Spaces:
Sleeping
Sleeping
File size: 11,107 Bytes
2badd2f |
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 |
from typing import List
import numpy as np
from engine.game.ai_compat import njit
from engine.game.fast_logic import batch_apply_action
@njit(cache=True)
def step_vectorized(
actions: np.ndarray,
batch_stage: np.ndarray,
batch_energy_vec: np.ndarray,
batch_energy_count: np.ndarray,
batch_continuous_vec: np.ndarray,
batch_continuous_ptr: np.ndarray,
batch_tapped: np.ndarray,
batch_live: np.ndarray,
batch_opp_tapped: np.ndarray,
batch_scores: np.ndarray,
batch_flat_ctx: np.ndarray,
batch_global_ctx: np.ndarray,
batch_hand: np.ndarray,
batch_deck: np.ndarray,
# New: Bytecode Maps
bytecode_map: np.ndarray, # (GlobalOpMapSize, MaxBytecodeLen, 4)
bytecode_index: np.ndarray, # (NumCards, NumAbilities) -> Index in map
):
"""
Step N game environments in parallel using JIT logic and Real Card Data.
"""
# Sync individual scores to global_ctx before stepping
for i in range(len(actions)):
batch_global_ctx[i, 0] = batch_scores[i]
batch_apply_action(
actions,
0, # player_id
batch_stage,
batch_energy_vec,
batch_energy_count,
batch_continuous_vec,
batch_continuous_ptr,
batch_tapped,
batch_scores,
batch_live,
batch_opp_tapped,
batch_flat_ctx,
batch_global_ctx,
batch_hand,
batch_deck,
bytecode_map,
bytecode_index,
)
class VectorGameState:
"""
Manages a batch of independent GameStates for high-throughput training.
"""
def __init__(self, num_envs: int):
self.num_envs = num_envs
self.turn = 1
# Batched state buffers
self.batch_stage = np.full((num_envs, 3), -1, dtype=np.int32)
self.batch_energy_vec = np.zeros((num_envs, 3, 32), dtype=np.int32)
self.batch_energy_count = np.zeros((num_envs, 3), dtype=np.int32)
self.batch_continuous_vec = np.zeros((num_envs, 32, 10), dtype=np.int32)
self.batch_continuous_ptr = np.zeros(num_envs, dtype=np.int32)
self.batch_tapped = np.zeros((num_envs, 3), dtype=np.int32)
self.batch_live = np.zeros((num_envs, 50), dtype=np.int32)
self.batch_opp_tapped = np.zeros((num_envs, 3), dtype=np.int32)
self.batch_scores = np.zeros(num_envs, dtype=np.int32)
# Pre-allocated context buffers (Extreme speed optimization)
self.batch_flat_ctx = np.zeros((num_envs, 64), dtype=np.int32)
self.batch_global_ctx = np.zeros((num_envs, 128), dtype=np.int32)
self.batch_hand = np.zeros((num_envs, 50), dtype=np.int32)
self.batch_deck = np.zeros((num_envs, 50), dtype=np.int32)
# Pre-allocated observation buffer (SAVES ALLOCATION TIME)
self.obs_buffer = np.zeros((num_envs, 320), dtype=np.float32)
# Load Bytecode Map
self._load_bytecode()
self._load_verified_deck_pool()
def _load_bytecode(self):
import json
try:
with open("data/cards_numba.json", "r") as f:
raw_map = json.load(f)
# Convert to numpy array
# Format: key "cardid_abidx" -> List[int]
# storage:
# 1. giant array of bytecodes (N, MaxLen, 4)
# 2. lookup index (CardID, AbIdx) -> Index in giant array
self.max_cards = 2000
self.max_abilities = 4
self.max_len = 64 # Max 64 instructions per ability
# Count unique compiled entries
unique_entries = len(raw_map)
# (Index 0 is empty/nop)
self.bytecode_map = np.zeros((unique_entries + 1, self.max_len, 4), dtype=np.int32)
self.bytecode_index = np.full((self.max_cards, self.max_abilities), 0, dtype=np.int32)
idx_counter = 1
for key, bc_list in raw_map.items():
cid, aid = map(int, key.split("_"))
if cid < self.max_cards and aid < self.max_abilities:
# reshape list to (M, 4)
bc_arr = np.array(bc_list, dtype=np.int32).reshape(-1, 4)
length = min(bc_arr.shape[0], self.max_len)
self.bytecode_map[idx_counter, :length] = bc_arr[:length]
self.bytecode_index[cid, aid] = idx_counter
idx_counter += 1
print(f" [VectorEnv] Loaded {unique_entries} compiled abilities.")
except FileNotFoundError:
print(" [VectorEnv] Warning: data/cards_numba.json not found. Using empty map.")
