99999
done = False
info = {}
for i in range(25):
if not done:
_, _, done, info = env.step(0) # Only adapt, never attack
check("Episode ends at turn 25", done is True)
check("Reason is turn limit", info.get("reason") == "Turn limit reached")
def test_actions_behave_correctly():
print("\n--- Test: Action Behavior ---")
env = MahoragaEnv()
env.reset()
state, _, _, _ = env.step(0)
check("Adapt PHYSICAL sets res to 40", state["resistances"]["physical"] == 40)
# Judgment Strike resets resistances
env.reset()
env.step(0)
env.step(0)
state, _, _, _ = env.step(3)
check("Judgment resets resistances", state["resistances"]["physical"] == 0)
check("Judgment damages enemy", state["enemy_hp"] < ENEMY_HP)
# Regeneration heals but does NOT reset resistances
env.reset()
env.step(0)
hp_before = env.agent_hp
res_before = env.resistances["PHYSICAL"]
state, _, _, _ = env.step(4)
check("Regeneration heals agent", state["agent_hp"] > hp_before - 120)
check("Regeneration does NOT reset resistances", state["resistances"]["physical"] == res_before)
def test_adaptation_stack():
print("\n--- Test: Adaptation Stack ---")
env = MahoragaEnv()
env.reset()
# Phase 1 enemy always attacks PHYSICAL
env.step(0)
check("Stack +1 on correct adapt", env.adaptation_stack == 1)
env.step(0)
check("Stack +2 on second correct adapt", env.adaptation_stack == 2)
env.reset()
env.step(1) # CE adapt vs PHYSICAL enemy
check("Stack stays 0 on wrong adapt", env.adaptation_stack == 0)
env.reset()
env.step(0)
env.step(0)
env.step(3)
check("Stack reset to 0 after Judgment", env.adaptation_stack == 0)
def test_invalid_action():
print("\n--- Test: Invalid Action ---")
env = MahoragaEnv()
env.reset()
try:
env.step(5)
check("Rejects action 5", False)
except ValueError:
check("Rejects action 5", True)
try:
env.step(-1)
check("Rejects action -1", False)
except ValueError:
check("Rejects action -1", True)
def test_heal_cooldown():
print("\n--- Test: Heal Cooldown ---")
env = MahoragaEnv()
env.reset()
state, _, _, info = env.step(4)
check("First heal works", info["heal_on_cooldown"] is False)
check("Cooldown set to 3", env.heal_cooldown_counter == 3)
_, _, _, info2 = env.step(4)
check("Heal blocked on turn 2 (cooldown)", info2["heal_on_cooldown"] is True)
_, _, _, info3 = env.step(4)
check("Heal blocked on turn 3 (cooldown)", info3["heal_on_cooldown"] is True)
_, _, _, info4 = env.step(4)
check("Heal available on turn 4 (cooldown expired)", info4["heal_on_cooldown"] is False)
def test_heal_preserves_resistances():
print("\n--- Test: Heal Preserves Resistances ---")
env = MahoragaEnv()
env.reset()
env.step(0)
env.step(0)
res_before = env.resistances.copy()
env.step(4)
check("PHYSICAL res preserved after heal", env.resistances["PHYSICAL"] == res_before["PHYSICAL"])
check("CE res preserved after heal", env.resistances["CE"] == res_before["CE"])
check("TECHNIQUE res preserved after heal", env.resistances["TECHNIQUE"] == res_before["TECHNIQUE"])
def test_judgment_burst_adaptation_match():
print("\n--- Test: Judgment Burst Via Adaptation Match ---")
env = MahoragaEnv()
env.reset()
# Adapt PHYSICAL (correct for Phase 1 enemy) then Judgment
env.step(0) # Adapt PHYSICAL, correct, stack=1, last_adapted=PHYSICAL
env.step(0) # Adapt PHYSICAL, correct, stack=2, last_adapted=PHYSICAL
enemy_hp_before = env.enemy_hp
