Skip to content

Latest commit

 

History

History
123 lines (80 loc) · 4.14 KB

python_binding.rst

File metadata and controls

123 lines (80 loc) · 4.14 KB

libtorrent python binding

Author: Arvid Norberg, [email protected]

building

Building the libtorrent python bindings will produce a shared library (DLL) which is a python module that can be imported in a python program.

building using setup.py

There is a setup.py shipped with libtorrent that can be used on windows. On windows the setup.py will invoke bjam and assume that you have boost sources at $BOOST_PATH. The resulting executable is self-contained, it does not depend any boost or libtorrent dlls.

On other systems, the setup.py is generated by running ./configure --enable-python-binding.

To build the Python bindings do:

  1. Run:

    python setup.py build
    
  2. As root, run:

    python setup.py install
    

building using boost build

To set up your build environment, you need to add some settings to your $BOOST_BUILD_PATH/user-config.jam.

Make sure your user config contains the following line:

using python : 2.3 ;

Set the version to the version of python you have installed or want to use. If you've installed python in a non-standard location, you have to add the prefix path used when you installed python as a second option. Like this:

using python : 2.6 : /usr/bin/python2.6 : /usr/include/python2.6 : /usr/lib/python2.6 ;

The bindings require at least python version 2.2.

For more information on how to install and set up boost-build, see the building libtorrent section.

Once you have boost-build set up, you cd to the bindings/python directory and invoke bjam with the apropriate settings. For the available build variants, see libtorrent build options.

For example:

$ bjam dht-support=on link=static

On Mac OS X, this will produce the following python module:

bin/darwin-4.0/release/dht-support-on/link-static/logging-none/threading-multi/libtorrent.so

using libtorrent in python

The python interface is nearly identical to the C++ interface. Please refer to the library reference. The main differences are:

asio::tcp::endpoint
The endpoint type is represented as a tuple of a string (as the address) and an int for the port number. E.g. ('127.0.0.1', 6881) represents the localhost port 6881.
lt::time_duration
The time duration is represented as a number of seconds in a regular integer.

The following functions takes a reference to a container that is filled with entries by the function. The python equivalent of these functions instead returns a list of entries.

  • torrent_handle::get_peer_info
  • torrent_handle::file_progress
  • torrent_handle::get_download_queue
  • torrent_handle::piece_availability

create_torrent::add_node() takes two arguments, one string and one integer, instead of a pair. The string is the address and the integer is the port.

session::apply_settings() accepts a dictionary with keys matching the names of settings in settings_pack. When calling apply_settings, the dictionary does not need to have every settings set, keys that are not present are not updated.

To get a python dictionary of the settings, call session::get_settings.

Retrieving session statistics in Python is more convenient than that in C++. The statistics are stored as an array in session_stats_alert, which will be posted after calling post_session_stats() in the session object. In order to interpret the statistics array, in C++ it is required to call session_stats_metrics() to get the indices of these metrics, while in Python it can be done using session_stats_alert.values["NAME_OF_METRIC"], where NAME_OF_METRIC is the name of a metric.

For an example python program, see client.py in the bindings/python directory.

A very simple example usage of the module would be something like this: