The death of a thousand cuts
-
...I should also add that the ArchiveVideo action works for individual snapshots (using Format = 1) with the files being stored in the images/ folder and available to browse from the Images page in the openLuup console.
On the openLuup console Sockets > SMTP page you should see the camera IP as a registered email sender, along with a count of the number of emails received. You obviously have to configure the camera to send these. On the POP3 page you can actually browse the messages sent as triggers, but the child security sensor should also behave as expected (if armed.) I have an old iPod touch set up to browse these internal emails via the POP3 server using the Mail app.
If you have lots of messages/images being sent, you can configure timed scene to use the openLuup plugin's SendToTrash and EmptyTrash actions to enforce file quotas based on filetype / age / number of files.
-
-
-
-
That looks great. When i tried to add new camera device. It added the two devices, but there is no un pw fields and the un pw action says no implementation. No joy.
-
-
No, my mistake, that's not the action you should use (that's supposed to change the camera password, I suppose, but it's not implemented here.) It's been a while.
The device files don't specify a default value for URL or DirectStreamingURL, so those don't exist. You can simply create them manually, either from Luup Test code or the +Create button on the camera's device variable page in openLuup.luup.variable_set ("urn:micasaverde-com:serviceId:Camera1", "URL", "whatever goes here", dev_no)
luup.variable_set ("urn:micasaverde-com:serviceId:Camera1", "DirectStreamingURL", "whatever goes here", dev_no)The contents are obviously camera-dependent. Username/password can, IIRC, be put either in the device's attribute fields
(which aren't created by default, either)or, as I do, directly in the URLs:- URL:
/cgi-bin/CGIProxy.fcgi?cmd=snapPicture2&usr=USR&pwd=PWD
- DirectStreamingURL:
/cgi-bin/CGIStream.cgi?cmd=GetMJStream&usr=USRr&pwd=PWD
Sorry, this is more manual than desirable.
I'll change the implementation file to create blank attributes and variables for the camera.
__
Edit: the latest development version of openLuup v20.5.6 does create these attributes and variables!
- URL:
-
-
Just a follow up on the CPU impact of this type of camera implementation. Here's the last 12 hours of CPU/memory usage on my RPi. I installed the camera device at around 10:45, about halfway through the graph.
You can see there's barely any difference, and, that which there is, it's simply due to running AltUI which requests the usual status updates from the engine (I don't usually run AltUI on this system.)
The video simply streams from the camera to the browser, and openLuup doesn't see the video stream. The only time it gets involved is in downloading an archive snapshot from the camera (or receiving email triggers with attached images.)
-
Success, thank you for your help. It looks really good.
-
While you’re at it can you make set attribute category_num to 6 so it’s recognized as a camera
TIA -
akbooerreplied to akbooer on May 6, 2020, 7:02 PM last edited by akbooer May 6, 2020, 3:03 PM
@akbooer said in The death of a thousand cuts:
The category system is rather flawed, so I tend to ignore it!
By way of example...
The category is traditionally defined by the device file. It turns out that
D_DigitalSecurityCamera1.xml
does not set it.D_DigitalSecurityCamera2.xml
, OTOH, does! -
Should I switch to D_DigitalSecurityCamera2?
-
...or just set the category manually. The extra controls offered by #2 are, IIRC, for steerable cameras. This isn’t implemented as it stands, so it makes no difference. If it’s an issue (to steer) it could be added.
-
@DesT said in The death of a thousand cuts:
@therealdb said in The death of a thousand cuts:
my own bot
Can you elaborate?
I wrote a bot for telegram. I already posted some pictures on the Vera forum:
Commands are in italian, but you'll get the point. First 3 are related to OpenSprinkler, then there’s a sensor connected to Vera for the entryway door, plus a short video (as GIF) from the camera. All the notifications are sent via Telegram, to me/my wife using position (since I used to work 50% of my time in a different part of the country before this COVID-19 mess, now I'm full remote for a couple of months).
The idea is to add frequently used manual actions here, and then get rid of the app
I'm now into forcing my wife to use it more, in order to add more features. For me, manual control of blinds in case of need is more than enough.
-
My dudes, I tried your HomeWave suggestion. I’m not having a great experience. It’s flaky and im having difficulty adding devices.
|-<:(
28/35