Am I right in assuming that - API wise - the bot only interacts with ponder.cat, and doesn’t make calls to the remote instance? (I’m wondering if there’s any barriers to it operating with communities that aren’t on a Lemmy instance).

Yes, that’s right. It should work fine on a non-Lemmy instance.

Does the bot resolve the human first, check what they moderate, and then resolve the community if they moderate it, or just always resolve the community, and then compare its moderators with who made the request? If its the latter, this could be a way for bad actors to crowbar a community onto your instance (assuming it doesn’t purge it if things don’t match up, of course).

It’s the latter. I think it’s okay. The same thing can happen on any instance where someone can search for a community from any other instance.

What would have happened if Otter had sent /add https://lemmy.ca/feeds/c/medicine.xml medicine@lemmy.ca ? Would this be like that time when someone put ‘google’ into google.com, and the Internet blew up?

It’s limited to one posting every 5 minutes per feed, so the damage would be limited, but you’re right that it would enter an infinite loop and post once every five minutes until someone put a stop to it.

Really great tool, thanks!

Thank you!

In the commands, will {instance} always be rss.ponder.cat?

create account on rss.ponder cat

Or do you make the communities and then we add feeds to them?

No to all. This particular tool is only for communities on other instances. It doesn’t interact with the big feeds on rss.ponder.cat.

rss.ponder.cat is for the all-RSS-post communities that I’ve been making. A lot of them will be pretty heavy on their posting, so some people may prefer to block the whole thing wholesale. I can add communities if people request it, but it’s something I want to be a little bit careful with, so as not to create too much spam.

This new tool is designed to add RSS feeds to communities outside of ponder.cat. Something like releases of a FOSS project, weather updates for a city, things like that. The moderators of those communities can use the bot to do whatever they want within their communities, without having to involve me.

Does each message need to have only one command?

No, you can issue multiple commands. It should work fine. Of course if it gives you any issues, you can let me know.

Edit: Otter already answered, I just didn’t see it. I’m leaving it for posterity, though.

I am fetching the RSS feeds for particular channels, for which I need the channel ID.

Google gives out not only the RSS feed, but also the channel ID, if you click the menu under “share”, as someone else pointed out to me a couple days ago. You are very confused about things. This thing about it being against the TOS is pure fantasy.

*fifth, with SO’s default sorting, after several screens full of wrong answers

It’s a wonderful answer and I appreciate knowing about it now, because that process definitely wasn’t the first idea that came to my mind. If only Google and SO had given that answer to me before I did the whole adventure I pasted into my post and comment.

It’s search. I’m copy-pasting the content going down from the top of the page, adding my own commentary.

The mystery is solved. People are trying to provide good answers to strangers for free because they are good and helpful. Google’s senseless reorganization of things is causing them not to function anymore. And then, Stack Overflow’s bad configuration is defeating those good people’s efforts to provide them to me.

I think we’ve learned that SO has succumbed to the same organizational syphilis that Google contracted a few years earlier.

I’ll click on the link you sent me, and start reading carefully, trying to find the answer. Here’s how it went. I am not faking or deliberately trying any wrong things here. I’m just grabbing the most central solution it’s presenting me on any given page, and trying it.

Here’s the progress:

or those who come across this after the latest Data API revision (31st January, 2024), they provided the forHandle query parameter to get data of a handle/username. You can refer to this answer for details here: stackoverflow.com/a/78074066/2665606 – Saqib Ahmed Commented Feb 29 at 10:01

Cool. I click on that question.

YouTube released a revision on 31st January 2024 to add a forHandle parameter in the channel list API that does exactly what OP asks.

You can call the channel list API with forHandle to get the channel ID and the upload playlist for that user/handle that you can subsequently use to fetch the videos.

GET https://www.googleapis.com/youtube/v3/channels

Query parameters:

forHandle=FolkartTr OR @FolkartTr OR %40FolkartTr (this allows the ‘@’ sign as well as URL encoding of it) key=<your API key> part=contentDetails

Cool.

$ wget -O - https://www.googleapis.com/youtube/v3/channels\?forHandle=Lindybeige
--2024-10-01 01:19:52--  https://www.googleapis.com/youtube/v3/channels?forHandle=Lindybeige
Resolving www.googleapis.com (www.googleapis.com)... 2a00:1450:4009:822::200a, 2a00:1450:4009:826::200a, 2a00:1450:4009:823::200a, ...
Connecting to www.googleapis.com (www.googleapis.com)|2a00:1450:4009:822::200a|:443... connected.
HTTP request sent, awaiting response... 403 Forbidden
2024-10-01 01:19:52 ERROR 403: Forbidden.

