Infomap is a network clustering algorithm based on the Map equation.

Hi, I'm a little confused about the effect of initial_partition of infomap.

I set the initial_partition for partial nodes folloing the example does in this link and it indeed improves the clustering performance.

There is a note in that link, i.e. 'The initial partition is saved between runs. If you want to use an initial partition for one run only, use run(initial_partition=partition)'. In my understanding, if I set initial partition with im.initial_partition = {1: 0, 2:0}, the final module ID of node1 and node2 will be kept same, right?

However, in my experiment, I found some module ID of nodes, which belongs to the same initial partition, are not same in the final results. For example, providing the initial partitions 1: 0, 2: 0, 3: 0, 4: 0, 5: 1, 6: 1, 7: 1 to im.initial_partition, the final module ID of 1, 2, 3 is 123, while the one of 4, 5 is 78 (different from 123). Of course, the frequency of such changing is not high.

Did I take a wrong way to provide the initial partitions, which results in the inconsistency? Or, maybe the initial_partition property cannot guarantee that the partition does not change?

Hello! I am using a windows 10 machine and I am trying to build infomap for python. I used both anaconda and the windows sub machine for Ubuntu and I used the command "pip install infomap" and "pip install --upgrade infomap", both give me the exact same error : failed to build the wheel. I am not sure the reason why, I have gcc installed and made sure that it is in the system variable. Would you mind explaining why this happened? Thank you!!

Hi, I have a undirected, fully-connected, weighted network with 264 nodes. The edge weights represent weights (importance of links) and not the distance between nodes. I ran the algorithm, but it gives me 264 clusters even after 100 trials. I am attaching the adjacency matrix for your consideration. Can you pinpoint what the issue might be?

averaged_matrix_weighted.csv.zip

My OS is Ubuntu 18.04 The g++ and make version as below

make --version
GNU Make 4.1
Built for x86_64-pc-linux-gnu

g++ --version
g++ (Ubuntu 7.5.0-3ubuntu1~18.04) 7.5.0

I was trying to make cpp examples to work

unzip infomap-1.9.0.zip
cd infomap-1.9.0/examples/cpp/minimal
make

Make the static library

make -C ../../.. lib
make[1]: Entering directory '/home/jeff/Downloads/infomap-1.9.0'
g++ -Wall -Wextra -Wno-unused-parameter -std=c++14 -O4 -fopenmp -DNS_INFOMAP -DAS_LIB -c src/io/version.cpp -o build/lib/io/version.o
g++ -Wall -Wextra -Wno-unused-parameter -std=c++14 -O4 -fopenmp -DNS_INFOMAP -DAS_LIB -c src/io/ProgramInterface.cpp -o build/lib/io/ProgramInterface.o
g++ -Wall -Wextra -Wno-unused-parameter -std=c++14 -O4 -fopenmp -DNS_INFOMAP -DAS_LIB -c src/io/Config.cpp -o build/lib/io/Config.o
g++ -Wall -Wextra -Wno-unused-parameter -std=c++14 -O4 -fopenmp -DNS_INFOMAP -DAS_LIB -c src/io/Network.cpp -o build/lib/io/Network.o
g++ -Wall -Wextra -Wno-unused-parameter -std=c++14 -O4 -fopenmp -DNS_INFOMAP -DAS_LIB -c src/Infomap.cpp -o build/lib/Infomap.o
g++ -Wall -Wextra -Wno-unused-parameter -std=c++14 -O4 -fopenmp -DNS_INFOMAP -DAS_LIB -c src/core/InfomapIterator.cpp -o build/lib/core/InfomapIterator.o
g++ -Wall -Wextra -Wno-unused-parameter -std=c++14 -O4 -fopenmp -DNS_INFOMAP -DAS_LIB -c src/core/ClusterMap.cpp -o build/lib/core/ClusterMap.o
g++ -Wall -Wextra -Wno-unused-parameter -std=c++14 -O4 -fopenmp -DNS_INFOMAP -DAS_LIB -c src/core/MapEquation.cpp -o build/lib/core/MapEquation.o
g++ -Wall -Wextra -Wno-unused-parameter -std=c++14 -O4 -fopenmp -DNS_INFOMAP -DAS_LIB -c src/core/InfomapBase.cpp -o build/lib/core/InfomapBase.o
g++ -Wall -Wextra -Wno-unused-parameter -std=c++14 -O4 -fopenmp -DNS_INFOMAP -DAS_LIB -c src/core/MetaMapEquation.cpp -o build/lib/core/MetaMapEquation.o
g++ -Wall -Wextra -Wno-unused-parameter -std=c++14 -O4 -fopenmp -DNS_INFOMAP -DAS_LIB -c src/core/BiasedMapEquation.cpp -o build/lib/core/BiasedMapEquation.o
g++ -Wall -Wextra -Wno-unused-parameter -std=c++14 -O4 -fopenmp -DNS_INFOMAP -DAS_LIB -c src/core/MemMapEquation.cpp -o build/lib/core/MemMapEquation.o
g++ -Wall -Wextra -Wno-unused-parameter -std=c++14 -O4 -fopenmp -DNS_INFOMAP -DAS_LIB -c src/core/StateNetwork.cpp -o build/lib/core/StateNetwork.o
g++ -Wall -Wextra -Wno-unused-parameter -std=c++14 -O4 -fopenmp -DNS_INFOMAP -DAS_LIB -c src/core/InfoNode.cpp -o build/lib/core/InfoNode.o
g++ -Wall -Wextra -Wno-unused-parameter -std=c++14 -O4 -fopenmp -DNS_INFOMAP -DAS_LIB -c src/utils/Log.cpp -o build/lib/utils/Log.o
g++ -Wall -Wextra -Wno-unused-parameter -std=c++14 -O4 -fopenmp -DNS_INFOMAP -DAS_LIB -c src/utils/FileURI.cpp -o build/lib/utils/FileURI.o
g++ -Wall -Wextra -Wno-unused-parameter -std=c++14 -O4 -fopenmp -DNS_INFOMAP -DAS_LIB -c src/utils/FlowCalculator.cpp -o build/lib/utils/FlowCalculator.o
Creating static library...
ar rcs lib/libInfomap.a build/lib/io/version.o build/lib/io/ProgramInterface.o build/lib/io/Config.o build/lib/io/Network.o build/lib/Infomap.o build/lib/core/InfomapIterator.o build/lib/core/ClusterMap.o build/lib/core/MapEquation.o build/lib/core/InfomapBase.o build/lib/core/MetaMapEquation.o build/lib/core/BiasedMapEquation.o build/lib/core/MemMapEquation.o build/lib/core/StateNetwork.o build/lib/core/InfoNode.o build/lib/utils/Log.o build/lib/utils/FileURI.o build/lib/utils/FlowCalculator.o Makefile
Wrote static library to lib/ and headers to include/
make[1]: Leaving directory '/home/jeff/Downloads/infomap-1.9.0'

