move_detect/edge/engine.py
유창욱 752afc3f00 feat(edge): heartbeat telemetry + monotonic timers + queue/camera robustness
- Split clocks: detector timers/cooldowns run on a MONOTONIC clock (immune to an
  NTP step on an RTC-less Pi); wall time is used only for the HMAC X-Timestamp
  and Event.occurred_at. Fixes false IMMOBILITY / dropped tracks after a clock jump.
- Heartbeat is time-based (30s, predictable regardless of load) and now carries
  temp/throttle/queue_pending/config_version telemetry.
- A dead (no-frame) camera counts as erroring instead of a healthy frame, so a
  silent monitoring outage no longer looks green.
- Offline queue: isolate one undecodable row instead of wedging every event behind
  it; classify 401 (stale clock) as transient, never a poison-drop.
- New edge/health.py: vcgencmd temp/throttle parsing, None off-Pi (no fake 0).
2026-07-25 19:20:23 +09:00

100 lines
5 KiB
Python

from dataclasses import dataclass
from edge.tracking import Track
from edge.events import Event, EventType, severity_for
from edge.detectors.fall import FallDetector
from edge.detectors.immobility import ImmobilityWatchdog
from edge.detectors.ghost import GhostIncidentDetector
from edge.detectors.zone import ZoneIntrusionDetector
@dataclass
class EngineConfig:
camera_id: str = "cam"
cooldown_s: float = 30.0
class EventEngine:
def __init__(self, fall: FallDetector, immobility: ImmobilityWatchdog,
ghost: GhostIncidentDetector, zone: ZoneIntrusionDetector,
cfg: EngineConfig):
self.fall = fall
self.immobility = immobility
self.ghost = ghost
self.zone = zone
self.cfg = cfg
self._last_emit: dict[tuple, float] = {}
def _allow(self, key: tuple, now: float) -> bool:
last = self._last_emit.get(key)
if last is not None and now - last < self.cfg.cooldown_s:
return False
self._last_emit[key] = now
return True
def forget(self, track_id: int) -> None:
"""트랙 소실 시 하위 감지기 + dedup 캐시의 트랙별 상태를 정리 (24/7 장치 메모리 누수 방지)."""
self.fall.forget(track_id)
self.immobility.forget(track_id)
self.zone.forget(track_id)
self._last_emit = {k: v for k, v in self._last_emit.items()
if not (len(k) >= 2 and k[1] == track_id)}
def _mk(self, etype: EventType, occurred_at: float, track_id=None, zone_id=None,
metrics=None, spatial=None) -> Event:
m = metrics or {}
duration_s = m.get("immobility_s") or m.get("waited_s") or 0.0
return Event(etype, severity_for(etype), self.cfg.camera_id, track_id,
zone_id, occurred_at, duration_s, m, spatial)
def process(self, active_tracks: list[Track], lost_tracks: list[Track],
now: float, frame_wh=None, wall_now: float | None = None) -> list[Event]:
# `now` is a MONOTONIC clock (never jumps): it drives every detector timer, hysteresis,
# and cooldown, so an NTP step cannot instantly satisfy immobile_s / expire all tracks /
# reset a genuine immobility countdown. `wall_now` is real wall-clock time, used ONLY to
# stamp Event.occurred_at (the server correlates and retains by wall time). Defaults to
# `now` so single-clock callers (unit tests) keep their existing occurred_at behavior.
occurred_at = wall_now if wall_now is not None else now
events: list[Event] = []
self.ghost.on_lost(lost_tracks, now)
for tr in active_tracks:
# frame_wh lets the zone detector normalize the foot to the dashboard's 0..1 zone
# space (see ZoneIntrusionDetector._foot). run_once supplies the live frame size.
zone_hits = self.zone.update(tr, now, frame_wh)
in_zone = bool(self.zone.active_zone_ids(tr.track_id))
sig, meta = self.fall.update(tr, now)
state = self.fall._states.get(tr.track_id)
self.ghost.note_state(tr.track_id, state.name if state else "STANDING")
if sig == "fall" and self._allow((EventType.FALL, tr.track_id), now):
events.append(self._mk(EventType.FALL, occurred_at, tr.track_id, metrics=meta))
elif sig == "no_motion" and self._allow((EventType.NO_MOTION, tr.track_id), now):
events.append(self._mk(EventType.NO_MOTION, occurred_at, tr.track_id, metrics=meta))
fired, imeta = self.immobility.update(tr, now, in_danger_zone=in_zone)
if fired and self._allow((EventType.IMMOBILITY, tr.track_id), now):
events.append(self._mk(EventType.IMMOBILITY, occurred_at, tr.track_id, metrics=imeta))
for hit in zone_hits:
# 협착(pinch) is the highest-danger zone (machinery entrapment) and must emit
# PINCH/HIGH — the old `restricted else WARNING` silently downgraded it. The
# dashboard offers exactly these three levels (see zones.php).
etype = {"pinch": EventType.PINCH,
"restricted": EventType.ZONE_RESTRICTED}.get(
hit["level"], EventType.ZONE_WARNING)
if self._allow((etype, tr.track_id, hit["zone_id"]), now):
events.append(self._mk(etype, occurred_at, tr.track_id, zone_id=hit["zone_id"]))
for g in self.ghost.update(active_tracks, now):
key = (EventType.GHOST, g["track_id"])
if self._allow(key, now):
events.append(self._mk(EventType.GHOST, occurred_at, track_id=g["track_id"],
spatial=g["spatial"],
metrics={"last_state": g["last_state"],
"waited_s": g["waited_s"]}))
# ghost.on_lost already consumed lost tracks' state above, so forgetting here is safe.
for tr in lost_tracks:
self.forget(tr.track_id)
return events