Spaces:
Running
Running
File size: 6,741 Bytes
2d88649 |
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 |
import numpy as np
from engine.game.game_state import GameState
class UnifiedObservationEncoder:
"""
Translates current GameState into various historic observation formats.
"""
@staticmethod
def encode(state: GameState, dim: int, player_idx: int = None) -> np.ndarray:
if player_idx is None:
player_idx = state.current_player
if dim == 8192:
return UnifiedObservationEncoder._encode_8192(state, player_idx)
elif dim == 320:
return UnifiedObservationEncoder._encode_320(state, player_idx)
elif dim == 128:
return UnifiedObservationEncoder._encode_128(state, player_idx)
else:
raise ValueError(f"Unsupported observation dimension: {dim}")
@staticmethod
def _encode_8192(state: GameState, player_idx: int) -> np.ndarray:
from ai.vector_env import VectorGameState as VGS
from ai.vector_env import encode_observations_vectorized
p = state.players[player_idx]
opp = state.players[1 - player_idx]
# Max ID for normalization is handled inside encoder
# Prepare inputs strictly matching VectorEnv.encode_observations_vectorized signature
# 1. Num Envs (1)
# 2. Batch Hand (1, 60)
# 3. Batch Stage (1, 3)
# 4. Batch Energy Count (1, 3)
# 5. Batch Tapped (1, 3)
# 6. Batch Scores (1,)
# 7. Opp Scores (1,)
# 8. Opp Stage (1, 3)
# 9. Opp Tapped (1, 3)
# 10. Card Stats (from VGS)
# 11. Global Context (1, 128)
# 12. Batch Live (1, 50)
# 13. Batch Opp History (1, 50)
# 14. Turn Number
# 15. Obs Buffer (1, 8192)
# --- Allocations ---
batch_hand = np.zeros((1, 60), dtype=np.int32)
batch_stage = np.full((1, 3), -1, dtype=np.int32)
batch_energy_count = np.zeros((1, 3), dtype=np.int32)
batch_tapped = np.zeros((1, 3), dtype=np.int32)
batch_live = np.zeros((1, 50), dtype=np.int32)
opp_stage = np.full((1, 3), -1, dtype=np.int32)
opp_tapped = np.zeros((1, 3), dtype=np.int32)
opp_history = np.zeros((1, 50), dtype=np.int32)
# --- Population ---
# Hand
h_len = min(len(p.hand), 60)
for i in range(h_len):
batch_hand[0, i] = p.hand[i]
# Stage
for i in range(3):
batch_stage[0, i] = p.stage[i]
batch_energy_count[0, i] = p.stage_energy_count[i]
batch_tapped[0, i] = 1 if p.tapped_members[i] else 0
opp_stage[0, i] = opp.stage[i]
opp_tapped[0, i] = 1 if opp.tapped_members[i] else 0
# Live Zone
# Assuming GameState has p.live_zone list or similar?
# GameState definition usually implies 'success_lives' are won lives.
# Active lives might be tracked elsewhere?
# If not available, leave as zeros.
# Checking GameState... usually just has success_lives. Active lives are transient in legacy?
# VectorEnv tracks them. Legacy might not.
# Scores
batch_scores = np.array([len(p.success_lives)], dtype=np.int32)
opp_scores = np.array([len(opp.success_lives)], dtype=np.int32)
# Global Context
g_ctx = np.zeros((1, 128), dtype=np.int32)
g_ctx[0, 0] = len(p.success_lives) # SC
g_ctx[0, 1] = len(opp.success_lives) # OS
g_ctx[0, 2] = len(p.discard) # TR
g_ctx[0, 3] = len(p.hand) # HD
g_ctx[0, 5] = p.energy_count # EN
g_ctx[0, 6] = len(p.main_deck) # DK
g_ctx[0, 8] = 5 # PHASE (Main) - Legacy default
# Opponent History (Trash top cards?)
op_h_len = min(len(opp.discard), 50)
for i in range(op_h_len):
# LIFO? VectorEnv usually assumes LIFO or FIFO depending on implementation.
# Usually end is top.
opp_history[0, i] = opp.discard[-(i + 1)]
# Output buffer
obs = np.zeros((1, 8192), dtype=np.float32)
if not hasattr(UnifiedObservationEncoder, "_vgs_cache"):
UnifiedObservationEncoder._vgs_cache = VGS(1)
vgs = UnifiedObservationEncoder._vgs_cache
encode_observations_vectorized(
1,
batch_hand,
batch_stage,
batch_energy_count,
batch_tapped,
batch_scores,
opp_scores,
opp_stage,
opp_tapped,
vgs.card_stats,
g_ctx,
batch_live,
opp_history,
state.turn_number,
obs,
)
return obs[0]
@staticmethod
def _encode_320(state: GameState, player_idx: int) -> np.ndarray:
# LEGACY 320 (First Speed-up Era)
# Replicates the encoding from ai/vector_env_legacy.py exactly.
# This era ONLY saw Self Stage and Self Score. Hand/Opp were 0.
obs = np.zeros(320, dtype=np.float32)
p = state.players[player_idx]
max_id_val = 2000.0 # Standard for VectorEnv
# Phase [5] = 1.0 (Mocking Main Phase index from Legacy VectorEnv)
obs[5] = 1.0
# Current Player [16]
obs[16] = 1.0
# Stage [168:204] (3 slots * 12 features)
# Note: Hand [36:168] remains 0.0 as in legacy training.
for i in range(3):
cid = p.stage[i]
base = 168 + i * 12
if cid >= 0:
obs[base] = 1.0 # Exist
obs[base + 1] = cid / max_id_val
# Legacy energy count was normalized by 5.0
obs[base + 11] = min(p.stage_energy_count[i] / 5.0, 1.0)
# Score [270] (Self Score normalized by 5.0 in legacy)
obs[270] = min(len(p.success_lives) / 5.0, 1.0)
return obs
@staticmethod
def _encode_128(state: GameState, player_idx: int) -> np.ndarray:
# 128-dim is the global_ctx vector
p = state.players[player_idx]
opp = state.players[1 - player_idx]
g_ctx = np.zeros(128, dtype=np.float32)
# Standard normalization from AlphaZero era
g_ctx[0] = len(p.success_lives) / 3.0
g_ctx[1] = len(opp.success_lives) / 3.0
g_ctx[2] = len(p.discard) / 50.0
g_ctx[3] = len(p.hand) / 50.0 # Normalized to deck size usually
g_ctx[5] = p.energy_count / 10.0
g_ctx[6] = len(p.main_deck) / 50.0
# Turn info
g_ctx[10] = state.turn_number / 20.0
g_ctx[11] = 1.0 if state.current_player == player_idx else 0.0
return g_ctx
|