A pre-release for field validation. HACS offers this only to users who have enabled beta versions for this repository. It is not a public release, and the change below still ships to everyone in the next monthly bundle as 3.43.0.
It supersedes v3.43.0-beta.11.
What changed since beta.11
-
New option: Camera notification cooldown (#703). A subject that lingers in view, or a camera whose motion sensor re-triggers, could produce several near-identical notifications within a minute, because the vision model words the same scene differently each time and caption deduplication compares wording. Set Camera notification cooldown (seconds) in Global Options (0 = off, up to 600; it sits just below Video Analyzer Notify Service) and the first notification from a camera sounds, while later ones from that camera within the window replace the same notification card quietly with the latest description and image. With face recognition on, the first Unknown Person reported inside the window still sounds, once. Each camera has its own fixed-length window; analysis, sensors, events and stored results are unchanged. Requested by @andymcmanus in #672.
Limits to know before testing: it needs a single companion-app notify service (
notify.mobile_app_*); without face recognition everything inside a window is a quiet update; Home Assistant does not report whether a push reached the phone, so quiet updates can follow an alert that was lost on the way; verified on iOS, not verified on Android. Full list: Notification cooldown. -
Recognized-people sensor and image could carry another analysis's names.
sensor.*_recognized_people,image.*_last_eventand the caption deduplication check read a per-camera "last recognized" slot after the summary model call, so when two analyses of one camera overlapped (a snapshot batch and an event recording, for example) they could be given the other batch's names. Each analysis now carries its own names.
Everything from beta.11 and earlier is included. See CHANGELOG.md under Unreleased.