Just a small update: The app can now open links directly to a detail page in either the Jolla Store or Storeman as well as opening arbitrary apps. I also added a bunch of text content and the backend is ready for contextual search (though not yet deployed and implemented in the app).
Note that the detail page opening might have to be patched out for the Jolla Store release because the rules don’t allow it (sigh).
I have a subdomain / part of my website dedicated to these kind of stuff:
It first tells you where to get the app stores, which once you should get and after that which apps can be used for what.
So if you are not into an app, you can still link to this for instance:
Feel free to link to any of them. I’ll put a link on the page to your new beginner’s app too.
Best regards
*Fuchur*
PS: Oh, and there is my YT channel in which I am covering a lot of SFOS stuff in this Playlist for newcomers (most videos pretty lengthy, but well that is my style):
Another small update: Search now works! It uses an open source AI embedding model allowing to query with natural language.
Which one? And what size? EDIT, english, only. Right?
It’s intfloat/multilingual-e5-large, size is 1024 and it supports multiple languages. The list of languages can be found here, though low resource languages will probably suffer when it comes to quality.
Is that a local AI or a server based one which will then call home and show the user who very likely wanted to make sure her/his data is not sent out anymore with the first app of SFOS s/he uses, that his/her data is given to an AI?
looks like a roberta, so could be run locally with something like llama.cpp, I guess?
EDIT: too big for a phone 
Fabrice Bellard has translation in TextSynth Server but even there the smallest model is 1.6GB m2m100_1_2B_q8
I’ve used this for personal translation projects, but running on a server I control. That could be embedded, but it’s too much data.
Server, the whole backend runs on my server. You can check the source code to see what “phones home”.
I don’t log anything outside errors, I literally have no idea what’s part of the request.
Just a FYI, I plan on making it disableable anyway.
SfosBeginnerGuideBackend/embeddings at master · RikudouSage/SfosBeginnerGuideBackend · GitHub correct?
If you are hosting this, I highly recomend using ts_server from bellard instead of using pytorch. it is much, much faster. the model I mention above is 100 languanges 8bit quantized … EDIT: if you want to try it out, I can spin up an instance.
I’d disable it and make it an option in the settings of the app if you really want to use this.
I would prefer a non-ai-search just based on meta keywords or content if you need that at all for a couple of informations like the once you will be putting into the app anyway.
Best regards
*Fuchur*
BTW. I see here an Bitwarden alternative in the list. Are there any alternatives for Keepass compatible password managers?
Continuing the discussion from Improving Ecosystem Visibility: The Case for a Web-Accessible Jolla Store:
Nice work! Is there a rendered web version available?
There are quite a few resources (wikis, guides) for users, so I’d like to link yours in the community-driven Sailfish OS Wiki: https://sailfishos.wiki/
Not yet there isn’t, I’m planning to create one, but the app takes priority. I’m also thinking of creating an Android app so that people currently on Android can read it as well.
I’ll take a look, thanks!
Use ownKeepass from Chum instead. That one is maintained!
How amazing that you got your answer already before I even noticed your question
pip pykeepass works fine, https://pypi.org/project/pykeepass/
before setting up venv install python3-lxml package and create venv with --system-site-packages else you need to also build lxml in venv which you dont want
https://sailfishos-chum.github.io/ does have recently updated in the right pane.