This repo contains the spectral coarse graining routines from igraph 0.9. These routines were removed from igraph 0.10.0 due to their numerical instability in unit tests and due to the fact that none of the current developers are familiar enough with the topic to be able to sort out the unit tests. If you are interested in helping out, let us know!
The contents of this repo are unmaintained; we aim to keep on ensuring that the library compiles with the most recent version of igraph, and we try to fix bugs if you find one, but no new features will be added.
The following commnds should be enough if you have a working igraph installation that does not contain the spectral coarse graining routines any more! This applies to igraph versions 0.10 and later.
Official igraph 0.9 releases will not work as these releases still contain the spectral coarse graining functions so the two libraries will conflict with each other.
Assuming that you have installed the appropriate version of igraph in a system-wide location, you can configure and build this project like this:
mkdir build && cd build
cmake ..
cmake --build .
cmake --install .If you have installed igraph in a non-standard location, you need to tell CMake where to look for it:
mkdir build && cd build
cmake -DCMAKE_PREFIX_PATH=/path/where/igraph/is/installed ..
cmake --build .
cmake --install .CMake builds a static library by default; if you need a shared library,
pass -DBUILD_SHARED_LIBS=ON as an argument to cmake.
You can also use this project as a subproject in another CMake project.
The documentation of the various functions are to be found in the source files. We do not build a separate HTML documentation for this library at the moment. If you are interested in helping out and making this repo Doxygen-compatible, let us know.
Feel free to open an issue if you think you have found a bug in this library, or if you think you can help us out with fixing the numerical instabilities of the unit tests. Note that the instabilities appear only with certain BLAS implementations on certain platforms, so if they work for you, it does not mean they are okay in general.
If you have general questions about the SCG methods, try asking around in the igraph Discourse group, but note that none of the current maintainers of igraph are familiar with spectral coarse graining, so chances are that you will need to wait for someone else to answer your questions.