added gap validation.

gap_validation/batch_spatial.py

@@ -0,0 +1,111 @@
+"""Run spatial NSE_S2 gap validation for all thesis sites (best BtI scenario per site)."""
+
+from __future__ import annotations
+
+import argparse
+import json
+import re
+from pathlib import Path
+
+from gap_validation.calendar import DEFAULT_GAP_LENGTHS, TRANSITIONS
+from gap_validation.run import run_validation
+
+# Primary season per site (matches scripts/export_thesis_tables.py).
+PRIMARY_SEASON = {
+    "forthgr": 2024,
+    "innsbruck": 2024,
+    "pitsalu": 2024,
+    "vindeln2": 2023,
+    "sunflowerjerez1": 2024,
+    "institutekarnobat": 2024,
+}
+
+
+def _site_positions(geojson: Path) -> dict[str, tuple[float, float]]:
+    data = json.loads(geojson.read_text(encoding="utf-8"))
+    out: dict[str, tuple[float, float]] = {}
+    for feat in data.get("features", []):
+        props = feat.get("properties") or {}
+        name = props.get("sitename")
+        coords = (feat.get("geometry") or {}).get("coordinates")
+        if not name or not coords or len(coords) < 2:
+            continue
+        lon, lat = float(coords[0]), float(coords[1])
+        out[str(name)] = (lat, lon)
+    return out
+
+
+def _parse_scenario(key: str) -> tuple[str, int | None, str]:
+    """``aggressive_sigma20`` → (strategy, sigma, bti)."""
+    mode = "itb" if key.endswith("_itb") else "bti"
+    base = key.replace("_itb", "")
+    m = re.match(r"^(aggressive|nonaggressive)_sigma(\d+)$", base)
+    if not m:
+        raise ValueError(f"Cannot parse scenario key: {key!r}")
+    strategy = m.group(1)
+    sigma = int(m.group(2))
+    return strategy, sigma if sigma == 30 else (None if sigma == 20 else sigma), mode
+
+
+def _best_bti_from_metrics(metrics_path: Path) -> str | None:
+    if not metrics_path.is_file():
+        return None
+    temporal = json.loads(metrics_path.read_text(encoding="utf-8")).get("temporal") or {}
+    best_key, best_nse = None, None
+    for k, v in temporal.items():
+        if not k.endswith("_itb") and isinstance(v, dict):
+            n = v.get("nse_pc")
+            if isinstance(n, (int, float)) and (best_nse is None or n > best_nse):
+                best_nse = n
+                best_key = k
+    return best_key
+
+
+def main() -> None:
+    ap = argparse.ArgumentParser(description="Batch spatial gap validation (six sites).")
+    ap.add_argument("--data-dir", type=Path, default=Path("data"))
+    ap.add_argument("--sites-geojson", type=Path, default=Path("data/sites.geojson"))
+    ap.add_argument("--skip-fusion", action="store_true")
+    ap.add_argument("--write-manifest-only", action="store_true")
+    ap.add_argument(
+        "--gap-days",
+        type=int,
+        action="append",
+        help="Filter gap lengths (default: all 15 and 30 in manifest).",
+    )
+    args = ap.parse_args()
+    positions = _site_positions(args.sites_geojson)
+    gap_filter = args.gap_days
+
+    for site, season in sorted(PRIMARY_SEASON.items()):
+        pos = positions.get(site)
+        if not pos:
+            print(f"[skip] No coordinates for {site}")
+            continue
+        metrics_path = args.data_dir / site / str(season) / "metrics.json"
+        scenario_key = _best_bti_from_metrics(metrics_path)
+        if not scenario_key:
+            print(f"[skip] {site} {season}: no metrics.json / BtI scenarios")
+            continue
+        strategy, sigma, mode = _parse_scenario(scenario_key)
+        sigma_kw = 30 if sigma == 30 else None
+        print(f"=== {site} {season} {scenario_key} ===")
+        out = run_validation(
+            site,
+            season,
+            pos,
+            strategy,
+            sigma_kw,
+            mode,
+            skip_manifest=False,
+            skip_fusion=args.skip_fusion,
+            write_manifest_only=args.write_manifest_only,
+            gap_days_filter=gap_filter,
+            transition_filter=None,
+            s2_calendar_strategy=strategy,
+        )
+        print(out)
+
+
+if __name__ == "__main__":
+    main()