The building and error message as below

g++ -Wall -O3 -std=c++14 -DNS_INFOMAP example.cpp -o example -I../../../include -L../../../lib -lInfomap
In file included from ../../../include/core/InfomapCore.h:17:0,
                 from ../../../include/Infomap.h:13,
                 from example.cpp:28:
../../../include/core/InfomapOptimizer.h:749:0: warning: ignoring #pragma omp parallel [-Wunknown-pragmas]
 #pragma omp parallel for schedule(dynamic) // Use dynamic scheduling as some threads could end early
 
../../../include/core/InfomapOptimizer.h:899:0: warning: ignoring #pragma omp critical [-Wunknown-pragmas]
 #pragma omp critical(moveUpdate)
 
example.cpp:30:29: error: variable or field ‘printClusters’ declared void
 void printClusters(infomap::Infomap& infomap)
                             ^~~~~~~
example.cpp:30:29: error: ‘Infomap’ is not a member of ‘infomap’
example.cpp:30:29: note: suggested alternative: ‘InfomapRun’
 void printClusters(infomap::Infomap& infomap)
                             ^~~~~~~
                             InfomapRun
example.cpp:30:45: error: expected primary-expression before ‘)’ token
 void printClusters(infomap::Infomap& infomap)
                                             ^
example.cpp: In function ‘int main(int, char**)’:
example.cpp:45:11: error: ‘Infomap’ is not a member of ‘infomap’
  infomap::Infomap infomapWrapper("--two-level -N2");
           ^~~~~~~
example.cpp:45:11: note: suggested alternative: ‘InfomapRun’
  infomap::Infomap infomapWrapper("--two-level -N2");
           ^~~~~~~
           InfomapRun
example.cpp:47:18: error: ‘infomapWrapper’ was not declared in this scope
  auto& network = infomapWrapper.network();
                  ^~~~~~~~~~~~~~
example.cpp:66:2: error: ‘printClusters’ was not declared in this scope
  printClusters(infomapWrapper);
  ^~~~~~~~~~~~~
Makefile:10: recipe for target 'example' failed
make: *** [example] Error 1

Is the infomap package on PyPi built properly?

Complete output (5 lines): Traceback (most recent call last): File "", line 1, in File "C:\Users\alexc\AppData\Local\Temp\pip-install-bw0bw2ol\infomap\setup.py", line 10, in import package_meta ModuleNotFoundError: No module named 'package_meta'

Hi,

I'm using this package as a dependency to another one I have forked. I'm trying to build a conda release for it and being able to install infomap through conda would help me.

I'll be attempting one myself but if others are keen, it would be extremely helpful.

Thank you for any assistance you can provide,

V

Describe the bug I am curious to know if the time complexity of the v1/grassberger branches is to blame for the extremely long projected run times for larger graphs (>100k nodes, millions of edges).

Using the master branch the graph in question is successfully clustered in around 30 minutes using 10 trials. However, for v1/grassberger-jelena, I aborted the clustering run after accruing 8 hours without the first trial being completed.

This was for a undirected, weighted graph defined in link-list format. The graph is highly modular, with perhaps 300 communities which become increasing sparse as subject become less frequently observed. (microbial community data)

I did not employ inner loop parallelization in either case.