env.step(3) # Judgment — last_adapted=PHYSICAL, enemy=PHYSICAL → BURST
damage_dealt = enemy_hp_before - env.enemy_hp
# burst 350 + stack 2*50 = 450
check("Burst when adapted to matching category (dmg=450)", damage_dealt == 450)
# Wrong adaptation — should NOT burst
env.reset()
env.step(1) # Adapt CE, but enemy is PHYSICAL
env.step(1) # Adapt CE again
enemy_hp_before = env.enemy_hp
env.step(3) # Judgment — last_adapted=CE, enemy=PHYSICAL → NO BURST
damage_dealt = enemy_hp_before - env.enemy_hp
# base 100 + stack 0*50 = 100
check("No burst when adapted to wrong category (dmg=100)", damage_dealt == 100)
def test_info_dict_fields():
print("\n--- Test: Info Dict Fields ---")
env = MahoragaEnv()
env.reset()
_, _, _, info = env.step(0)
required_keys = ["damage_taken", "damage_dealt", "correct_adaptation", "adaptation_stack", "heal_on_cooldown"]
all_present = all(k in info for k in required_keys)
check("Info dict has all required fields", all_present)
check("correct_adaptation is True for matching adapt", info["correct_adaptation"] is True)
check("damage_taken > 0", info["damage_taken"] > 0)
check("damage_dealt = 0 for adapt action", info["damage_dealt"] == 0)
def test_hp_configuration():
print("\n--- Test: HP Configuration ---")
env = MahoragaEnv()
state = env.reset()
check("Agent HP = 1200", state["agent_hp"] == 1200)
check("Enemy HP = 1000", state["enemy_hp"] == 1000)
# ========== ENEMY TESTS ==========
def test_subtype_mapping():
print("\n--- Test: Subtype Mapping ---")
for attack_type in ATTACK_TYPES:
check(f"{attack_type} has subtypes", attack_type in SUBTYPES)
check(f"{attack_type} has 3 subtypes", len(SUBTYPES[attack_type]) == 3)
check("PHYSICAL has PIERCE", "PIERCE" in SUBTYPES["PHYSICAL"])
check("CE has BEAM", "BEAM" in SUBTYPES["CE"])
check("TECHNIQUE has DELAYED", "DELAYED" in SUBTYPES["TECHNIQUE"])
def test_ignore_armor_bypass():
print("\n--- Test: Ignore Armor Bypass ---")
res = new_resistances()
res["PHYSICAL"] = 50
normal_dmg = compute_enemy_damage("PHYSICAL", res, ignore_armor=False)
pierce_dmg = compute_enemy_damage("PHYSICAL", res, ignore_armor=True)
check("Ignore armor does more damage than normal at 50 res", pierce_dmg > normal_dmg)
check("Normal damage = 60", normal_dmg == 60)
check("Armor bypass damage = 72", pierce_dmg == 72)
res_zero = new_resistances()
normal_zero = compute_enemy_damage("PHYSICAL", res_zero, ignore_armor=False)
pierce_zero = compute_enemy_damage("PHYSICAL", res_zero, ignore_armor=True)
check("Same damage at 0 resistance", normal_zero == pierce_zero)
def test_curriculum_phase_1():
print("\n--- Test: Curriculum Phase 1 (Turns 1-5) ---")
enemy = CurriculumEnemy()
for turn in range(1, 6):
attack = enemy.get_attack(turn_number=turn)
check(f"Turn {turn} is PHYSICAL", attack["category"] == "PHYSICAL")
check(f"Turn {turn} has damage", attack["damage"] == 120)
check(f"Turn {turn} has ignore_armor field", "ignore_armor" in attack)
check(f"Turn {turn} subtype valid", attack["subtype"] in SUBTYPES["PHYSICAL"])
def test_curriculum_phase_2():
print("\n--- Test: Curriculum Phase 2 (Cycle, No Deviation) ---")
# Use a fresh enemy and go directly to Phase 2 turns
enemy = CurriculumEnemy()
import utils.constants as c
old_dev = c.PHASE_2_DEVIATION
c.PHASE_2_DEVIATION = 0 # Temporarily disable randomness
