Incorrect implementation of event detection for Nest cameras

[jrluw]

Event detection for Nestcams (as currently implemented) doesn't seem to work, at least in conjunction with homeassistant. Is this because the requirements for "is_ongoing" are too stringent? (Maybe the Nest API is very precise about events, and an event is only "ongoing" for a few milliseconds.) If that's the case, should this be fixed so that "ongoing" is more lenient? (E.g., cache the old last_event, check whether it differs from the new last_event. If so, and the time between the two is small enough, is_ongoing returns true.)

[jrluw]

Would be great if you can say whether it's supposed to work. If not, I can try to fix it.

[jkoelker]

I'm not sure, I don't use nestcams at all as I don't have the hardware. That said if its not working for whatever reason, and someone patches it, I'd accept the patches.

[ruthless0ne]

@jrluw did you ever get anywhere with this?

[awarecan]

@jrluw can you try python-nest 4.0.1 and home-assistant 0.71.0b0 (rc branch)

[jonas73x]

this still doesn't seem to work for me at least. I have a Nest Cam IQ indoor and
print (napi.structures[0].cameras[0].person_detected)
print (napi.structures[0].cameras[0].motion_detected)
always returns False, despite me standing in front of the camera and the Nest app sending me "person detected notifications". Is this expected to work now? And, thanks for an awesome python module. love it.

[awarecan]

I have heard some other users said people detection not working when familiar face enabled on IQ and Hello. But motion detection should work.

The only camera I have is Nest Indoor, which is working fine.

Can you post the raw HTTP response in here?

[jonas73x]

thank your for your quick reply. unfortunately i don't think i know how to get the raw HTTP response. tips very welcome otherwise i'll try to figure it out.

[jonas73x]

not sure if this is what you are looking for but i did try to run some CURL calls and got this back:
event: put
data: {"path":"/devices/cameras/WHAT_LOOKS_TO_BE_A_TOKEN/last_event/has_person","data":true}

[awarecan]

So, there is no begin_time or end_time in the JSON. In that case we won't return person_detected.

Can you check

print(napi.structures[0].cameras[0].last_event.has_person)

See if it can back to False after you trigger it to True, or how quick it does.

[jonas73x]

print(napi.structures[(napi.structu 0].cameras[0].last_event.has_person) turns to True as it's supposed to and seems to turn back to False about 45 seconds after the person has moved out of the camera view. And also assuming no other motion/event is detected. I can get the start end time with
https://developer-api.nest.com/devices/cameras/device_id/last_event/start_time
and
https://developer-api.nest.com/devices/cameras/device_id/last_event/end_time

[jonas73x]

i think the problem is the time zone. checking both the start_time/end_time urls above as well as
napi.structures[0].cameras[0].last_event.start_time
napi.structures[0].cameras[0].last_event.end_time
they are in GMT time-zone while I'm actually in GMT+2.
assuming now = datetime.datetime.now(self.end_time.tzinfo) correctly picks up GMT +2 which then screws up the is_ongoing comparison. my address is correctly set in my account and the Nest app timeline shows the correct time so this seems to be an API issue. any suggestions/ideas how to fix it would be greatly appreciated.

[awarecan]

Maybe https://developer-api.nest.com/structures/structure_id/time_zone

[jonas73x]

the strange thing is if i curl that URL i get:
event: put
data: {"path":"/structures/XXXX/time_zone","data":"Europe/Stockholm"}
which is correct, yet the times are wrong by the 2 hours.