Oh. I need an API key. Okay, that one’s useless.

I read another answer.

If I understood correctly, your problem is that you can’t do anything from such a c/ channel id with the Channels: list of the YouTube Data API v3. If you’re just looking for the channel id linked to this id then because as YouTube Data API v3 doesn’t work for this, I would recommend you to use my open-source YouTube operational API, indeed by requesting https://yt.lemnoslife.com/channels?cId=FolkartTr you’ll receive a JSON with id equals to the channel id linked to the provided cId value.

That means nothing to me.

I hit back. We’re back at the original page you linked me to:

To obtain the channel id you can view the source code of the channel page and find either data-channel-external-id=“UCjXfkj5iapKHJrhYfAF9ZGg” or “externalId”:“UCjXfkj5iapKHJrhYfAF9ZGg”.

UCjXfkj5iapKHJrhYfAF9ZGg will be the channel ID you are looking for.

Already covered. It doesn’t work.

An easy answer is, your YouTube Channel ID is UC + {YOUR_ACCOUNT_ID}. To be sure of your YouTube Channel ID or your YouTube account ID, access the advanced settings at your settings page

And if you want to know the YouTube Channel ID for any channel, you could use the solution @mjlescano gave.

https://www.googleapis.com/youtube/v3/channels?key={YOUR_API_KEY}&forUsername={USER_NAME}&part=id If this could be of any help, some user marked it was solved in another topic right here.

Are you getting sick of reading these? So am I! I want to remind that DDG gave me the answer at the top of the page, as a tool that would solve the problem for me.

At any channel page with “user” url for example http://www.youtube.com/user/klauskkpm, without API call, from YouTube UI, click a video of the channel (in its “VIDEOS” tab) and click the channel name on the video. Then you can get to the page with its “channel” url for example https://www.youtube.com/channel/UCfjTOrCPnAblTngWAzpnlMA.

Edit:

Above is not working any more. But we can open Developer Tools (cmd + option + I) and try to find the URL there. Search by channel_id for some channels, it will show you, but NOT for all the channels.

By the way, if this is your own channel – you can go here https://developers.google.com/youtube/v3/docs/channels/list and make request with part snippet and mine true.

Oh, I found this answer. Thank you, just works!

Edit:

Just in case you need UC channel id of any channel by the “YouTube handle”, you can also use ‘API Explorer’ on the right of this page https://developers.google.com/youtube/v3/docs/channels/list and enter forHandle with your ‘API key’ (let me share screenshots below)

Sure, let’s try the Cmd+Option+I solution.

Hey! Look at that. www.youtube.com wants to use my microphone. I can:

• Allow while visiting this site
• Allow this time
• Never allow

I think we’re done here. For all I know I would have been able to find it in the network tab of developer tools while searching for the channel, but I think the point is made. I didn’t say that the answer didn’t exist anywhere on SO, I said that the things I was trying because either Google or SO were telling me they were the answer were not working.

Edit: I reread your comment and got confused. For me, the screenshot you’re showing has 26 points, and shows up below all of the answers that I showed above, which have 259, 79, 32, and 30 points respectively. How many points does it have on your page, to show as the second highest answer? I didn’t deliberately stop reading right before the answer. I absolutely made a sincere effort to find the answer on that page, documenting my progress as I went.

Ed Zitron diagnosed the causes and timing of the rot:

https://www.wheresyoured.at/the-men-who-killed-google/

I think that over the next few years Sam Altman is going to learn the same lessons that events have been trying to teach Elon Musk since circa 2021.

1. You didn’t build that. The people that work for you did.
2. Being a big hero is contingent on you and your behavior, and can change.
3. Those people who are giving you all this money aren’t your comrades. When your usefulness is at its end, they won’t give you a second thought.

How about tlnet? I just made !tlnet@rss.ponder.cat.

It is and I could. I’d be fine doing it, but why not just read them when they come up in !fediverse@lemmy.world? I’ve been trying not to create communities that are going to be duplicates or spam, or split the user base between one way of reading articles and another way of reading articles. Do you want it as a DM, maybe?

I think an even better way would be software that can follow the original Wordpress feed, if they have one, but Lemmy can’t do that right now.

Not a problem at all. I think a better way to do that will be to let moderators of existing communities add the bot to their existing communities. Someone asked about doing that, and it’s easy to set up the bot to make it possible, so I think I’ll just do that instead. I don’t need to create a duplicate community for anything that’s already got one.

I’m fine with the existing structure, with one community per periodical. I tried !coding_blogs@rss.ponder.cat and !science_streams@rss.ponder.cat and it looks like some people are into that type of structure, but I’m thinking mostly in terms of one-periodical communities or moderators from off-instance communities being able to add things.

