A valid Kodo license is required if you wish to use this project.
Please request a license by filling out the license request form.
Kodo is available under a research- and education-friendly license, you can see the details here.
If you try to configure without a valid license, then you will get an error!
kodo-python contains a set of high-level Python bindings for the Kodo Network Coding C++ library. The bindings provide access to basic functionality provided by Kodo, such as encoding and decoding data. The examples folder contains sample applications showing the usage of the Python API.
If you have any questions or suggestions about this library, please contact us at our developer mailing list (hosted at Google Groups):
First of all, follow this Getting Started guide to install the basic tools required for the compilation (C++14 compiler, Git, Python).
The compilers used by Steinwurf are listed at the bottom of the buildbot page.
These steps may not work with your specific Linux distribution, but they may guide you in the right direction.
First, acquire the required packages from your package management system:
sudo apt-get update sudo apt-get install python build-essential libpython-dev python-dev
If you are using Python 3, you'll need to install libpython3-dev
instead.
Install the latest Command Line Tools and/or XCode to compile this project. The latest tested version is 10.2.1, so please upgrade if your CLT or XCode is older than that.
Python 2.7.10 is pre-installed on OSX, but some required Python headers are missing, so you need to install a more recent Python version from Homebrew (or Macports).
You can choose to install Homebrew's Python 2 which will become the default
Python on your system, so you can call waf using the python
command:
brew install python@2 python waf configure python waf build
You can also choose Homebrew's Python 3, but then you must always use the
python3
command for invoking waf:
brew install python3 python3 waf configure python3 waf build
First of all, you need to install Visual Studio 2017. There are many variants, but you should basically get the same C++14 compiler: VS Express, Community, Professional, Enterprise or the standalone Build Tools might all work.
You can choose to install Python 2.7 or Python 3.7+. It is very important to install 32-bit Python for a 32-bit VS toolchain and you need 64-bit Python for a 64-bit VS toolchain. Some Visual Studio versions only provide a 32-bit toolchain, so this might be the only option.
If you installed Python 2.7, then you need to set the VS90COMNTOOLS
environment variable to point to the folder that actually contains
vcvarsall.bat
(this depends on the version of VS2017 that you installed):
C:\Program Files (x86)\Microsoft Visual Studio\2017\WDExpress\VC\Auxiliary\Build C:\Program Files (x86)\Microsoft Visual Studio\2017\Community\VC\Auxiliary\Build C:\Program Files (x86)\Microsoft Visual Studio\2017\Professional\VC\Auxiliary\Build C:\Program Files (x86)\Microsoft Visual Studio\2017\BuildTools\VC\Auxiliary\Build
If you installed Python 3.7+, then you need to set the VS140COMNTOOLS
environment variable to point to the folder that actually contains
vcvarsall.bat
(this depends on the version of VS2017 that you installed):
C:\Program Files (x86)\Microsoft Visual Studio\2017\WDExpress\VC\Auxiliary\Build C:\Program Files (x86)\Microsoft Visual Studio\2017\Community\VC\Auxiliary\Build C:\Program Files (x86)\Microsoft Visual Studio\2017\Professional\VC\Auxiliary\Build C:\Program Files (x86)\Microsoft Visual Studio\2017\BuildTools\VC\Auxiliary\Build
After setting the appropriate environment variable, you need to make sure that
Python's distutils can find a valid location for vcvarsall.bat
, because
waf will call that batch file to obtain some compiler flags. If you can execute
the following test script without getting an exception, then you should be able
to configure kodo-python using waf:
if __name__ == "__main__": from distutils import log log.set_threshold(log.DEBUG) from distutils.msvccompiler import MSVCCompiler dist_compiler = MSVCCompiler() dist_compiler.initialize() print("Compile options:") print(dist_compiler.compile_options) print("LDFLAGS:") print(dist_compiler.ldflags_shared)
If you have any issues with the test script, then most likely you need to
apply this one-liner patch to msvc9compiler.py
in Python27\Lib\distutils
or Python37\Lib\distutils
: https://bugs.python.org/file45916/vsforpython.diff
The problem is that the location of vcvarsall.bat
has changed with
respect to the Common Tools folder, so distutils cannot find it without
the patch. The issue is explained here: https://bugs.python.org/issue23246
You need to build the Python bindings from source.
First, clone the project:
git clone [email protected]:steinwurf/kodo-python.git
Configure and build the project:
cd kodo-python python waf configure python waf build
After building the project, you should find the resulting kodo.so
,
kodo.dylib
or kodo.pyd
file here (the actual path and extension
depend on your OS):
build/linux/kodo.so build/darwin/kodo.dylib build/win32/kodo.pyd
You can copy this file to the same folder as your Python scripts, or you can copy it to your PYTHONPATH (so that you can import it from anywhere).
Then you can import the module in your Python script:
>>> import kodo
With the enable_codecs
option, you can configure kodo-python to only enable
some desired codecs and disable all others. For example:
python waf configure --enable_codecs=rlnc
Run python waf --help
to list the available codecs. You can even
select multiple codecs with a comma-separated list:
python waf configure --enable_codecs=rlnc,fulcrum
The compilation process might take a long time on certain Linux systems if less than 4 GB RAM is available. The g++ optimizer might consume a lot of RAM during the compilation, so if you see that all your RAM is used up, then you can try to constrain the number of parallel jobs to only one during the build step:
python waf build -j 1
With this change, a fast compilation is possible with only 2 GB RAM.
This issue is specific to g++ (which is the default compiler on Linux), and the RAM usage and the compilation time can be much better with clang. The code produced by clang is also fast.
If the compilation does not work with g++, then you can install clang like this (on Ubuntu and Debian):
sudo apt-get install clang
Then you can configure the project to use clang++:
CXX=clang++ python waf configure
The detailed instructions for compiling the project on the Raspberry Pi are found in our Raspberry guide.