self.bytecode_map = np.zeros((1, 64, 4), dtype=np.int32)
self.bytecode_index = np.zeros((1, 1), dtype=np.int32)
def _load_verified_deck_pool(self):
import json
try:
# Load Verified List
with open("verified_card_pool.json", "r", encoding="utf-8") as f:
verified_data = json.load(f)
# Load DB to map CardNo -> CardID
with open("data/cards_compiled.json", "r", encoding="utf-8") as f:
db_data = json.load(f)
self.verified_card_ids = []
# Map numbers to IDs
card_no_map = {}
for cid, cdata in db_data["member_db"].items():
card_no_map[cdata["card_no"]] = int(cid)
for v_no in verified_data.get("verified_abilities", []):
if v_no in card_no_map:
self.verified_card_ids.append(card_no_map[v_no])
# Fallback
if not self.verified_card_ids:
print(" [VectorEnv] Warning: No verified cards found. Using ID 1.")
self.verified_card_ids = [1]
else:
print(f" [VectorEnv] Loaded {len(self.verified_card_ids)} verified cards for training.")
self.verified_card_ids = np.array(self.verified_card_ids, dtype=np.int32)
except Exception as e:
print(f" [VectorEnv] Deck Load Error: {e}")
self.verified_card_ids = np.array([1], dtype=np.int32)
def reset(self, indices: List[int] = None):
"""Reset specified environments (or all if indices is None)."""
if indices is None:
indices = list(range(self.num_envs))
# Optimization: Bulk operations for indices if supported,
# but for now loop is fine (reset is rare compared to step)
# Prepare a random deck selection to broadcast?
# Actually random.choice is fast.
for i in indices:
self.batch_stage[i].fill(-1)
self.batch_energy_vec[i].fill(0)
self.batch_energy_count[i].fill(0)
self.batch_continuous_vec[i].fill(0)
self.batch_continuous_ptr[i] = 0
self.batch_tapped[i].fill(0)
self.batch_live[i].fill(0)
self.batch_opp_tapped[i].fill(0)
self.batch_scores[i] = 0
# Reset contexts
self.batch_flat_ctx[i].fill(0)
self.batch_global_ctx[i].fill(0)
# Initialize Deck with Verified Cards (Random 50)
# Fast choice from verified pool
if len(self.verified_card_ids) > 0:
dk = np.random.choice(self.verified_card_ids, 50)
self.batch_deck[i] = dk
# Initialize Hand (Draw 5 from deck)
# Simple simulation: Move top 5 deck to hand
self.batch_hand[i, :5] = self.batch_deck[i, :5]
# Shift deck? Or just pointer?
# For this benchmark we assume infinite deck or simple pointer logic via opcodes.
# But the 'hand' array needs to be populated for gameplay to start.
self.turn = 1
def step(self, actions: np.ndarray):
"""Apply a batch of actions across all environments."""
step_vectorized(
actions,
self.batch_stage,
self.batch_energy_vec,
self.batch_energy_count,
self.batch_continuous_vec,
self.batch_continuous_ptr,
self.batch_tapped,
self.batch_live,
self.batch_opp_tapped,
self.batch_scores,
self.batch_flat_ctx,
self.batch_global_ctx,
self.batch_hand,
self.batch_deck,
self.bytecode_map,
self.bytecode_index,
)
# Simplified turn advancement
# In real VectorEnv, this would be managed by the engine rules
pass
def get_observations(self):
"""Return a batched observation for RL models."""
return encode_observations_vectorized(
self.num_envs,
self.batch_stage,
self.batch_energy_count,
self.batch_tapped,
self.batch_scores,
self.turn,
self.obs_buffer,
)
@njit(cache=True)
def encode_observations_vectorized(
num_envs: int,
batch_stage: np.ndarray, # (N, 3)
batch_energy_count: np.ndarray, # (N, 3)
batch_tapped: np.ndarray, # (N, 3)
batch_scores: np.ndarray, # (N,)
turn_number: int,
observations: np.ndarray, # (N, 320)
):
# Reset buffer (extremely fast on pre-allocated)
observations.fill(0.0)
max_id_val = 2000.0 # Normalization constant
for i in range(num_envs):
# --- 1. METADATA [0:36] ---
# Phase (Simplify: Always Main Phase=1 for now in vector env)
# Phase 1=Start, 2=Draw, 3=Main... Main is index 3+2=5?
# GameState logic: phase_val = int(phase) + 2. Main is 3. So 5.
observations[i, 5] = 1.0
# Current Player [16:18] - Always Player 0 for this vector view
observations[i, 16] = 1.0
# --- 2. HAND [36:168] ---
# VectorEnv doesn't track hand yet. Leave 0.0.
# --- 3. SELF STAGE [168:204] (3 slots * 12 features) ---
for slot in range(3):
cid = batch_stage[i, slot]
base = 168 + slot * 12
if cid >= 0:
observations[i, base] = 1.0
observations[i, base + 1] = cid / max_id_val
observations[i, base + 2] = 1.0 if batch_tapped[i, slot] else 0.0
# Mock attributes (since we don't have full DB access inside JIT yet)
# In real imp, we'd pass arrays like member_costs
observations[i, base + 3] = 0.5 # Default power
# Energy Count
observations[i, base + 11] = min(batch_energy_count[i, slot] / 5.0, 1.0)
# --- 4. OPPONENT STAGE [204:240] ---
# Not tracked in partial vector env yet.
# --- 5. LIVE ZONE [240:270] ---
# Not tracked in partial vector env yet.
# --- 6. SCORES [270:272] ---
observations[i, 270] = min(batch_scores[i] / 5.0, 1.0)
return observations
|