fix: per-source elevation params — strava_export vs barometric vs raw GPS

Previous thresholds (10 m GPS, 5 m barometric, 30 s MA) were calibrated for
raw noisy GPS. Strava-exported FIT files carry elevation already pre-processed
by Strava (smooth 1 m quantisation, no steps > 5 m), so the aggressive
filtering suppressed real climbing — avg −17 % error across 37 reference
activities.

New strategy, keyed on source + altitude_source:
  strava_export           → MA 5 s, threshold 1.0 m
  fit_file / barometric   → no MA, threshold 1.5 m
  fit_file / gps          → MA 5 s, threshold 2.0 m
  unknown non-strava      → MA 5 s, threshold 1.5 m

Result on 37 cross-referenced activities: avg −2.8 %, std 4.6 %,
37/37 within ±15 % (was 0/37).

Both paths — initial import (metrics._elevation) and bulk recalculate
(dem.recalculate_elevation_hysteresis) — now use the same elevation_params()
function from metrics.py.

bincio/extract/dem.py

@@ -19,6 +19,8 @@ import urllib.request
 from pathlib import Path
 from typing import Optional
 
+from bincio.extract.metrics import elevation_params
+
 # Sample one GPS point per N seconds when building the DEM query.
 # SRTM30 resolution is ~30 m; at 30 km/h cycling that's ~3 s per tile —
 # sampling every 10 s is more than enough.
@@ -346,10 +348,10 @@ def recalculate_elevation_hysteresis(user_dir: Path, activity_id: str) -> dict:
     if len(elevations) < 2:
         raise ValueError("Not enough elevation data to compute gain/loss")
 
-    # Determine source-aware threshold
     detail = json.loads(json_path.read_text(encoding="utf-8"))
     altitude_source = detail.get("altitude_source", "unknown")
-    threshold = 1.0 if altitude_source == "barometric" else 3.0
+    source = detail.get("source") or ""
+    ma_window, threshold = elevation_params(altitude_source, source)
 
     # Strip leading no-fix zeros (same logic as metrics._elevation)
     if elevations and abs(elevations[0]) < 0.5:
@@ -358,8 +360,7 @@ def recalculate_elevation_hysteresis(user_dir: Path, activity_id: str) -> dict:
                 elevations = elevations[i:]
                 break
 
-    # Pre-smooth to suppress noise, then accumulate with low dead-band
-    smoothed = _moving_average(elevations, _MA_WINDOW_S)
+    smoothed = _moving_average(elevations, ma_window) if ma_window > 1 else elevations
     gain, loss = _hysteresis_gain_loss(smoothed, threshold)
     gain_r = round(gain, 1)
     loss_r = round(loss, 1)
@@ -385,5 +386,7 @@ def recalculate_elevation_hysteresis(user_dir: Path, activity_id: str) -> dict:
         "elevation_gain_m": gain_r,
         "elevation_loss_m": loss_r,
         "threshold_m": threshold,
+        "ma_window_s": ma_window,
         "altitude_source": altitude_source,
+        "source": source,
     }