libtorrent | libTorrent BitTorrent library

Assuming you mean "get" when you say "pull" and "call" (i.e. you have a torrent_handle and you want to know what its save path is); you can ask for the torrent_status by calling handle.status() and the torrent status object has a save_path member.

It's documented here and here.

Kindly follow the below given YouTube link. This would be helpful to you.

https://youtu.be/kFG2YHVGZLs

In short, just add the following two lines before "!apt install python3-libtorrent"

!python -m pip install --upgrade pip setuptools wheel

!python -m pip install lbry-libtorrent

First of all, there are two BEPs definining merkle-tree based torrents. BEP30 and BEP52. Neither enables sharing a file based on its merkle hash alone as they require some additional data to build a minimal viable torrent such as the piece size or the file size. To support multiple files, file names and other things some additional metadata is needed which ultimately gets encapsulated in the infohash. So torrent clients generally deal in infohashes, not file hashes.

You can bypass some of these requirements by choosing a fixed filename and fixed piece size and then generate a canonical metadata representation for your file. In libtorrent that could be done through the create_torrent() API by creating a v2-only torrent (BEP52 without backwards compatibility).

that call (std::map::find()) was added in C++14 (see docs). I see you pass in -std=c++14 on the command line as well. Are you sure your GCC supports C++14? It seems a bit old for that.

The current stable branch of libtorrent only requires C++11 support, if that is the branch you're building, there may be something wrong with the compiler support detection here. If you are building from libtorrent master, it requires proper C++14 support. So in that case you may want to use the stable release.

I ended up solving my problem by utilizing the 'file_priorities' property in params. This still downloads the files until the metadata is retrieved, however, so I just saved them to a temporary directory.

The official documentation only lists some high-level DHT APIs. The lower-level implementation is marked as TORRENT_EXTRA_EXPORT which I believe means it's only meant for testing. The ping method specifically is private.

Since you're using java you could use mldht (I'm the author), which exposes lower-level functionality too.

you're setting seed_mode to True. Which will make libtorrent assume you have all the files for the torrent, hence the state being "seeding".

Once you get a peer, and it requests a block, libtorrent will actually go and open the file, and it will actually verify the piece hash as well (to avoid uploading corrupt data). However, if the file isn't there, the torrent will transition into checking mode, where it can no longer trust that it is seeding. It will check all the files.

The first step in checking the files is checking resume data (which you transition into). However, this is typically very quick and is followed by transitioning into checking-files state.

This doesn't seem to happen for some reason. To trouble shoot that you should pop_alerts() and print them to a log, and possibly enable more alerts, like torrent_log alerts.

However, it appears the first problem you have is that you don't have the files you're trying to seed. It's very hard to diagnose this without the alert log though.

the session_settings type and set_settings() function on session are deprecated (and have been for quite a while). The reference documentation online (https://libtorrent.org) is for the most recent stable release, so you won't find them documented there.

Instead, use settings_pack and apply_settings() on the session. Or even better, pass in your settings pack to the session constructor.

In the C++ interface, settings_pack is a class with a fairly simple interface, but in the python binding it's just a plain dictionary.

To set up a settings pack in python, you do this:

Usually, this is represented with targets and links between them.

Your setup could look like this: