File size: 12,972 Bytes
0d5371b | 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 | """
Stormer Result Evaluation and Visualization.
Computes per-channel RMSE and ACC (Anomaly Correlation Coefficient)
for model predictions against ground truth at the corresponding lead times.
Usage:
python scripts/result.py
"""
import numpy as np
import matplotlib.pyplot as plt
import os
import sys
from pathlib import Path
root_path = Path(__file__).parent.parent
sys.path.append(str(root_path))
import glob
import h5py
from datetime import datetime, timedelta
from tqdm import tqdm
from onescience.utils.YParams import YParams
from matplotlib import rcParams
rcParams['mathtext.fontset'] = 'stix'
rcParams['axes.linewidth'] = 0.9
rcParams['xtick.major.width'] = 0.9
rcParams['ytick.major.width'] = 0.9
def get_metadata(data_dir, channels):
"""Read variable names and time_step from HDF5 attrs."""
h5_files = sorted(glob.glob(os.path.join(data_dir, "data", "*.h5")))
with h5py.File(h5_files[0], "r") as f:
ds = f["fields"]
all_variables = [
v.decode() if isinstance(v, bytes) else v for v in ds.attrs["variables"]
]
time_step = int(ds.attrs["time_step"])
channel_indices = [all_variables.index(v) for v in channels]
# Find all prediction files (now include lead time suffix)
total_files = [f for f in os.listdir('./result/output/') if f.endswith('.npy')]
total_files.sort()
return total_files, channel_indices, time_step
def filename_to_datetime(filename_base):
"""Convert YYYYMMDDHH base filename to datetime."""
return datetime.strptime(filename_base, "%Y%m%d%H")
def parse_pred_filename(filename):
"""Parse prediction filename like '2003010206_lead72h.npy'.
Returns:
base_time: datetime of input time
lead_hours: lead time in hours
"""
name = filename.replace('.npy', '')
parts = name.split('_lead')
base_str = parts[0]
lead_str = parts[1].replace('h', '')
base_time = datetime.strptime(base_str, "%Y%m%d%H")
lead_hours = int(lead_str)
return base_time, lead_hours
def get_ground_truth(data_dir, target_time, channel_indices, time_step):
"""Load ground truth from HDF5 at the target datetime.
Args:
data_dir: path to data directory containing data/*.h5 files
target_time: datetime of the target time step
channel_indices: indices of desired channels
time_step: hours between consecutive frames
Returns:
label: (C, H, W) numpy array, or None if not found
"""
year = target_time.year
year_start = datetime(year, 1, 1)
hours_since_year_start = (target_time - year_start).total_seconds() / 3600
t_idx = int(hours_since_year_start / time_step)
h5_path = os.path.join(data_dir, 'data', f'{year}.h5')
if not os.path.exists(h5_path):
# Try next/last year (for year boundary)
for adj_year in [year - 1, year + 1]:
alt_path = os.path.join(data_dir, 'data', f'{adj_year}.h5')
if os.path.exists(alt_path):
h5_path = alt_path
if adj_year < year:
# Target is early in year, data from previous year
prev_start = datetime(adj_year, 1, 1)
year_len = int((datetime(adj_year + 1, 1, 1) - prev_start).total_seconds() / 3600)
t_idx = year_len // time_step + int(
(target_time - datetime(year, 1, 1)).total_seconds() / 3600 / time_step
)
break
else:
return None
try:
with h5py.File(h5_path, "r") as f:
T_total = f["fields"].shape[0]
if t_idx >= T_total or t_idx < 0:
return None
label = f["fields"][t_idx] # [C_total, H, W]
label = label[channel_indices] # [C_selected, H, W]
return label
except Exception:
return None
def compute_metrics(total_files, channel_indices, time_step, data_dir, clim_mean):
"""Compute per-channel RMSE and ACC."""
n_channels = len(channel_indices)
if os.path.exists('./result/rmse.npy') and os.path.exists('./result/acc.npy'):
print("π Loading cached metrics...")