a6 = enemy.get_attack(turn_number=6)
check("Phase 2 step 1 = PHYSICAL", a6["category"] == "PHYSICAL")
a7 = enemy.get_attack(turn_number=7)
check("Phase 2 step 2 = CE", a7["category"] == "CE")
a8 = enemy.get_attack(turn_number=8)
check("Phase 2 step 3 = TECHNIQUE", a8["category"] == "TECHNIQUE")
a9 = enemy.get_attack(turn_number=9)
check("Phase 2 step 4 = PHYSICAL (cycle)", a9["category"] == "PHYSICAL")
c.PHASE_2_DEVIATION = old_dev # Restore
def test_curriculum_phase_3():
print("\n--- Test: Curriculum Phase 3 (Target Lowest) ---")
enemy = CurriculumEnemy()
# Phase 3 starts at turn 16
resistances = {"PHYSICAL": 80, "CE": 40, "TECHNIQUE": 0}
attack = enemy.get_attack(turn_number=16, resistances=resistances)
check("Phase 3 targets lowest resistance (TECHNIQUE)", attack["category"] == "TECHNIQUE")
resistances2 = {"PHYSICAL": 0, "CE": 80, "TECHNIQUE": 40}
attack2 = enemy.get_attack(turn_number=17, resistances=resistances2)
check("Phase 3 targets lowest (PHYSICAL)", attack2["category"] == "PHYSICAL")
def test_enemy_dict_format():
print("\n--- Test: Enemy Dict Format ---")
enemy = CurriculumEnemy()
attack = enemy.get_attack(turn_number=1)
check("Returns dict", isinstance(attack, dict))
check("Has 'category'", "category" in attack)
check("Has 'subtype'", "subtype" in attack)
check("Has 'damage'", "damage" in attack)
check("Has 'ignore_armor'", "ignore_armor" in attack)
check("category is PHYSICAL", attack["category"] == "PHYSICAL")
check("subtype is valid", attack["subtype"] in SUBTYPES["PHYSICAL"])
def test_rl_observation_uses_3_types_only():
print("\n--- Test: RL Observation Uses Only 3 Types ---")
env = MahoragaEnv()
env.reset()
state, _, _, _ = env.step(0)
check("last_enemy_attack_type is a valid RL type",
state["last_enemy_attack_type"] in ATTACK_TYPES)
check("Resistances use 3 keys only",
set(state["resistances"].keys()) == {"physical", "ce", "technique"})
# ========== REWARD TESTS ==========
def test_adaptation_reward():
print("\n--- Test: Adaptation Reward ---")
env = MahoragaEnv()
env.reset()
_, reward, _, info = env.step(0)
breakdown = info["reward_breakdown"]
check("Correct adapt gives +0.8", breakdown["adaptation"] == 0.8)
env.reset()
_, reward, _, info = env.step(1)
breakdown = info["reward_breakdown"]
check("Wrong adapt gives 0.0", breakdown["adaptation"] == 0.0)
def test_damage_rewards():
print("\n--- Test: Damage Rewards ---")
env = MahoragaEnv()
env.reset()
_, _, _, info = env.step(0)
breakdown = info["reward_breakdown"]
check("Survival reward is negative (took damage)", breakdown["survival"] < 0)
env.reset()
_, _, _, info = env.step(3)
breakdown = info["reward_breakdown"]
check("Combat reward is positive (dealt damage)", breakdown["combat"] > 0)
env.reset()
_, _, _, info = env.step(0)
breakdown = info["reward_breakdown"]
check("Combat reward is 0 for adapt action", breakdown["combat"] == 0.0)
def test_heal_penalty():
print("\n--- Test: Heal Penalty (Anti-Cowardice) ---")
env = MahoragaEnv()
env.reset()
_, _, _, info = env.step(4)
breakdown = info["reward_breakdown"]
check("Heal at high HP penalized (-1.0)", breakdown["anti_cowardice"] == -1.0)
env.reset()
for _ in range(7):
env.step(1)
check("Agent HP is low enough", env.agent_hp < 0.7 * MAX_HP)
_, _, _, info = env.step(4)
breakdown = info["reward_breakdown"]
