Update RTC channel-specific normalization helper

nasar_dataset.py

@@ -22,19 +22,36 @@ import numpy as np
 import pandas as pd
 
 
-def _normalize_rtc(arr: np.ndarray, norm_stats: tuple[float, float]) -> np.ndarray:
-    """Clip to [p1, p99] and rescale to [0, 1]. Zeros stay zero."""
-    p1, p99 = norm_stats
-    valid = arr > 0
-    arr = np.where(valid, np.clip((arr - p1) / (p99 - p1), 0.0, 1.0), 0.0)
-    return arr.astype(np.float32)
+RTC_COPOL_NORM_STATS = (0.150255, 2.387287)
+RTC_XPOL_NORM_STATS = (0.02, 1.09)
+DEFAULT_RTC_NORM_STATS = (RTC_COPOL_NORM_STATS, RTC_XPOL_NORM_STATS)
+RTCNormStats = tuple[float, float] | tuple[tuple[float, float], tuple[float, float]]
+
+
+def _normalize_rtc(arr: np.ndarray, norm_stats: RTCNormStats) -> np.ndarray:
+    """Clip to channel p1/p99 and rescale to [0, 1]. Zeros stay zero.
+
+    ``norm_stats`` can be either a shared ``(p1, p99)`` tuple or
+    ``((co_p1, co_p99), (x_p1, x_p99))`` for channel-specific RTC normalization.
+    """
+    out = arr.astype(np.float32, copy=True)
+    first = norm_stats[0]
+    if isinstance(first, (tuple, list, np.ndarray)):
+        channel_stats = norm_stats
+    else:
+        channel_stats = (norm_stats,) * out.shape[0]
+
+    for ch, (p1, p99) in enumerate(channel_stats[: out.shape[0]]):
+        valid = out[ch] > 0
+        out[ch] = np.where(valid, np.clip((out[ch] - p1) / (p99 - p1), 0.0, 1.0), 0.0)
+    return out
 
 
 def _prepare_rtc(
     arr: np.ndarray,
     log_scale: bool,
     alpha: float,
-    norm_stats: tuple[float, float] | None,
+    norm_stats: RTCNormStats | None,
 ) -> np.ndarray:
     """Apply the same RTC preprocessing used during NASAR pretraining."""
     arr = arr.astype(np.float32, copy=True)
@@ -145,7 +162,7 @@ class NASARRTCDataset(_NASARBase):
         years: list[int] | None = None,
         random_temporal: bool = True,
         min_quality_score: float | None = 0.53,
-        norm_stats: tuple[float, float] | None = (0.14, 2.38),
+        norm_stats: RTCNormStats | None = DEFAULT_RTC_NORM_STATS,
         shard_size: int = 10_000,
     ) -> None:
         super().__init__(