Add VAM (climbing velocity) metric and per-duration curve

Extract pipeline now computes two VAM metrics per activity (cycling,
running, hiking, walking):
- climbing_vam_mh: VAM on ascending segments only, using 30 s forward
  lookahead to classify climbing vs. flat/descent (stored in detail JSON)
- vam_curve: [[duration_s, vam_mh], ...] best VAM per standard duration
  (60 s – 1 h), sliding window on 30 s smoothed elevation, only windows
  with ≥ 10 m net gain count (stored in summary + detail)

Athlete JSON aggregates vam_curve across all activities (all_time,
last_365d, last_90d), same structure as power_curve.

Frontend:
- ActivityDetail shows "Climbing VAM" stat (grouped with elevation)
- AthleteView adds a "VAM Curve" tab that appears only when the athlete
  has climbing data; renders VamChart (new component, mirrors MmpChart)

vam_curve stripped from combined global feed; kept in user year shards
for season-based on-the-fly aggregation in VamChart.

Requires bincio reextract to backfill existing activities.

bincio/extract/writer.py

@@ -101,6 +101,8 @@ def write_activity(
         "mmp": metrics.mmp,
         "best_efforts": metrics.best_efforts,
         "best_climb_m": metrics.best_climb_m,
+        "climbing_vam_mh": metrics.climbing_vam_mh,
+        "vam_curve": metrics.vam_curve,
         "laps": [_serialise_lap(lap) for lap in activity.laps],
         "timeseries_url": f"activities/{activity_id}.timeseries.json" if timeseries else None,
         "source": source,
@@ -257,6 +259,7 @@ def build_summary(
         "mmp": metrics.mmp,
         "best_efforts": metrics.best_efforts,
         "best_climb_m": metrics.best_climb_m,
+        "vam_curve": metrics.vam_curve,
         "source": _infer_source(activity),
         "privacy": privacy,
         "detail_url": f"activities/{activity_id}.json",
@@ -300,6 +303,25 @@ def write_athlete_json(summaries: list[dict], output_dir: Path, athlete_config:
     mmps_365 = [s["mmp"] for s in summaries if s.get("mmp") and _is_outdoor(s) and s["started_at"] >= cutoff_365]
     mmps_90  = [s["mmp"] for s in summaries if s.get("mmp") and _is_outdoor(s) and s["started_at"] >= cutoff_90]
 
+    # ── VAM curve aggregation ─────────────────────────────────────────────────
+
+    def _merge_vam_curves(vam_lists: list[list[list[int]]]) -> list[list[int]]:
+        best: dict[int, int] = {}
+        for vc in vam_lists:
+            for d, v in vc:
+                if d not in best or v > best[d]:
+                    best[d] = v
+        return [[d, v] for d, v in sorted(best.items())]
+
+    _VAM_SPORTS = {"cycling", "running", "hiking", "walking"}
+
+    def _has_vam(s: dict) -> bool:
+        return bool(s.get("vam_curve")) and s.get("sport") in _VAM_SPORTS and _is_outdoor(s)
+
+    all_vams = [s["vam_curve"] for s in summaries if _has_vam(s)]
+    vams_365  = [s["vam_curve"] for s in summaries if _has_vam(s) and s["started_at"] >= cutoff_365]
+    vams_90   = [s["vam_curve"] for s in summaries if _has_vam(s) and s["started_at"] >= cutoff_90]
+
     # ── Personal records aggregation ──────────────────────────────────────────
     # records[sport][distance_km] = {time_s, activity_id, started_at, title}
     # best_climb[activity_id] = {climb_m, started_at, title}
@@ -368,6 +390,11 @@ def write_athlete_json(summaries: list[dict], output_dir: Path, athlete_config:
             "last_365d": _merge_mmps(mmps_365) if mmps_365 else None,
             "last_90d":  _merge_mmps(mmps_90)  if mmps_90  else None,
         },
+        "vam_curve": {
+            "all_time":  _merge_vam_curves(all_vams) if all_vams else None,
+            "last_365d": _merge_vam_curves(vams_365)  if vams_365  else None,
+            "last_90d":  _merge_vam_curves(vams_90)   if vams_90   else None,
+        },
         "records": {
             sport: _serialise_sport_records(records[sport])
             for sport in SPORTS