starshipwinepineapple

I imagine it's like the Active Noise Canceling (ANC) mode of headphones which will monitor background noise and then calculate the inverse sound wave and play that sound to you which in effect mostly cancels out the background noise (works better for certain frequencies than others). For a microphone they could do the same but instead of playing you the inverse sound, they could play it through the microphone for whomever is listening.

My advice (from above)

• fork projects on whatever platform they are hosted and keep the forks on that platform.
• only put your own original code on codeberg.

So whatever code you are looking at, keep it on that platform if you are forking it.

If you have original code, then you can license it however you want and put it on codeberg if it's an OSI approved license.

For starters, read licensing from codeberg (including their licensing page that is linked from there. Granted youre not choosing the licenses but there is good info to learn there.

Now codeberg wants free licensing. A good rule of thumb is if it's an Open Source Initiative (OSI) license then it's good. For a list of OSI licenses you can check out their page.

There may be other free and open source licenses that people use but if you know nothing about licensing then stick with the OSI ones until you get your feet wet.

You'll also want to make sure that whatever project you're using isn't using dual licensing with non-free parts. If it looks like they have multiple licenses described in their README.md or LICENSE then as a beginner i would just run away as those can be complex and you may not be able to put that code on codeberg

HOWEVER, my advice as someone who uses codeberg as my main forge is this:

• fork projects on whatever platform they are hosted and keep the forks on that platform.
• only put your own original code on codeberg.

My rationale is that part of the open source mentality is being able to give back your code contributions. Putting your code on another forge will make it so you're not able to do that unless you later add a different git remote to your fork on that forge and then merge (im ignoring the rare unicorn projects that accept contributions via email). That isn't hard to do but it's extra steps. If you fork a github project on github then you can still have a local copy in your PC and if the original policy is removed/nuked/relicensed/etc then you could dig into the licensing and see if you can put your local copy on codeberg.

What are you trying to do, specifically? And why? AUR is a collection of third party package build scripts to install software. You don't download the entire thing on any system.

Navidrome exposes the folder structure via the subsonic api, and any compatible app can then display folders if they want. Both symfonium (mobile) and feishin (desktop) have folder support.

Are there any good apps for android for navidrome/jellyfin respectively?

Symfonium is paid but by far the best i tried. Will work with both via subsonic api.

how easy can I add songs to them/do they pull metadata from somewhere (like jellyfin does for movies)

I use navidrome and can't speak to jellyfin. Navidrome is intentionally read-only for security reasons so you add music to where navidrome can see it and it'll auto-scan them in. For metadata my recommendation is musicbrainz picard. A lot of people recommend beets and it can be good but it's a big learning curve. Musicbrainz picard has a nice gui and easier to get music matched for beginners.

how do they (or any other options) compare in terms of ease of setup/maintainability?

From what i see people use jellyfin because they don't want "yet another container" and most people already have jellyfin for other media content. So that would be easiest route. Navidrome isn't hard to set up though, and overall has a better feel for me.

do you have any overall recommendations?

Look into LRCGET to get the timesynced and/or plaintext lyrics of your songs. There's also a navidrome plugin that will automatically fetch the lyrics when a song is played

Edit: for desktop id recommend feishin. It'll work with both through subsonic api

I would say the bigger issue is needing more people willing to be package maintainers. Their docs do cover how to move packages from AUR to an official package but it does involve needing to be sponsored by existing package maintainers and doing extra work. In the end it didn't seem worth it for my package that meets the criteria to be /extra.

And with the AUR malware issue, if these AUR maintainers can't be bothered to maintain their packages and let them get orphaned, then how would you expect those maintainers to meet that higher standard of work?

Like any open source project, the limiting factor is the volunteers. And, well, I'm just not going to tell the arch linux volunteer package maintainers they need to do more when I haven't been willing to do the work they are doing

Might be your instance. Try blog and spot and .com without spaces. The blog is also linked within OPs article

Their april 15th blog post explicitly calls it a backdoor and mentions it was very well hidden. I'm interested to see what comes of this

Article really buries the lede here by not even mentioning mitigation for all 3 vulnerabilities:

Update UniFi Network application to Version 10.1.89 or later.

In your docker-compose.yaml you need to add in your directory if you haven't already

services:
  immich-server:
    volumes:
      # Do not edit the next line. If you want to change the media storage location on your system, edit the value of UPLOAD_LOCATION in the .env file
      - ${UPLOAD_LOCATION}:/usr/src/app/upload
      - /etc/localtime:/etc/localtime:ro
+     # Add NAS pictures and videos folders to use as External Libraries
+     - "/volume1/data/ Pictures:/usr/src/app/library/pictures"

Then in immich you need to add this as an external library. Click your profile icon -> administration -> external library -> create library. It will ask for a path and you need to use the library mount point within docker which was /usr/src/app/library/pictures from my example. Click add path

That should get you good to go

You're right to question this.

In machine learning Accuracy means the correct % of overall classifications. There's some other terms like:

• Precision which is the % of correctly identified positives divided by the number of positive classifications. A high precision score would mean that of everyone who flagged as a match you had relatively few who were not actual shoplifters.
• Recall (true positive rate) which is the % of correctly identified positives divided by all actual positives. A high recall score measures how many shoplifters you caught and would minimize false negatives, but at the cost of more false positives.

So in the case of classification of shoplifters ideally you would focus on Precision as false positives are undesired, but if a company doesn't care about false positives as much as getting the shoplifters they'd focus on Recall. In either event, Accuracy is a poor metric to use or advertise in an imbalanced data set like shoplifting as most customers are not shoplifters so even if the model didn't classify anyone as a shoplifter they'd still be 99+% accurate.