Decentralized DHT search site for IPFS
If you don't know what Torrent Paradise is, see the website.
This is a repository of all the tools I use to build and run torrent-paradise.ml. The 'code name' of the project is nextgen (next gen torrent search), so don't be surprised if it comes up somewhere.
Maybe, open an issue. Be sure to demonstrate an effort that you tried to solve the problem yourself.
Here's what the setup looks like rn: - VPS, Debian Buster, 4 GB RAM - user with username nextgen on the server - my laptop w/ Linux - Go toolchain installed - node v10.15 & npm - Python 3 (required only for index-generator/fix-metajson.py)
The programs create their own tables in the DB that they need. Database name is "nextgen". You need to create the materialized views (fresh and search). You can find some useful SQL code in snippets.sql.
Each of the daemons (api, crawl-rss, seedleech-daemon) is its own standalone Go package and resulting binary. You have to compile the binaries yourself. There are systemd .service files available for each of the daemons.
The torrent collection is a mashup of the (now no longer provided) TPB dumps, my own DHT spidering efforts, and magnetico community database dumps.
The easiest way to get your own site up and running is to start with my .csv dump. It should be easy to import into any kind of system. It contains seed/leech counts too (!). If I were to import it, I'd modify import-magnetico-db.
Generation of the IPFS index will prob take a long time, a machine with high single-core perf recommended (ipfsearch runs on node.js)
go buildin spider/ to compile and scp the binary it to the server. You can use the systemd service file in
spider/spider.serviceto start the spider on server boot.
go buildin seedleech-daemon/ to compile and scp the binary it to the server. You can use the systemd service file in
Use sqlite3 on a the decompressed dump to generate a .csv file. Format: infohash,name,length(bytes). Optionally quoted.
Then use the go binary in import-magnetico-db to do the import.
The directory website gets deployed to IPFS, static gets deployed to the server. Static calls the API, the IPFS version doesn't.
Before working on something, open an issue to ask if it would be okay. I would love to KISS.