Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
  • Sep 22 04:24
    RUANHAOANDROID commented #136
  • Sep 22 04:20
    RUANHAOANDROID commented #136
  • Sep 21 17:06
    lastzero commented #142
  • Sep 21 12:19
    ripnetuk commented #142
  • Sep 21 12:15
    ripnetuk commented #142
  • Sep 21 12:01
    ripnetuk commented #142
  • Sep 20 20:50
    lastzero commented #142
  • Sep 20 20:37
    ripnetuk commented #142
  • Sep 20 20:35
    lastzero labeled #142
  • Sep 20 20:35
    lastzero assigned #142
  • Sep 20 20:35
    lastzero commented #142
  • Sep 20 15:52
    lastzero labeled #132
  • Sep 20 15:44
    lastzero commented #132
  • Sep 20 15:35
    lastzero labeled #132
  • Sep 20 08:54
    ripnetuk opened #142
  • Sep 20 00:50
    lastzero commented #129
  • Sep 20 00:35
    lastzero commented #133
  • Sep 19 22:47

    lastzero on master

    Make sure TakenAt and TakenAtLo… Merge branch 'develop' (compare)

  • Sep 19 22:46

    lastzero on develop

    Make sure TakenAt and TakenAtLo… (compare)

  • Sep 19 21:36

    lastzero on master

    Fix go.mod/sum after upgrade to… Implement time zone support for… Merge branch 'develop' (compare)

Michael Mayer
@lastzero
This is our new community chat for visitors and non-developers (we also use the #photoprism channel on gophers.slack.com). Feel free to ask questions, we'll reply as soon as possible!
Ildar Asanov
@frops
Hello. I run docker exec -ti photoprism photoprism index, but got error fatal error: runtime: out of memory
Michael Mayer
@lastzero
Does it work with less / other photos? Maybe there is something "wrong" with one of the files... the JPEG decoder of Go is a bit special sometimes.
Sebastian Kunert
@skunert
Hey, I would like to contribute something on the frontend side. Any suggestions what to do next?
Michael Mayer
@lastzero
We're going to refactor the frontend soon, see https://github.com/photoprism/photoprism/wiki/UI-Components
Once this is done, I think there are smaller tasks that can be done independently
Do you have any ideas / suggestions?
Sebastian Kunert
@skunert
I would have been interested into the location page, which is currently not implemented. But if a refactoring is in progress, I can wait.
Michael Mayer
@lastzero
Would you suggest using Google Maps or OpenStreetmap?
Sebastian Kunert
@skunert
Hmm I think for a proof of concept it would be enough to use the one that is most freely available. We could take a look at https://openlayers.org/. I came across it recently and apparently it offers support for differend providers like bing, openstreetmap and google maps.
Michael Mayer
@lastzero
Code footprint and performance will be important for many users, especially on mobile. Does https://leafletjs.com/ have a smaller size than OpenLayers?
Google Maps is probably complicated because of the licensing.
Note that tests can currently fail because of updated image classification labels... I will push a new docker image later today when I have proper internet (in the train right now).
Michael Mayer
@lastzero
‪Anyone knows a compact database of similar nouns and locations for implementing our search? If you search for fish, you should also find sharks. Same for NYC and New York.‬
Michael Mayer
@lastzero
‪Project update: We've improved image classification (now based on NASNet Mobile) and added auth to the built-in TiDB server. New demo will be released soon!‬
Michael Mayer
@lastzero
Pushed a new container image. You need to run "docker-compose build" and restart everything.
Tests are green again: https://travis-ci.org/photoprism/photoprism
Michael Mayer
@lastzero
We'll try to extract our search ontology from YAGO, a free semantic knowledge base. Anyone experience with OWL, YAGO and parsing?
Michael Mayer
@lastzero
@skunert Proof-of-conept for maps is included in our latest demo: https://demo.photoprism.org/places
Sebastian Kunert
@skunert
Looks really nice!
Michael Mayer
@lastzero
Thank you 🌈
A bit to high on iOS and Android, desktop looks good to me. Maybe something for a PR.
Michael Mayer
@lastzero
See also https://groups.google.com/a/photoprism.org/forum/#!topic/developers/9zoFpcGXc_E
Sebastian Kunert
@skunert
Sadly, I am quite busy at the moment. But when there is time, I am happy to contribute some more :)
Michael Mayer
@lastzero
Check out our Twitter account for project updates and screenshots: https://twitter.com/browseyourlife
twatzl
@twatzl

Hello everyone. I stumbled upon this project a few days ago and plan to contribute a bit, so I thought I might introduce myself a bit. I hope this is the right place for it.

Who am I? Software Developer from Austria. I have been developing software part time for the last three years and have started learning programming in 2011.

Why I want to contribute? I have a big photo collection (about 2TB and growing) and not found any suitable tool up until now. Photoprism even in it's early state seems like the perfect tool for my needs. Additionally I was thinking about implementing pretty much the same thing as photoprism is right now. Especially I think it would be fun having a tool which enables you to try out different "plugins" on your photo collection. e.g. automatic dust spot removal, tagging, etc.
Also I love go as a programming language.

What I plan to contribute? Field testing Photoprism on (parts of) my collection. Providing my knowledge for example about nextcloud integration or in form of test files, but I also want to actively contribute code if my time allows for that.

Michael Mayer
@lastzero
@twatzl Sorry for not getting back to you earlier... didn't see the notification. Did you have a chance to test PhotoPrism already? Tables and queries are not yet optimized for very large photo collections as we first implement the basics and optimize later. We're currently indexing 500 GB of photos collected in 19 years. Indexer is running since 1 or 2 days on a very small computer (NVIDIA Jetson Nano, see our twitter account). No errors so far!
twatzl
@twatzl

Well yeah I tested it when I first found the project, but I only tested it using a subset of 1GB of photos. Setup was super easy. I used docker and had it up and running in <5 mins.
Just wanted to run it again, but of course now it refuses to work for whatever reason. haha.

I am still quite busy with my bachelors thesis (zero shot tagging of images) as it turned out "just using" tensorflow models isn't as easy as I thought :D
but I hope that I have some time to contribute soon (tm)

Michael Mayer
@lastzero
@twatzl Note that the port changed to 2342 and the photo paths inside the container changed too, see https://raw.githubusercontent.com/photoprism/photoprism/master/docker/photoprism/docker-compose.yml
Using the same port in the container and on the host makes it easier to test and remember
Docs were outdated for a couple of days but nobody complained... glad I noticed it. Not much time to write docs at the moment, they certainly need improvements too.
(There are docs in Docker Hub too which I didn't check - might still be outdated.)
Michael Mayer
@lastzero
Docker Hub and Setup page in our Wiki / Developer Guide now updated too. Please report wrong / missing Information if you notice it.
twatzl
@twatzl
Using the compose file now I at leas can connect again. maybe there was something wrong with the docker image i was using. I have no idea. I didn't get any errors or anything, so i have no idea what went wrong. anyway now it works again (at leas i can reach the UI again)

I just found this news article today. since you are from germany you can hopefully understand german anyway.
Short summary: A study was conducted and found out that people take a lot of photos, but do neither back them up properly nor do they do anything with them.

https://ooe.orf.at/stories/3003954/

Michael Mayer
@lastzero
Thank you for sharing!
Develop branch is currently "broken" due to database configuration issues... Theresa added new tests last week... they work locally but break on CI... will be fixed ASAP
Michael Mayer
@lastzero
Tests are fixed and Go upgraded to 1.12.7 (you'll need to rebuild your Docker image/container) ✨
Test coverage is also up... 51% now. Theresa will add more tests this week.
Michael Mayer
@lastzero
The Jetson Nano indexed 80760 files / 471 GB in ~252.5h. That's ~11s per file. We're going to optimize that along with search performance.
Thore Weilbier
@weilbith
Hey there 👋
Just discovered Photoprism and I'm very curios. Do you use the IPTC Standard to tag images? 🤔
twatzl
@twatzl

@lastzero

The Jetson Nano indexed 80760 files / 471 GB in ~252.5h. That's ~11s per file. We're going to optimize that along with search performance.

Do you know exactly what part of indexing takes longest? Does it make a difference if it is a raw or a jpg file?

Michael Mayer
@lastzero
Sorry, we've been on vacation... Development continues in September, but I'll answer any questions in the meantime
We read and index EXIF tags... AFAIK that's the IPTC standard. Writing tags to image files is currently not possible. We want to use XMP sidecar files for that.
Michael Mayer
@lastzero
Converting RAW to JPEG takes long, creating thumbnails / resized images takes a bit and image classification... I'd say in that order. Note that the Jetson is pretty small / slow compared to a regular desktop computers.