The quality of search engines has gone down so much for technical questions.
I’m looking for a way to index sites like stack exchanges, reddit, quora, and research papers. Would this be possible to do this locally with metadata?
This is probably going to be an unpopular opinion here, but I don't imagine that everyone selfhosting their own search engine is going to be very scalable...
In addition to YaCy and the varieties of Searx (both of which perform better for me than any of the commercial search engines), it’s not even out of the question to do this yourself, if you’re willing to start with the most recent Common Crawl dump and do some spidering in between releases. I don’t recommend it, unless you want to learn for yourself why search engines often give such miserable results, but it’s possible.
However, that’s the issue, here. Can you self-host a search engine? Sure, if you want to maintain the storage to back it. That depends on how deep your pockets go…
I’m glad you ask! I have been working on a peer-to-peer search engine. The goal is to index websites that are on IPFS. It’s an MVP but you can already try the demo and run your node to make your data searchable for the whole network (you would need to generate html files from your data and put it on ipfs first)
Even the main search engines don’t index the entire internet of content these days and their databases are truly massive already. Writing a basic web crawler to produce a search index isn’t all that hard (I used to do it as a programming exercise for applicants) but dealing with the volume of data of the entire internet and storing it to produce a worthwhile search engine however is just not feasible on home hardware, it would be TB’s at least. It wouldn’t just be a little worse it would dramatically worse unless you put substantial resources to it including enormous amounts of network bandwidth which would have your ISP questioning your “unlimited 1 gbps fibre” contract. It would probably take years to get decent and always be many months out of date at best.
Doesn’t seem practical to try to self host based on the need to download and index every single page of the internet its a truly massive scale problem.
StackExchange dumps are available for Kiwix, the project that allows to use a local dump of Wikipedia. You can find all the available dumps there, including the StackExchange ones. You can even build your own search engine through libs allowing to use those zim files (the dumps), if you want.
I mean, you can easily self host a meta-search engine like Searx, Searx-ng, Whoogle, etc. I run Searx-ng and it sends your queries to multiple engines and aggregates the results for you.
To host your own search engine, you’d need to crawl and index every site. It’s certainly doable, but it would take a lot of time /effort.
I agree. Selfhosting a true search engine is way too much work and infeasible for individuals. Meta search engines however are very feasible and a great option.
Add comment