check("Heal at low HP NOT penalized (0.0)", breakdown["anti_cowardice"] == 0.0)
def test_terminal_reward():
print("\n--- Test: Terminal Reward ---")
info_win = {"damage_taken": 0, "damage_dealt": 100, "correct_adaptation": False}
state_win = {"agent_hp": 500, "enemy_hp": 0}
rewards = compute_rewards(info_win, state_win, 3, done=True)
check("Win terminal = +10.0", rewards["terminal"] == 10.0)
info_loss = {"damage_taken": 100, "damage_dealt": 0, "correct_adaptation": False}
state_loss = {"agent_hp": 0, "enemy_hp": 500}
rewards = compute_rewards(info_loss, state_loss, 0, done=True)
check("Loss terminal = -8.0", rewards["terminal"] == -8.0)
rewards_nd = compute_rewards(info_win, state_win, 0, done=False)
check("Not done terminal = 0.0", rewards_nd["terminal"] == 0.0)
def test_reward_breakdown_in_info():
print("\n--- Test: Reward Breakdown In Info ---")
env = MahoragaEnv()
env.reset()
_, reward, _, info = env.step(0)
check("Info has reward_breakdown", "reward_breakdown" in info)
breakdown = info["reward_breakdown"]
expected_keys = {"survival", "combat", "adaptation", "anti_cowardice", "efficiency", "terminal", "opportunity"}
check("Breakdown has all 7 components", set(breakdown.keys()) == expected_keys)
total = sum(breakdown.values())
check("Total reward = sum of components", abs(reward - total) < 1e-9)
def test_opportunity_penalty():
print("\n--- Test: Opportunity Penalty ---")
env = MahoragaEnv()
env.reset()
# Build 2 stacks via correct adaptation
env.step(0) # stack=1
_, _, _, info = env.step(0) # stack=2
breakdown = info["reward_breakdown"]
check("Opportunity penalty when stack>=2 and not striking (-0.5)", breakdown["opportunity"] == -0.5)
# Now strike — no penalty
env.reset()
env.step(0) # stack=1
env.step(0) # stack=2
_, _, _, info3 = env.step(3) # Judgment Strike
breakdown3 = info3["reward_breakdown"]
check("No opportunity penalty when striking", breakdown3["opportunity"] == 0.0)
def test_no_reward_for_nothing():
print("\n--- Test: No Free Rewards ---")
env = MahoragaEnv()
env.reset()
env.step(4)
_, reward, _, info = env.step(4)
breakdown = info["reward_breakdown"]
check("No adaptation reward on wasted turn", breakdown["adaptation"] == 0.0)
check("No combat reward on wasted turn", breakdown["combat"] == 0.0)
check("No efficiency bonus on wasted turn", breakdown["efficiency"] == 0.0)
if __name__ == "__main__":
print("=" * 50)
print(" MahoragaEnv Merged System Tests")
print("=" * 50)
# Core mechanics
test_resistance_update()
test_clamp_logic()
test_damage_formula()
test_judgment_damage()
test_episode_termination()
test_actions_behave_correctly()
test_adaptation_stack()
test_invalid_action()
test_heal_cooldown()
test_heal_preserves_resistances()
test_judgment_burst_adaptation_match()
test_info_dict_fields()
test_hp_configuration()
# Enemy system
test_subtype_mapping()
test_ignore_armor_bypass()
test_curriculum_phase_1()
test_curriculum_phase_2()
test_curriculum_phase_3()
test_enemy_dict_format()
test_rl_observation_uses_3_types_only()
# Rewards
test_adaptation_reward()
test_damage_rewards()
test_heal_penalty()
test_terminal_reward()
test_reward_breakdown_in_info()
test_opportunity_penalty()
test_no_reward_for_nothing()
print("\n" + "=" * 50)
print(f" Results: {PASS} passed, {FAIL} failed")
print("=" * 50)
if FAIL > 0:
sys.exit(1)
|