rumba

Pardon the mess, this turned out to be quite a lot of text.

No vector DB, no paid API, no full articles. Plain Python pulling free RSS feeds, plus Ollama running on my own 2070 Super. RSS already gives you the headline and a short blurb, so there are never megabytes to sift through. Every Ollama call is one story at a time, so the context window isn’t an issue. Llama3.1:8b for summaries and scoring. Gemma4:e4b for clickbate work, nomic-embed-text for headline duplicate vectors.

Find duplicates:

Python: strip each headline down to keywords (drop words like “the”, “says”, “update”) and only compare headlines that share an uncommon word. Two headlines match if they share enough keywords.

Python: sends batches of 128 headlines to Ollama

Ollama: turns each headline into a vector using nomic-embed-text

Python: compares the vectors, and anything over the config threshold counts as the same story. The vectors just live in memory for the run, no database needed. If Ollama is down, the keyword match still works on its own.

Regen short summaries:

Ollama: Generate a 2 sentence summary from the feed text if available

Get rid of clickbait:

Ollama: Detect clickbait from the headline only. Classify this news headline. “informative”: it states the news itself (who did what, what happened, what was found). Dramatic or emotional wording is fine if the fact is stated. “teaser”: it withholds the key fact to make you click: e.g. “The Surprising Reason…”, “The Class That…”, “Why X…”, “How X became…”, “What we know about…”, “Is X about to…?”, “Then things got weird.”

Unwrap clickbait and rewrite the headline if possible, otherwise drop the story

Ollama: A clickbait headline hides a key fact. Question it leaves open: {withheld}

Does the text below explicitly answer that exact question? Mentioning the topi talking around it, or answering a different question does not count.

“answer”: the answer quoted or closely paraphrased from the text, or “” if the text doesn’t give it. “headline”: if answered, a plain, factual replacement headline (at most 90 cha states that answer, using only facts stated in the text. Keep the text’s certainty: …

Respond only with JSON: {“answer”: “”, “headline”: “”}

Headline: {title} Text: {text}

Score each headline

Ollama: return a summary and an importance score in the same call:

SCALE = (“1-3 niche, local or trivial; 4-6 notable; 7-8 major national or inte “9-10 historic events or emergencies affecting millions right now.”)

Python: rank = importance + a bonus for more outlets carrying it - points for age

Don’t put US articles in World unless they have non-US countries involved

Ollama: Which countries is this news story about? List the country where it takes place and every other country that is a real party to it (acting, targeted, negotiating, fighting, trading, sanctioned). Use common English country names, e.g. “United States”, “United Kingdom”. Ignore countries only mentioned in passing or as background. Always name at least one country: if the text doesn’t say wh give the country where it most likely takes place.

Respond only with JSON: {“countries”: [”…"]}

Headline: {title} Text: {text}

Python: if the only country is the US, it’s left out of World

Everything gets cached per story, so each one only goes through Ollama once.

Ollama container, 2070 super, 3-7b, meta and qwen

I use it to ingest tons of news rss, find dupes, summarize, and then post that output to mesh a mesh network.

I also use it for a self hosted mealie to figure out recipe->ingredient.

I also use it for receipts. I translate pdf -> txt, feed the text to qwen, ask it who is the vendor, what is the price, what it a reasonable memo for the receipt, and give a dozen categories, which is best.

It’s an old-school Beanie Baby.

Someone crafted, it looked amazing, everyone wanted one, they probably started calling them good luck charms and sold a bunch more.

They used to make guns out of umbrella parts.

awesome, I just setup traefik / le on an old sort domain of mine and make everything inside support https.

Honestly, I was hoping to add client only certs on my jellyfin and open it to friends an family, but none of their streaming devices support it.

Just don’t get too close to TV’s or streaming boxes. Wildly different levels of hell

I keep one Physical Pihole and one in a container, be sure yeah, this happens.

I’ve made cotton candy with them. 60% of the flavor survived the heating.

OK Stegadon is pretty damned cool.

If that documentation was awful, I’d REALLY like to see your take on NixOS :)

Hey I drove to the library, picked up all these things you needed, got dinner here ya go, free!

You drove? man that’s lazy…

He used AI to clean up translation and save time after he spent a fuck ton of time curating and delivering us a helpful product. Calling him out as lazy is an awful take.

Yup, if there was ever a decent use for AI, this is it. Lemmy can (and will) hate the shit out of it, but it took a little burden off the shoulders of someone doing us a great service.

I can’t argue with the Ants, I love how they formed the lines on the graphs with their own bodies.

That’s fine for browser-based watching, literally no one in my group watches via the browser. Even on android it’d be a fight. Grandma’s not going to go on to a browser to auth her session.

The clients need to support it. If it were just backend, I’d fork it myself.

The authentication is lacking 2fa and has a half hearted attempt at fail2ban

If you try to properly implement either of those, the standard device clients won’t work anymore.

Plex provides default SSL.

The relay is actually a bit more useful.

You can be on a carrier grade NAT with no real external IP.

It’s more akin to running a VPS somewhere and SSH tunneling your home server through it.

They also cashe the entirety of the TVDB and EPG Services.

I’m not sore about most of this with jellyfin, and I am trying to primarily use it, but I really miss some of the features. But realistically, adding 2FA to the clients would be a huge benefit. trying to replace 2FA with wish.com fail2ban feels particularly dirty.

