Given an input torrent, it will scan your collection for the files in the torrent. If all (or most) the files are found, a folder with links to all the files will be created and the torrent added to the torrent client.
All you need to do is download the torrents and AutoTorrent plays mix and match to make it possible to seed as much as possible across trackers.
- Linux, BSD, OSX - Something not windows
- rTorrent, Deluge and Transmission
- Python 2.6, 2.7, 3.3, 3.4
- Shell / SSH / Putty
From PyPi (stable):
virtualenv autotorrent-env autotorrent-env/bin/pip install autotorrent
From GitHub (develop):
virtualenv autotorrent-env autotorrent-env/bin/pip install git+https://github.com/JohnDoee/autotorrent.git#develop
Get the configuration file
wget -Oautotorrent.conf https://github.com/JohnDoee/autotorrent/raw/develop/autotorrent.conf.example
Upgrade from previous version¶
Upgrading from PyPi (stable)
autotorrent-env/bin/pip install --upgrade autotorrent
Upgrading from Github (develop)
autotorrent-env/bin/pip install git+https://github.com/JohnDoee/autotorrent.git#develop --upgrade --force-reinstall
All settings can be found and changed in autotorrent.conf, this file must reside in the same folder as autotorrent is executed from.
- db - Path to the database file
- store_path - Folder where the virtual folders seeded, resides
- ignore_files - A comma seperated list of files that should be ignored (supports wildcards)
- add_limit_size - Max size, in bytes, the total torrent size is allowed to vary
- add_limit_percent - Max percent the total torrent size is allowed to vary
- link_type - What kind of link should AutoTorrent make? the options are hard and soft.
- scan_mode - options are unsplitable, normal and exact. These can be used in combination. See the scan_mode section for more information.
the add_limit_* variables allow for downloading of e.g. different NFOs and other small files that makes a difference in the torrents.
- client - torrent client to use, choices are: rtorrent, deluge and transmission
- url - URL to rtorrent, must be to the XMLRPC server or SCGI server.
- label - Label added to torrents when added to rtorrent (used in rutorrent only)
the url supports both SCGI directly and XMLRPC via HTTP.
To use scgi, prefix the url with scgi instead of http, e.g. scgi://127.0.0.1:10000/
To use unix socket for scgi, make an url with no ip:port and instead a path, e.g. scgi:///tmp/rtorrent.socket
- host - an ip:port pair, e.g. 127.0.0.1:12345
- username - deluge rpc username
- password - deluge rpc password
A list of disks where to build the search database from.
The keys must be sequential, i.e. disk1, disk2, disk3 etc.
There are currently three scan modes supported by AutoTorrent. These modes can be used in combination and should all improve the end result.
The modes are named normal, exact and unsplitable. They can be combined by adding a comma
between them, e.g.
It takes the filename and size and tries to find files with same name and size.
This mode cannot handle duplicate filename/size pairs.
The perfect way to move torrent client as it tries to set the download path to the old path.
This mode does not allow for missing files and is intended to re-add non-renamed back to a torrent client.
This mode takes scene releases and extracted dvd/bluray isos into consideration and relies on the folder it thinks is the main / head folder. Perfect for cross-seeding scene releases.
This mode tries to hashcheck files with the exact name as wanted, but the size might be different (up to 10% different). If pieces match, then it is resized to fit original size and written to the destination directory.
Make sure there is enough space in the target directory.
This mode tries to hashcheck files with the exact size as wanted, but the name might be different.|
This mode tries to hashcheck files with a size within 10% of the original. If pieces match, then it is resized to fit original size and written to the destination directory.
Make sure there is enough space in the target directory.
This mode is very slow as it will try a lot of files.
Start by installing and configuring.
Step 1, build the database with
autotorrent -r, this may take some
Step 2, have some torrents ready and run
autotorrent -a folder/with/torrents/*.torrents, this command will
spit out how it went with adding the torrents.
And you’re good to go.
Q: How are files with relative path in the configuration file, found?
The paths should be relative to the configuration file, e.g. /home/user/autotorrent-env/autotorrent.conf, then store_path=store_paths/X/ resolves to /home/user/autotorrent-env/store_path/
Q: I have three sites I cross-seed between, how do you suggest I structure it?
Say, you have site X, Y and Z. You want to seed across the sites as they share lots of content. You download all your data into /home/user/downloads/. For this you will need three configuration file, one for each site.
AutoTorrent is installed into /home/user/autotorrent-env/.
Only store_path is recommended to vary between the configuration files (the others are optional).
- store_path for site X - /home/user/autotorrent-env/store_paths/X/
- store_path for site Y - /home/user/autotorrent-env/store_paths/Y/
- store_path for site Z - /home/user/autotorrent-env/store_paths/Z/
disks paths can be:
Q: Can I use the same Database file for several configuration files?
Yes, if they have the same disks. Don’t worry about adding the store_path to the disks, AutoTorrent will figure it out.
Q: What problems can occur?
One big problem is that the files are not checked for their actual content, just if their filename matches and size matches. If AutoTorrent tries to use a file that is not complete, then you can end up sending loads of garbage to innocent peers, alhough they should blackball you quite fast.
Q: I want to cross-seed RARed scene releases, what do you think about that?
The actual .rar files must be completely downloaded and the same size. Things that can vary are: nfos, sfvs, samples and subs.
The releases must also have an sfv in the same folder as the rar files files.
Q: What are hardlinks and what are the risks or problems associated with using them?
MIT, see LICENSE