return np.load('./result/rmse.npy'), np.load('./result/acc.npy')
clim_mean = clim_mean[0, :, :, :] # (C, H, W)
channel_rmse = np.zeros(n_channels)
acc_numerator = np.zeros(n_channels)
acc_pred_sq = np.zeros(n_channels)
acc_label_sq = np.zeros(n_channels)
valid_count = 0
for file in tqdm(total_files, unit="files", desc="Computing metrics"):
base_time, lead_hours = parse_pred_filename(file)
target_time = base_time + timedelta(hours=lead_hours)
# Load ground truth at target time
label = get_ground_truth(data_dir, target_time, channel_indices, time_step)
if label is None:
continue
# Load prediction
pred = np.load(f'result/output/{file}').squeeze() # (C, H, W)
if pred.shape != label.shape:
# Handle shape mismatch
if pred.ndim == 4:
pred = pred[0]
if pred.shape != label.shape:
continue
# RMSE
channel_rmse += np.sqrt(np.mean((label - pred) ** 2, axis=(1, 2)))
# ACC (Anomaly Correlation Coefficient)
label_anom = label - clim_mean
pred_anom = pred - clim_mean
acc_numerator += np.sum(pred_anom * label_anom, axis=(1, 2))
acc_pred_sq += np.sum(pred_anom ** 2, axis=(1, 2))
acc_label_sq += np.sum(label_anom ** 2, axis=(1, 2))
valid_count += 1
if valid_count == 0:
print("β οΈ No valid predictions found.")
return np.zeros(n_channels), np.zeros(n_channels)
channel_rmse /= valid_count
channel_acc = acc_numerator / (np.sqrt(acc_pred_sq * acc_label_sq) + 1e-8)
np.save('./result/rmse.npy', channel_rmse)
np.save('./result/acc.npy', channel_acc)
return channel_rmse, channel_acc
def show_result_table(channels, channel_rmse, channel_acc):
"""Print formatted RMSE/ACC table."""
w = 40
print(f"\nβ{'β' * (w + 2)}β¬{'β' * 14}β¬{'β' * 14}β")
print(f"β {'Channel':<{w}} β {'RMSE':>12} β {'ACC':>12} β")
print(f"β{'β' * (w + 2)}βΌ{'β' * 14}βΌ{'β' * 14}β€")
# Show first 10 + key vars, then average
display_idx = list(range(min(10, len(channels))))
# Add key vars if not in first 10
for key in ['geopotential_500', 'temperature_850', '2m_temperature']:
if key in channels:
idx = channels.index(key)
if idx not in display_idx:
display_idx.append(idx)
for i in display_idx:
ch = channels[i]
print(f"β {ch:<{w}} β {channel_rmse[i]:>12.4f} β {channel_acc[i]:>12.4f} β")
print(f"β{'β' * (w + 2)}βΌ{'β' * 14}βΌ{'β' * 14}β€")
print(f"β {'Average':<{w}} β {np.mean(channel_rmse):>12.4f} β {np.mean(channel_acc):>12.4f} β")
print(f"β{'β' * (w + 2)}β΄{'β' * 14}β΄{'β' * 14}β")
def plot_prediction(label, pred, var, filename):
"""Plot truth, prediction, and difference for a single variable."""
fig, axes = plt.subplots(1, 3, figsize=(15, 4))
xtick_labels = ['180Β°W', '90Β°W', '0Β°', '90Β°E', '180Β°E']
ytick_labels = ['90Β°S', '45Β°S', '0Β°', '45Β°N', '90Β°N']
xticks = np.linspace(0, label.shape[-1] - 1, 5)
yticks = np.linspace(0, label.shape[-2] - 1, 5)
vmin = min(label.min(), pred.min())
vmax = max(label.max(), pred.max())
diff = label - pred
rmse = np.sqrt(np.mean(diff ** 2))
diff_abs_max = np.abs(diff).max()
plot_configs = [
{'data': label, 'title': 'Truth', 'cmap': 'viridis',
'vmin': vmin, 'vmax': vmax},
{'data': pred, 'title': 'Prediction', 'cmap': 'viridis',
'vmin': vmin, 'vmax': vmax},
{'data': diff, 'title': f'Difference (RMSE={rmse:.2f})',
'cmap': 'RdBu_r', 'vmin': -diff_abs_max, 'vmax': diff_abs_max},
]
for ax, cfg in zip(axes, plot_configs):
im = ax.imshow(cfg['data'], cmap=cfg['cmap'],
vmin=cfg['vmin'], vmax=cfg['vmax'])
ax.set_title(cfg['title'], fontsize=12, pad=4)
ax.set_xlabel('Longitude')
ax.set_ylabel('Latitude')
ax.set_xticks(xticks)
ax.set_xticklabels(xtick_labels)
ax.set_yticks(yticks)
ax.set_yticklabels(ytick_labels)
plt.colorbar(im, ax=ax, orientation='horizontal')
fig.suptitle(var, fontsize=14, fontweight='bold', y=0.98)
plt.savefig(filename, dpi=300, bbox_inches='tight')
plt.close()
def plot_loss_curves(train_loss, valid_loss):
"""Plot training and validation loss curves."""