Are there any that you would cherry-pick that you think you would personally use? I’d be perfectly willing to add them, if so.

I don’t think that the Austin or Texas communities are useful as communities. Do you mind if I delete them?

Are there other feeds from your OPML that you would really like to have in Lemmy?

Than there is problem with I don’t trust media will write the truth anyway, so giving them few bucks will probably not change that. But it is important for us to know what other people know.

Are there any media sources that I’m hosting feeds for which you feel that way about? I think the problem is much worse in a lot of free content, and I’ve been trying to bring in honest and high-quality sources when I’m adding news sources.

While this is outside of our current discussion, they need to find better model.

If it is a daily newspaper, maybe paywal new articles and release after sone reasonable time (like a week, or month… or a year).

I don’t understand, can you explain more?

Edit: I understand now. That’s outside the scope of my abilities… I would like to be able to offer a paid subscription with a deal that provides access to a wide variety of paywalled content, like a site license at a university, but I think that’s also outside the scope of my abilities. You’re right that they need a better model.

I like your idea of separating feeds, to keep paywalled content out of my feed.

It seems like a good compromise. I certainly understand that if someone’s decided not to read paywalled content, putting a lot of it into their Lemmy feed in a way that’s difficult to disable isn’t a good thing to do. I think separating the paywalled content into a separate user so it’s easy to block is probably a good pragmatic solution.

Most of the most popular RSS communities are free. I like some of them that are paywalled and a little way down the list, like !thenewyorker@rss.ponder.cat and !theatlantic@rss.ponder.cat, but most of the top ones are free. One of the really nice things about one community per source is that you know which ones to subscribe to and which ones you’d have to pay for that you can block.

If you don’t know the New York Times has a paywall, and you click on a link to them, that’s a learning experience for you at this point. I think some of the griping about paywalls is just entitled. It’s okay if people made content for you and they want to get paid. At the same time, I’m not trying to spam people who don’t want paywall content. If I can make a quality-of-life improvement for people who don’t want to get burned by paywalls on random links from places they’ve never heard of, then fine.

I also want to give shout-outs to some feeds that are way, way down and trying to charge money for very high quality stuff:

• !eenews@rss.ponder.cat
• !grist@rss.ponder.cat
• !themarshallproject@rss.ponder.cat
• !technologyreview@rss.ponder.cat
• !foreignpolicy@rss.ponder.cat

I completely agree. Maybe my phrasing was careless. I wasn’t trying to be critical of the pace of accepting PRs or anything. I only meant that I think more flexibility in the frontend would help, instead of needing any minor UI change to go all the way through a cycle all the way up to you, incorporating it into the core codebase, and then filtering back down to an upgrade by the instance admin. But please don’t take it as blaming you for any of that situation. I was raising it in the effort to propose a solution and also to advocate against people just complaining about the moderation tools and then moving on, and waiting for you to make them happy.

I did look at the backend plugin system PR, although sadly not enough yet to have any opinion or feedback on it. I do think a frontend plugin system, of sorts, could help a lot. I’m not sure when I will have time but I will try to put together something on this instance to show what I’m talking about, and if I do wind up doing it and it’s well received, I am completely open to putting it together as a fixed-up and official PR for the main codebase.

I’ll make an AI chatbot that only wants to talk about Dark Souls. How about that, as a compromise?

That is fair. Might be worthwhile talking to instance admins and core devs about how best to make use of it? Putting it behind some admin approval or administration might be the best way.

That’s a good idea. And then, if it turns into a mess of botspam, it’s not my fault.

This is a very good idea. Slashdot hit on something very good with their scoring and moderation system, and then the lessons were forgotten in the systems that came after Slashdot.

There are so many small tweaks like this tagging idea that would improve things, that I think the way to do it is a frontend “plugin” system that can accommodate them without it being a big fandango of tweaking the core to include fifty of these little nubbins all over the codebase.

Nothing that’s ordained from on high will ever be completely perfect. If you can have a little text box on your custom frontend, where you can tweak your own UI features, and then talk with the instance admin about importing one of those little tweak plugins for the whole site, or for your own community only, that sounds like a huge step forward to me. If you’ve ever added little tweaks to Wordpress, with the custom CSS boxes or by adding a few lines to functions.php, I envision this being similar.

This is all much more ambitious than I was thinking of when I asked this original question, but it would be a lot of fun to work on. If the result was it going back up into Lemmy, and each instance having its own collection of tweaks and a thriving community of people working on them without needing to disrupt the pace of development on the core, that would be a pretty excellent outcome.