Expose images through proxy for specific UNIX timestamps

[EliSchleifer]

Is your feature request related to a problem? Please describe.

I am looking to integrate Protect images with a LLM to scan for packages. I want to see images from 30 seconds after and 30 seconds before an event.

I don't think that the current proxy allows that since it is limited to video. Would be great to provide direct access something like

/unifiprotect/thumbnail/{nvrid}/{cameraId}/timestamp/{UNIX_TS}

Describe alternatives you've considered

Alternative is to use the IP address directly for anonymous snapshotting which is less secure and more annoying.

Additional context

No response

Code of Conduct

• I agree to follow this project's Code of Conduct

Are you willing to resolve this issue by submitting a Pull Request?

• Yes, I have the time, and I know how to start.
• Yes, I have the time, but I don't know how to start. I would need guidance.
• No, I don't have the time, although I believe I could do it if I had the time...
• No, I don't have the time and I wouldn't even know how to start.

[RaHehl]

There is already a way to retrieve snapshots at a specific timestamp:
https://github.com/uilibs/uiprotect/blob/main/src/uiprotect/data/devices.py#L2006

For your use case, I’d personally recommend downloading the video clips and then extracting snapshots using ffmpeg. Since this method offloads the snapshot generation from the UniFi hardware to a separate machine, it’s likely to generate less load on the UniFi system.

Example using ffmpeg to extract one frame per second:

ffmpeg -i input_video.mp4 -vf fps=1 snapshot_%03d.jpg

This command will produce snapshot_001.jpg, snapshot_002.jpg, and so forth, for each second of the video. You can adjust the fps value as needed.

[RaHehl]

Ah, I think I misunderstood you. You don’t want multiple screenshots from 30 seconds; you just need two in total. In that case, I would simply use the existing method in the camera object.

[EliSchleifer]

is get snapshot exposed by the http proxy like the event and video feed /start/end endpoints? Maybe that is what’s missing or needs documentation I did resort to pulling the video and pulling frames but it feels bulkier .

…

________________________________ From: Raphael Hehl ***@***.***> Sent: Thursday, December 12, 2024 11:25:01 AM To: uilibs/uiprotect ***@***.***> Cc: Eli Schleifer ***@***.***>; Author ***@***.***> Subject: Re: [uilibs/uiprotect] Expose images through proxy for specific UNIX timestamps (Issue #321) Ah, I think I misunderstood you. You don’t want multiple screenshots from 30 seconds; you just need two in total. In that case, I would simply use the existing method in the camera object. — Reply to this email directly, view it on GitHub<#321 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AAJVCPTVVMN4ZM6VU7BIMID2FHPI3AVCNFSM6AAAAABTKBS7M2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDKMZZHAZTMNBUGI>. You are receiving this because you authored the thread.Message ID: ***@***.***>

[RaHehl]

I think I understand what you are aiming for, but I don’t fully grasp the reasoning behind it yet. From what I see, the context here is a standalone Python library, not something specifically tailored for Home Assistant. This library doesn’t include a web server, which makes the idea less clear to me. The proxy implementations you’re referring to appear to be in Home Assistant, as shown in the link: https://github.com/home-assistant/core/blob/dev/homeassistant/components/unifiprotect/views.py#L67.

Are you aiming to connect this to an LLM within Home Assistant? That would make more sense to me, and in that case, I believe a feature request would be better placed there for the unifiprotect component, within the Home Assistant project.

As for alternative approaches, you might consider solutions like the AI Port on UniFi Protect https://techspecs.ui.com/unifi/cameras-nvrs/up-ai-port or using Frigate https://frigate.video/ with a Google Coral

[EliSchleifer]

I see. I can aim for a fix in the Unifi hassio integration. Didn’t realize where in the stack proxy was implemented: since a method is already exposed by python code. Should be easy enough i actually use frigate and it is insufficient for these purposes. Thx

…

________________________________ From: Raphael Hehl ***@***.***> Sent: Friday, December 13, 2024 1:31:43 AM To: uilibs/uiprotect ***@***.***> Cc: Eli Schleifer ***@***.***>; Author ***@***.***> Subject: Re: [uilibs/uiprotect] Expose images through proxy for specific UNIX timestamps (Issue #321) I think I understand what you are aiming for, but I don’t fully grasp the reasoning behind it yet. From what I see, the context here is a standalone Python library, not something specifically tailored for Home Assistant. This library doesn’t include a web server, which makes the idea less clear to me. The proxy implementations you’re referring to appear to be in Home Assistant, as shown in the link: Home Assistant Proxy Code. Are you aiming to connect this to an LLM (Large Language Model) within Home Assistant? That would make more sense to me, and in that case, I believe a feature request would be better placed there for the unifiprotect component, within the Home Assistant project. As for alternative approaches, you might consider solutions like the AI Port on UniFi Protect https://techspecs.ui.com/unifi/cameras-nvrs/up-ai-port or using Frigate https://frigate.video/ with a Google Coral — Reply to this email directly, view it on GitHub<#321 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AAJVCPXZQ57HCNV7JZSKGRT2FJ5M7AVCNFSM6AAAAABTKBS7M2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDKNBQGY2TCNJUGQ>. You are receiving this because you authored the thread.Message ID: ***@***.***>