Fediverse

Looks like your comment got triple posted. The lemmy instances are getting slammed lol

Yes but that search doesn't take you to the instance that you are logged into already. Which is one of my main goals with this site. While that did give me the inspiration for this and has the power of Google behind it, it lacks knowledge about how the fediverse actually works.

Why do I learn about this now that I pretty much don't care?

Not yet but I can add this feature

Great heads up

I just don’t get why people would join SOCIAL media if they don’t want to their posts public. Its like having a group meeting in the middle of a public square and complain people saw them

Maybe they should just use discord or or private communities instead

Lemmy is not private in any way. In fact, while the OP's project is really cool and admirable, there is already https://fedi-search.com, which searches top Lemmy instances (and kbin, Mastodon, and peertube) using regular operators behind the scenes in Google, Bing or DDG (whichever you prefer), because those search engines already crawl Lemmy along with the rest of the internet, unless the site owners block crawling via robots.txt.

I've already got some complaints about that. You can see one of the issues raised on GitHub.

At the moment, I'm only picking up mastodon posts that are federated to Lemmy, but you can't choose Mastodon as a preferred-instance, yet. When and/if I decide to add Mastodon support, I'll reach out to the admins over there to get feedback first.

Edit and note to any server admin: If you want to block the crawler from hitting your site, just add lemmy-search to your robots.txt and crawling will be prevented. But this doesn't stop cross-federation posts from being picked up on another instance.

TBH they can block them from search in the preferences. Otherwise anything on the Web is searchable and findable.

Are the search result links not opening on your home instance? If so please raise an issue on GitHub with your home instance URL and I'll investigate. But you should be able to select your home instance from the drop-down and then search and all of the results will have their links direct to that instance. (This does require that your instance has been indexed by the search engine, which as of this moment I've only indexed 278 of the nearly 1000? Lemmy instances out there.)

It might not have been crawled yet. The search engine will periodically search for new content but this isn't instant. So it may take a day or two to find it.

Due to a one of my primary goals of having links open in your home instance, you can only search a single instance at a time.

There's an open issue about this though but it requires some changes to Lemmy itself to work.

Ya, now if everyone can stop finding bugs! So I can take some time off. /jk

Same. I guess it is some Cloudflare service to prevent DDoS attacks or something

See one of my other replies. But that was a thought originally. Just hook into the original database instead of crawling using the APIs. Problem is, the table structure required to search is much different than that of a community form. At least if you want to do searches quickly. It takes me almost 5-10 seconds just to process 50 posts at the moment, and I'm doing those in batches... but ya maybe in the future I can talk to the Lemmy devs and see about merging these two projects?

"some search string instance:lemmy.world".

Keywords are:

instance:<instance name>

community:!<community name>@<instance name>

and

author:@<author name>@<instance name>.