junipertcy / bipartitesbm

A Bayesian model+algorithm for community detection in bipartite networks

Hi there 🐙

I am a postdoctoral scholar at the BioFrontiers Institute at CU Boulder. My current work is supported by The Kavli Foundation and will develop an analytical approach to infer functional connectivity from neurophysiological data. You can nerd-snipe me by network science, neuroscience, computational topology, and convex optimization!

For more info, please see my Google Scholar or CV.

bipartitesbm's Issues

Use the tempfile library

Use the tempfile STL for the *.tmp files generated during the algorithm. (A TODO for myself, again)

OSError: [Errno 7] Argument list too long

When pass strings to the mcmc binary, the argument length is limited by the system. One can check the value via getconf ARG_MAX (and that's usually 2621440, meaning 2MB of arguments).

This may point to a potential solution. Or, save the arguments as tempfiles.

Support Dask

Support Dask.future (https://docs.dask.org/en/latest/futures.html) as well, make the loky option as the default.

Re-structure the code

Re-structure the code in the form that will be introduced in the paper. (This is a TODO for myself)

cannot find 'boost/program_options.hpp'

bipartiteSBM/engines/bipartiteSBM-MCMC/src/mcmc_main.cc:17:10: fatal error: boost/program_options.hpp: No such file or directory
#include <boost/program_options.hpp>
^~~~~~~~~~~~~~~~~~~~~~~~~~~

add travis-ci for this project

as title

Questions about synthetic networks

Hi!

I noticed the "generate synthetic networks" section in the documents seems incomplete. And may I ask the reference for the "ers=pc/K+(1−p)(1−c)/K(K−1)", does "c" here refers to "c_out" or "c_in" or the sum of all edges "c_out+c_in"? Thanks!

Best,
Siwei

Replace pathos to other libraries

Currently, only Python2 is supported.

Install issue

Hi!
When i run the following command line
sh scripts/compile_engines.sh
I receive this error:
cripts/compile_engines.sh: 3: scripts/compile_engines.sh: pushd: not found
scripts/compile_engines.sh: 4: scripts/compile_engines.sh: pushd: not found

Can you explain me why?
Thank you

Error when #nodes ≈ #(probed-)groups

Thanks @erikj540, who found error with these inputs,

edgelist = [[0, 3], [0, 4], [0, 5], [1, 3], [1, 4], [1, 5], [2, 6], [2, 7], [2, 8]]
types =  [1, 1, 1, 2, 2, 2, 2, 2, 2]

Signs and symptoms:

AssertionError: [ERROR] inconsistency between mb. indexes and #blocks.

Pytest problem on linux

As title.

No module named 'biSBM'

Hello Yen,

I have sent you an email regarding the issue. In case you missed it, I post here also.

I followed the steps on the installation webpage and it seemed all successful. However, when I ran the test files, I got an error: No module named 'biSBM'. Do you have any idea when this happened and how could I fix it?

Thx a lot in advance!

I am a mac user.

problems at installation with blockmodel.cc

Here is the full output of sh scripts/compile_engines.sh

-- The C compiler identification is GNU 5.4.0
-- The CXX compiler identification is GNU 5.4.0
-- Check for working C compiler: /usr/bin/cc
-- Check for working C compiler: /usr/bin/cc - works
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Detecting C compile features
-- Detecting C compile features - done
-- Check for working CXX compiler: /usr/bin/c++
-- Check for working CXX compiler: /usr/bin/c++ - works
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Performing Test COMPILER_SUPPORTS_CXX14
-- Performing Test COMPILER_SUPPORTS_CXX14 - Success
-- Found Boost: /usr/include (found suitable version "1.58.0", minimum required is "1.40") found components: program_options
-- Performing Test HAVE_STEADY_CLOCK
-- Performing Test HAVE_STEADY_CLOCK -- success
-- Configuring done
-- Generating done
-- Build files have been written to: /home/olaf/Descargas/bipartiteSBM/engines/bipartiteSBM-MCMC
Scanning dependencies of target mcmc
[ 11%] Building CXX object src/CMakeFiles/mcmc.dir/mcmc_main.cc.o
[ 22%] Building CXX object src/CMakeFiles/mcmc.dir/metropolis_hasting.cc.o
[ 33%] Building CXX object src/CMakeFiles/mcmc.dir/output_functions.cc.o
[ 44%] Building CXX object src/CMakeFiles/mcmc.dir/graph_utilities.cc.o
[ 55%] Building CXX object src/CMakeFiles/mcmc.dir/blockmodel.cc.o
/home/olaf/Descargas/bipartiteSBM/engines/bipartiteSBM-MCMC/src/blockmodel.cc: In member function ‘void blockmodel_t::agg_merge(std::mt19937&, int, int)’:
/home/olaf/Descargas/bipartiteSBM/engines/bipartiteSBM-MCMC/src/blockmodel.cc:234:104: error: ‘blockmodel_t::bmove_’ is not a variable in clause ‘private’
/home/olaf/Descargas/bipartiteSBM/engines/bipartiteSBM-MCMC/src/blockmodel.cc:246:44: error: ‘blockmodel_t::bmoves_’ is not a variable in clause ‘shared’
src/CMakeFiles/mcmc.dir/build.make:134: fallo en las instrucciones para el objetivo 'src/CMakeFiles/mcmc.dir/blockmodel.cc.o'
make[2]: *** [src/CMakeFiles/mcmc.dir/blockmodel.cc.o] Error 1
CMakeFiles/Makefile2:113: fallo en las instrucciones para el objetivo 'src/CMakeFiles/mcmc.dir/all'
make[1]: *** [src/CMakeFiles/mcmc.dir/all] Error 2
Makefile:103: fallo en las instrucciones para el objetivo 'all'
make: *** [all] Error 2
scripts/compile_engines.sh: 4: cd: can't cd to engines/bipartiteSBM-KL/

Refactor the code

Try to improve the code via new technologies.
https://datawhatnow.com/things-you-are-probably-not-using-in-python-3-but-should/

When #nodes is very few, the algorithm can be buggy.

As title. This happens when testing with the Mexican drug trafficking dataset.