I can’t decide if you don’t know what you’re talking about or you’re just trying to troll me.

I’m sorry, but writing down the data from your organizational program and re-entering it all from scratch is NOT a backup solution.

If you have such scant data to do that, you didn’t need to have nextcould installed in the first place.

18% of people running next cloud are not backing it up.

Not OP, I’ve kinda had a middle of the road experience with it.

I run JF and Plex on the same shares.

I dropped 10k tracks on it and a bunch of audiobooks, my stuff is 100% tagged.

I use tailscale to get to the server because here’s no Nat Holepunching going on.

I try to use it as much possible for audio, but some days, I just give in and use plexamp (like a guilty pleasure)

cons:

• It has issues with displaying some of the songs, they’re tagged right but you just can’t find some of it. They’re all Discogs coded, so there’s not even a lot of extra characters.
• It doesn’t always remember where in a book I am,
• It has no idea about collections of book files.
• Search is very slow, (yes there is a plugin for this, yes it’s complicated enough I haven’t tried it yet)
• Scrolling a large list is stupid low, it should just stream everything text into ram and bring thumbs in on demand
• Finamp: Finamp is barely a wrapper for the JF engine to the point that they can’t implement effects or crossfade without the feature being added in JF first. But JF is just using a ready-to-go library to play music, so changes to JF require upstream library updates. Audio development feels stagnant.
• Finamp scrolling loads one letter at a time. Scroll to Z? you get to wait, A…B…C…D…E…F…G…H…I…J…K…L…M…N…O…P…Q…R…T…U…V…W…X…Y…Z, no skipsies. It literally takes me a couple of minutes to go to songs that start with Z.
• Plugin installs are complicated and poorly documented, and compatibility with versions is dicey
• Finamp: If you lose the network in the middle of a song, you can soft-lock the app.
• Finamp: occasionally crashes if left for a long play session on my late-model Android phone.
• No options to cast.
• No listening through a NAT without port forwarding (which is dicey without a security team)
• No 2FA
• Finamp?: Shuffle is too random, you can get the same song to play twice in a couple of minutes. it needs to pull at least a couple of hours of list and shuffle that, rather than random play.

pros:

• It’s free
• It works good enough-ish for a daily car ride.
• It has some form of limited home-grown fail2ban
• The developers are super nice people.
• I exported my Plex playlists and used some Python to turn them into m3u lists, which worked fine. (Would be a cool feature to import from Plex)
• Playlist and Shuffle work mostly fine.

There are a LOT of pros and cons.

Pros:

• Developed by a professional, multi-disciplined full-time team with some security oversight.
• Hosted caching of The Movie DB for faster lookups
• Provision of SSL communication to and from your server without any special setup
• FREE EPG data caching
• Centralized server management from the web
• Low-speed relay for those stuck behind CGNAT.
• A REALLY solid mobile media player (sorry, but plexamp beats the pants off the JF alternatives)
• Centralized Login for your friends and family with email-based password reset
• 2FA already set up
• A nice reflector gauge to see if you’re ports are open and what your limits are
• Great client support on a LOT of devices
• Search is fast out of the box, even with extensive collections
• Their clients tend to do a better job supporting all the decoding features on every player

Cons:

• Not free
• Not Open
• They have a lot of your historical data and will eventually sell it when they sell the company. This is not going to be optional. That data is worth a lot and they likely already have enough EULA rights to sell it to whoever asks. Imagine if the MPAA gets in on the fun.
• Their security history is quite dicey
• The lifetime membership will eventually be enshitified as it’s not economically sound in the long run
• They constantly change the terms of the agreement.
• They constantly remove features people are using
• They constantly push to share data between users
• They constantly push Ads
• They are making previously free features pay.
• Their investors are starving, which makes them a liability.
• Their clients are generally slower.

Aww come on guys, my JF boner can only handle so much /s

Seriously though, why did they even give you the option to disagree, you know they’re just going to force it 3-6 months.

👋

Fantastic, I don’t think I’ve ever seen you give on anything before.

This is cool, are you incapable of not getting the last word in? My schedule just opened up today and man, I’ve got all day.

Nice gaslight you got there. You do you man keep trolling.

As I said gentry, you do you.

All you need do is look at the histories of Playon, Evernote, LastPass, and the hundreds of other titles that have changed their business model to be less free.

If you hear your engine knocking, you should plan for your car to break down.

Their free tier has always been about the same IRRC. ACL’s are basic only. Limited, but still generous users and devices.

If the drop device counts on free or eliminate ACL’s we’d also advocate leaving them.

Everything you said is true.

Also of note, tailscale is making a shit load off money off corporate America. Every one of us who are deciding engineers that dips our toes in goes holy fuck that’s cool and immediately pushes to implement it in a professional capacity.

And, when Tailscale reduces its free offering, it’ll be time to move to Headscale (or elsewhere).

We (most) are not advocating leaving Plex because it is not FOSS. We’re advocating leaving them because they are changing the terms in ways that have repeatedly suggested that it is circling the drain and feasting on its current userbase in a very Google/Apple/Microsoft way.

We’re not getting anything out of people leaving Plex. There’s no stock here, the community is not so small that it needs all the people from Plex. It’s a humanitarian effort, probably a neurodivergent one, but still humanitarian.

LOL

Hope for the best, plan for the worst,

or don’t plan, you do you!