Request to add pdlib to PECL.

[apoorv569]

I am using Nextcloud-AIO and want to install the Face Recognition app which requires this extension as a dependency. Please add this to PECL so me and anybody else can use this.

I got recommended to request this here, https://github.com/nextcloud/all-in-one/discussions/1551

[Csantos88]

+1 Vote for add pdlib in PECL PHP Repos

[goodspb]

@matiasdelellis really grand that you could help with this. thanks.

I have written to PECL but never got feedback. I don't know why and don't know how to proceed. >.<

[boboche]

+1, its 2023, there's no open source DAM with SIMPLE install, automatic AI object tagging / classification that simply installs and work out of the box without having to fight with some docker-fu or configuration headaches. Nextcloud is a super nice package and has everything in place to go next level.

But what should be a simple add-on from the app store + some simple post-config (i.e. select which model to download/apply + add. cfg, so that it doesn't make a huge VM or initial download for people who don't need it in the first place) is a in-shell php/snap/scripting snafu. Great and "normal" for coders, unfortunately not for the intended target audience tho.

[szaimen]

Hi, I just wanted to mention that AiO has this now: https://github.com/nextcloud/all-in-one/tree/main/community-containers#how-to-add-containers. So the community could potentially add the facerecognition container as additional container there. Feel free to ping me if you should need help on this!

[PrivatHomeServer]

Hello Simon, I have watched your activities regarding the community containers on Friday. Looks very promising. I'm not the absolute IT expert - so I went the way of @sunnyd24 and finished part 1 with the memory allocation. Now encouraged by your post @szaimen I wanted to look at what we need to do to get community docker for "facerecognition". For the other json files, I looked once. "image ".../..." --> is missing in this case since we always run a build (docker build -t facerecognition https://github.com/matiasdelellis/facerecognition-external-model.git)

What do we need to do to make it a community docker here now? For me still the only issue with Nextcloud AOI, because the "regognize" solution is not usable for faces.

I would like to see a way here.

[szaimen]

"image ".../..." --> is missing in this case since we always run a build (docker build -t facerecognition https://github.com/matiasdelellis/facerecognition-external-model.git)

What do we need to do to make it a community docker here now?

Indeed someone would need to create an image out of it and publish it on docker hub.

For that you might have a look at https://github.com/szaimen/aio-fail2ban/blob/main/.github/workflows/docker-build.yml and submit such a workflow at https://github.com/matiasdelellis/facerecognition-external-model.git

---

Regarding the json, I just leave this here as an example:

{
    "aio_services_v1": [
        {
            "container_name": "nextcloud-aio-facerecognition",
            "display_name": "Computing container for facerecognition",
            "documentation": "https://github.com/nextcloud/all-in-one/tree/main/community-containers/facerecognition",
            "image": "matiasdelellis/aio-facerecognition",
            "image_tag": "v1",
            "internal_port": "5000",
            "restart": "unless-stopped",
            "environment": [
                "TZ=%TIMEZONE%",
                "API_KEY=some-super-secret-api-key"
            ],
            "aio_variables": [
                "nextcloud_memory_limit=4096M"
            ],
            "nextcloud_exec_commands": [
                "php /var/www/html/occ app:install facerecognition",
                "php /var/www/html/occ config:system:set facerecognition.external_model_url --value nextcloud-aio-facerecognition:5000",
                "php /var/www/html/occ config:system:set facerecognition.external_model_api_key --value some-super-secret-api-key",
                "php /var/www/html/occ face:setup -m 5",
                "php /var/www/html/occ face:setup -M 4G",
                "php /var/www/html/occ occ face:background_job &"
            ]
        }
    ]
}

[PrivatHomeServer]

@szaimen thank you very much for your comments. Superficially I have understood the topic. However, I am too little IT-person to be able to really implement it. Unfortunately.

From my point of view, this add-on is essential for the photo application to stand out against solutions like immich.

[benji0522]

wanted

[marcselis]

+1

[Sven286]

Because it all took more than a year to even have a glimph of face recognition feature. In the mean time a lot of development happened at Immich. I switched back to there. Good luck with nextcloud. /unsubscribe

[elvismercado]

+1

[gabglus]

+1

[otola]

+1

[nbluemer]

I also need pdlib in order to run face recognition on Nextcloud AIO.

[rastographics]

Because it all took more than a year to even have a glimph of face recognition feature. In the mean time a lot of development happened at Immich. I switched back to there. Good luck with nextcloud. /unsubscribe

After following this issue for a year, this is exactly the same reply I came here to make before unsubscribing from this issue.

Immich is amazing. You can point it at your nextcloud photos directory too, and just use it for read-only browsing of your files while still using nextcloud to manage the file organization and syncing.

[szaimen]

Actually there is some progress already in https://github.com/matiasdelellis/facerecognition-external-model/pull/7 but I am waiting for @matiasdelellis to review the PR.

[szaimen]

After https://github.com/matiasdelellis/facerecognition-external-model/pull/7 is merged, mainly only https://github.com/nextcloud/all-in-one/pull/3999 is missing now :)

[szaimen]

mainly only nextcloud/all-in-one#3999 is missing now :)

This is now released with v7.10.0 Beta. Testing and feedback is welcome! See https://github.com/nextcloud/all-in-one#how-to-switch-the-channel and https://github.com/nextcloud/all-in-one/tree/main/community-containers/facerecognition

[luzfcb]

@matiasdelellis @szaimen I started testing v7.10.0 Beta. Is there any way to improve performance of facerecognition-external-model? For me, it analyzes 1 image every 16 seconds. My server runs on Ryzen 5600G, 32 GB of RAM, and Nvidia GTX 1650, so I suppose it's more than enough to have a very good performance.

Well, this comment is completely offtopic of this issue. Is there somewhere we could continue discussions?

[szaimen]

Well, this comment is completely offtopic of this issue. Is there somewhere we could continue discussions?

I suppose here? https://github.com/matiasdelellis/facerecognition/discussions

[matiasdelellis]

Hi @luzfcb

My server runs on Ryzen 5600G, 32 GB of RAM, and Nvidia GTX 1650, so I suppose it's more than enough to have a very good performance.

Mainly it's great that you have an Nvidia GTX 1650 😬 , but the docker container is not compiled to use cuda. You should install CUDA on your host, and build the container to make use of it. 😉 It will be much faster. 😬 It shouldn't be complicated, but I can't help you because I never had a video card. 😞

On the other hand, you can reduce the size of temporary images on settings. By default this container is in 4Mpx, and with 2Mps you will get it will be 50% faster and you will get practically the same result.

I left 4mpx as the default because the larger the images, the more likely you are to find all your faces, but in practice, you don't get more than 10% new faces.

Finally no matter how long the process takes, it is designed to do the job progressively. We try not to take shortcuts but do better.

p.s: Not even Google gives you the results so quickly when you upload many images... 😅 Let's be pacient.. 😉

[szaimen]

Lets continue here: https://github.com/matiasdelellis/facerecognition/discussions/716