Fix: CsI(Tl) non-linear response correction + detector calibration overhaul

Root cause of Am-241 misidentification: the Radiacode 103's CsI(Tl) crystal
shifts low-energy peaks upward (59.5 keV → 71.6 keV for Am-241) due to
non-proportional scintillation response. The model was trained on theoretical
peak positions and couldn't match the shifted real peaks.

Changes:
- Add inverse CsI(Tl) non-linear correction to inference pipeline
  (radiacode_monitor.py, web/config.py, test_detection.py)
  E_apparent = E_true * (1 + 0.37 * exp(-E_true/100))
  Corrects channel mapping so peaks appear at theoretical energies
- Fix energy calibration: DetectorConfig now uses E = 0.33 + 2.97*ch
  with 1023 channels, matching the real detector (was energy_min=20,
  skip_first_channel=True, different channel width)
- Add K-escape peaks for CsI(Tl) iodine X-ray escape (E - 28.5 keV)
- Add asymmetric peak shapes for low-energy tails (< 200 keV)
- Add log1p normalization in dataset and inference (replaces max-norm)
- Add background-subtracted training mode (subtract_background flag)
- Add low-signal augmentation (0.01-5 Bq activities, 30-300s durations)
- Update docker-compose.yml: batch_size=32, duration=30-300s,
  CSI_NONLINEAR_ALPHA/BETA env vars for detect and web
- Web dashboard: apply CsI correction to displayed spectra
- Various UI fixes (Chart.js width, zoom/pan, isotope lines)

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

web/app/config.py

@@ -1,4 +1,5 @@
 import os
+import numpy as np
 from pathlib import Path
 
 STATE_PATH = Path(os.environ.get("STATE_PATH", "/data/monitor_state.json"))
@@ -11,8 +12,47 @@ ISOTOPE_INDEX_PATH = Path(os.environ.get("ISOTOPE_INDEX_PATH", "/models/vega_iso
 ENERGY_OFFSET = float(os.environ.get("ENERGY_CALIBRATION_OFFSET", "0.33"))
 ENERGY_SLOPE = float(os.environ.get("ENERGY_CALIBRATION_SLOPE", "2.97"))
 NUM_CHANNELS = 1023  # Last channel (1023) is overflow bin, excluded from display
+ENERGY_MIN = 30.0  # keV - detector lower limit
+ENERGY_MAX = 3000.0  # keV - detector upper limit (3 MeV)
+
+# CsI(Tl) non-linear response correction parameters
+# Matches the detector's non-proportional scintillation response
+CSI_NONLINEAR_ALPHA = float(os.environ.get("CSI_NONLINEAR_ALPHA", "0.37"))
+CSI_NONLINEAR_BETA = float(os.environ.get("CSI_NONLINEAR_BETA", "100.0"))
+
+
+def correct_csi_nonlinear(spectrum, num_channels=1023):
+    """Apply inverse CsI(Tl) non-linear response correction.
+
+    Remaps spectrum channels so peaks appear at their theoretical energy
+    positions, correcting for the detector's non-proportional scintillation
+    response that shifts low-energy peaks upward.
+    """
+    alpha = CSI_NONLINEAR_ALPHA
+    beta = CSI_NONLINEAR_BETA
+    output_channels = np.arange(num_channels, dtype=np.float64)
+    e_true = ENERGY_OFFSET + ENERGY_SLOPE * output_channels
+    e_apparent = e_true * (1 + alpha * np.exp(-e_true / beta))
+    source_channels = (e_apparent - ENERGY_OFFSET) / ENERGY_SLOPE
+    source_channels = np.clip(source_channels, 0, num_channels - 1.001)
+    lower = np.floor(source_channels).astype(int)
+    upper = np.minimum(lower + 1, num_channels - 1)
+    frac = source_channels - lower
+    return spectrum[lower] * (1 - frac) + spectrum[upper] * frac
 
 
 def energy_axis():
-    """Generate energy axis in keV from channel numbers."""
-    return [round(ENERGY_OFFSET + ENERGY_SLOPE * i, 2) for i in range(NUM_CHANNELS)]
+    """Generate energy axis in keV from channel numbers, clipped to detector range."""
+    axis = [round(ENERGY_OFFSET + ENERGY_SLOPE * i, 2) for i in range(NUM_CHANNELS)]
+    return [e for e in axis if ENERGY_MIN <= e <= ENERGY_MAX]
+
+
+def energy_mask():
+    """Return boolean mask of channels within detector energy range."""
+    return [ENERGY_MIN <= ENERGY_OFFSET + ENERGY_SLOPE * i <= ENERGY_MAX for i in range(NUM_CHANNELS)]
+
+
+def clip_to_range(arr):
+    """Clip array to detector energy range using energy mask."""
+    mask = energy_mask()
+    return [arr[i] for i in range(len(arr)) if mask[i]]