feat(nav,stats): Events sidebar entry + 'Overall Peak' excludes false triggers

Two related operator-facing improvements after the nav reorg.

1) Events as a top-level sidebar entry.

The /sfm page (fleet-wide event database) was demoted to Settings →
Developer in the previous reorg.  Bringing it back to main nav as
"Events" — operators do reach for the cross-project, sortable
event list, so it earns a top-level slot.

Sidebar now (7 items):
  Dashboard · Devices · Projects · Events · Tools · Job Planner · Settings

Settings → Developer card pointing at /sfm is removed.  /sfm page
title/subtitle updated from "SFM Event Data" to just "Events".  URL
unchanged.

2) "Peak PVS" KPI tile becomes "Overall Peak" and excludes false
   triggers from the calculation.

When operators ask "what's the biggest event at this location/unit/
project?" they mean the biggest REAL event, not the biggest sensor
glitch.  A single mis-flagged false trigger could otherwise dominate
the tile (the 14.13 in/s spike at Loc 1 was a prime example).

backend/services/sfm_events.py:
- _compute_stats() skips false_trigger=True events when computing
  peak_pvs / peak_pvs_at / peak_pvs_serial.  Continues counting them
  in false_trigger_count so the separate "False Triggers" tile still
  reflects what got filtered out.  last_event unchanged (recency, not
  magnitude).
- Same change automatically propagates to events_for_unit() and
  vibration_summary_for_project() — both call _compute_stats().

Templates: "Peak PVS" → "Overall Peak" in 3 KPI tile locations
(vibration_location_detail.html, partials/projects/vibration_summary
.html, unit_detail.html).  The physical-quantity name "Peak Vector
Sum" in the event-detail modal stays — that's the actual physics
term, not a summary stat.

Verified end-to-end: Overall Peak renders on real data; peak event
false_trigger flag confirmed False.

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

backend/services/sfm_events.py

@@ -548,7 +548,14 @@ def _empty_stats() -> dict:
 
 
 def _compute_stats(events: list[dict]) -> dict:
-    """Roll up summary stats from a merged event list.  Cheap O(N) pass."""
+    """Roll up summary stats from a merged event list.  Cheap O(N) pass.
+
+    The "Overall Peak" stat (peak_pvs) EXCLUDES events flagged as false
+    triggers — operators care about the highest REAL event, not the
+    biggest sensor glitch.  false_trigger_count still includes them so
+    operators can see how many were filtered out.  last_event uses
+    every event regardless (it's about activity recency, not magnitude).
+    """
     if not events:
         return _empty_stats()
 
@@ -559,19 +566,22 @@ def _compute_stats(events: list[dict]) -> dict:
     false_trigger_count = 0
 
     for ev in events:
-        pvs = ev.get("peak_vector_sum")
-        if pvs is not None and (peak_pvs is None or pvs > peak_pvs):
-            peak_pvs = pvs
-            peak_pvs_at = ev.get("timestamp")
-            peak_pvs_serial = ev.get("serial")
+        is_false_trigger = bool(ev.get("false_trigger"))
+        if is_false_trigger:
+            false_trigger_count += 1
+
+        # Peak calculation: skip flagged false triggers.
+        if not is_false_trigger:
+            pvs = ev.get("peak_vector_sum")
+            if pvs is not None and (peak_pvs is None or pvs > peak_pvs):
+                peak_pvs = pvs
+                peak_pvs_at = ev.get("timestamp")
+                peak_pvs_serial = ev.get("serial")
 
         ts = ev.get("timestamp")
         if ts and (last_event is None or ts > last_event):
             last_event = ts
 
-        if ev.get("false_trigger"):
-            false_trigger_count += 1
-
     return {
         "event_count":         len(events),
         "peak_pvs":            peak_pvs,