Link: https://app.radicle.xyz/nodes/rosa.radicle.xyz/rad:z3j3VDRonEWNoasPTcHbiXzMWkLbo

Hello, since it's complicated to index DHTs, I figured it'd be more efficient to build an index of fingerprints from real data once.

So I've been collecting releases hashes for this index. It can be used for various purposes:

• check the integrity of your own files (bit rot is a real thing)
• identify BTv2 torrent files that contain specific files (a database of torrent files is required)
• locate alive IPFS swarms to join more easily (no need to read all your data multiple times to recompute various CIDs yourself)

The collection contains around 1K releases and weights 40MB.
I've prioritized scene Bluray rips of movies (1080p / 2160p).
No infohash will be included, as these are not reproducible enough.

I'm using a basic script to add a new release (filename must match the official release name). I'm using others to discover scene releases in a filesystem; retrieve official release names from files using the srrdb api (crc32 search); collect torrents from Prowlarr and H&R them (although I'd prefer crowd-source directly from the community!).

The index is stored on git to allow collaboration. It is hosted using Radicale to avoid centralization and reduce hosting pressures.

If you are interested, join and add your own hashes to the collection in Radicle patches! (see instructions in the README)

Let me know what you think, suggest improvements or discuss similar projects you know about!

This is definitely up my alley, I gave up on keeping all my media in my torrent client indefinitely for seeding because of the performance, so I've long dreamed of making some way to reconnect loose files back to torrents so I can seed them.

Seems I could maybe build something on top of this? I tried running magnetico for a while (going so far as to add postgres support to help it scale) but it quickly grows far larger than I want to manage.

My next idea is to make a file scanner that maintains a list of file paths and several common hashes, then do a dht crawl and only save stuff that matches. Then I can hopefully automatically add and remove torrents to a client that has read-only access to the files as needed (remove if plenty of seeders, keep for a while if no or low seeders and rotate through prioritizing stuff that needs seeds)

I'm wondering if there's some useful overlap between what you're doing and my goals but I think I need to dig into it more.

Hi, yes it definitely sounds similar for the media files database side. Using a DHT crawler, you can identify new torrents matching specific file tree roots (so only works for bittorrent v2, which is not used so much for now), and update swarms statistics (S/L).

Sweet your using radicle

Very cool! Sorry I have nothing to add to it, but this is a great idea, and radicle is a great fit for this.

Sounds cool, wonder if only make sense for vid only. What if smaller files like txt, img gallery

What if file name/path change? Scale resolution up/down? Genre/Subgroup in future?

I don't think it would scale for many millions of files.

Hi, not sure I get your point. A release has a fixed name and resolution. Regarding genre classification, I estimated if was too complex to determine reliably. And I wanted to avoid storing metadata.

Dates folders are there to reduce directory sizes, similarly to a merkle tree with only one level. This is due to performance limitations with large directories in most filesystems and also in Git. Also it still allows for easy manual search (only the date is required).

Note: there can still be several releases for a single movie (several resolutions and sources, LIMITED/REPACK etc).

I think you are also tried to solve same video but different name, release, resolution by different uploaders problem.

For now it's only scene releases, so there's no duplicate (not sure if this is what you mean).