mask = ~(np.isnan(train_loss) | np.isnan(valid_loss))
train_loss = train_loss[mask]
valid_loss = valid_loss[mask]
if len(train_loss) == 0:
print("β οΈ No loss data to plot.")
return
fig, ax = plt.subplots(figsize=(5, 3.5))
colors = {'train': '#2563EB', 'valid': '#EA580C'}
epochs = np.arange(1, len(train_loss) + 1)
ax.plot(epochs, train_loss, color=colors['train'], linewidth=1.5, label='Train')
ax.plot(epochs, valid_loss, color=colors['valid'], linewidth=1.5,
label='Valid', linestyle='--')
min_idx = np.argmin(valid_loss)
ax.scatter(epochs[min_idx], valid_loss[min_idx],
color=colors['valid'], s=40, zorder=5, edgecolors='white')
ax.annotate(f'Best: {valid_loss[min_idx]:.3f}',
xy=(epochs[min_idx], valid_loss[min_idx]),
xytext=(10, 10), textcoords='offset points',
fontsize=8, color=colors['valid'],
arrowprops=dict(arrowstyle='-', color=colors['valid'], lw=0.5))
ax.set(xlabel='Epoch', ylabel='Loss', xlim=(0, len(train_loss) + 1))
ax.legend(frameon=False, loc='upper right')
ax.grid(True, linestyle='--', alpha=0.3)
ax.spines[['top', 'right']].set_visible(False)
plt.tight_layout()
plt.savefig('./result/loss.png', dpi=300, bbox_inches='tight')
plt.close()
print("β
Loss curves saved to './result/loss.png'")
if __name__ == "__main__":
current_path = os.getcwd()
sys.path.append(current_path)
config_file_path = os.path.join(current_path, 'conf/config.yaml')
cfg = YParams(config_file_path, 'model')
cfg_data = YParams(config_file_path, "datapipe")
os.makedirs('./result/', exist_ok=True)
# ---- Plot loss curves ----
train_loss_file = f"{cfg.checkpoint_dir}/trloss.npy"
valid_loss_file = f"{cfg.checkpoint_dir}/valoss.npy"
if os.path.exists(train_loss_file) and os.path.exists(valid_loss_file):
train_loss = np.load(train_loss_file)
valid_loss = np.load(valid_loss_file)
plot_loss_curves(train_loss, valid_loss)
else:
print("β οΈ Loss files not found β skipping loss plot.")
# ---- Compute metrics ----
data_dir = cfg_data.dataset.data_dir
total_files, channel_indices, time_step = get_metadata(
data_dir, cfg_data.dataset.channels
)
if len(total_files) == 0:
print("β οΈ No prediction files in './result/output/' β skipping metrics.")
sys.exit(0)
# Load climate mean for ACC
h5_files = sorted(glob.glob(os.path.join(data_dir, "data", "*.h5")))
with h5py.File(h5_files[0], "r") as f:
mu = f["global_means"][:]
clim_mean = mu[:, channel_indices, :, :]
channel_rmse, channel_acc = compute_metrics(
total_files, channel_indices, time_step, data_dir, clim_mean
)
show_result_table(cfg_data.dataset.channels, channel_rmse, channel_acc)
# ---- Plot example predictions ----
test_year = cfg_data.dataset.test_time[0]
eg_files = [f for f in total_files if f.startswith(f'{test_year}')][:3]
key_vars = ['2m_temperature', 'geopotential_500', 'temperature_500']
available_vars = [v for v in key_vars if v in cfg_data.dataset.channels]
channel_index_map = {v: cfg_data.dataset.channels.index(v) for v in available_vars}
if eg_files:
print(f"\nπ Plotting example predictions for: {available_vars}")
for file in eg_files:
base_time, lead_hours = parse_pred_filename(file)
target_time = base_time + timedelta(hours=lead_hours)
label = get_ground_truth(data_dir, target_time, channel_indices, time_step)
if label is None:
print(f" β οΈ No ground truth for {file}")
continue
pred = np.load(f'result/output/{file}').squeeze()
if pred.ndim == 4:
pred = pred[0]
for var in available_vars:
idx = channel_index_map[var]
out_file = f'./result/{file[:-4]}_{var}.png'
if pred.shape == label.shape:
plot_prediction(label[idx], pred[idx], var, out_file)
print(f' β
{out_file}')
print("\nβ
Evaluation complete.")
|