Show HN: Torrent search engine but the database is fetched locally from IPFS

giga.cat

9 points·etagate··4 comments

Hello HN!

This website doesn't host any record/magnet, but it fetches a dump (csv file) from IPFS and then it builds a database locally on browser with 'flexsearch' so you can do text searches and find what you need.

By default, for the showcase, it's wired to a copy of a dump from TNT village (dead italian forum) by its CID, but you can add your own dump to IPFS and then share the dump by attaching the resulting CID as a parameter to the link like: giga.cat?list=Qm.....XyZ

It is nothing serious, it's not perfect, it was my first time using js with dozens of async/await.

I was bored a month ago so I decided to learn more about IPFS and I was surprised that this awesome project is also being implemented with javascript, so you can use it even without installation. I tested if was feasible to share a whole csv file of (title, magnet, ...) and how many records were acceptable because of loading time increase. Well it takes 7s to fetch, parse and populate from a 15MB csv file with ~100k lines. Not that usable, it's just an upper limit. I think it could work well for small communities.

Wow this is interesting. Congrats on the launch!

For those who don't know IPFS, or how this would be used, could you share an example on how this would be useful (vs. other providers)?

Instead of CSV, you could serve an sqlite database directly and query it using sql.js or absurd-sql.

Very interesting, I could apply this for a specific list. Thanks. Btw I designed this specific website for non-tech users to let them share their dumps. Populating and dumping a database it's a bit hard for a non-it user, while I see it's very common to share csv files

Reply on news.ycombinator.com