To Reproduce

./Infomap edges.in -v -u -z -i link-list out_dir

Desktop (please complete the following information):

• OS: Linux, kernel > 5.2 (Fedora or Ubuntu)
• Version v1/grassberger-jelena
• Compiler: Gcc >7

hi - I am trying to use the pip installed version of the python package, but the code from the jupyter notebook doesn't seem to work. First, the initial import fails:

In [6]: from infomap import infomap
   ...:
---------------------------------------------------------------------------
ImportError                               Traceback (most recent call last)
<ipython-input-6-0c8597e1552c> in <module>()
----> 1 from infomap import infomap

ImportError: cannot import name 'infomap'

However, it works with from infomap import Infomap

The bigger problem is that the addLink function seems to be missing from Infomap:

In [7]: infomapWrapper = infomap.Infomap("--two-level")
   ...:
   ...: # Add link weight as an optional third argument
   ...: infomapWrapper.addLink(0, 1)
   ...:
   ...:
---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
<ipython-input-7-3febb79f95d2> in <module>()
      2
      3 # Add link weight as an optional third argument
----> 4 infomapWrapper.addLink(0, 1)

~/anaconda3/envs/py3/lib/python3.6/site-packages/infomap.py in <lambda>(self, name)
   2834     for _s in [InfomapBase]:
   2835         __swig_getmethods__.update(getattr(_s, '__swig_getmethods__', {}))
-> 2836     __getattr__ = lambda self, name: _swig_getattr(self, Infomap, name)
   2837     __repr__ = _swig_repr
   2838

~/anaconda3/envs/py3/lib/python3.6/site-packages/infomap.py in _swig_getattr(self, class_type, name)
     78     if method:
     79         return method(self)
---> 80     raise AttributeError("'%s' object has no attribute '%s'" % (class_type.__name__, name))
     81
     82

AttributeError: 'Infomap' object has no attribute 'addLink'

This is also evident from a listing of the object attributes. Has the API changed for the wrapper?

Hi, I download this tools and can make successful, But when I enter examples/python, and make some error come out,Here is the error: ./infomap_wrap.cpp:2452: error: ‘PyCObject_Import’ was not declared in this scope ./infomap_wrap.cpp: In function ‘void SWIG_Python_SetModule(swig_module_info*)’: ./infomap_wrap.cpp:2521: error: ‘PyCObject_FromVoidPtr’ was not declared in this scope ./infomap_wrap.cpp:2512: warning: unused variable ‘swig_empty_runtime_method_table’ ./infomap_wrap.cpp: In function ‘swig_type_info* SWIG_Python_TypeQuery(const char*)’: ./infomap_wrap.cpp:2544: error: ‘PyCObject_AsVoidPtr’ was not declared in this scope ./infomap_wrap.cpp:2549: error: ‘PyCObject_FromVoidPtr’ was not declared in this scope ./infomap_wrap.cpp: In function ‘std::vector<std::basic_string<char, std::char_traits, std::allocator >, std::allocator<std::basic_string<char, std::char_traits, std::allocator > > >* std_vector_Sl_std_string_Sg____getitem____SWIG_0(std::vector<std::basic_string<char, std::char_traits, std::allocator >, std::allocator<std::basic_string<char, std::char_traits, std::allocator > > >, PySliceObject)’: ./infomap_wrap.cpp:4462: error: cannot convert ‘PySliceObject*’ to ‘PyObject*’ for argument ‘1’ to ‘int PySlice_GetIndices(PyObject*, Py_ssize_t, Py_ssize_t*, Py_ssize_t*, Py_ssize_t*)’ ./infomap_wrap.cpp: In function ‘void std_vector_Sl_std_string_Sg____setitem____SWIG_0(std::vector<std::basic_string<char, std::char_traits, std::allocator >, std::allocator<std::basic_string<char, std::char_traits, std::allocator > > >, PySliceObject, const std::vector<std::basic_string<char, std::char_traits, std::allocator >, std::allocator<std::basic_string<char, std::char_traits, std::allocator > > >&)’: ./infomap_wrap.cpp:4471: error: cannot convert ‘PySliceObject*’ to ‘PyObject*’ for argument ‘1’ to ‘int PySlice_GetIndices(PyObject*, Py_ssize_t, Py_ssize_t*, Py_ssize_t*, Py_ssize_t*)’ ./infomap_wrap.cpp: In function ‘void std_vector_Sl_std_string_Sg____delitem____SWIG_1(std::vector<std::basic_string<char, std::char_traits, std::allocator >, std::allocator<std::basic_string<char, std::char_traits, std::allocator > > >, PySliceObject)’: ./infomap_wrap.cpp:4480: error: cannot convert ‘PySliceObject*’ to ‘PyObject*’ for argument ‘1’ to ‘int PySlice_GetIndices(PyObject*, Py_ssize_t, Py_ssize_t*, Py_ssize_t*, Py_ssize_t*)’ ./infomap_wrap.cpp: In function ‘std::deque<unsigned int, std::allocator >* std_deque_Sl_unsigned_SS_int_Sg____getitem____SWIG_0(std::deque<unsigned int, std::allocator >, PySliceObject)’: ./infomap_wrap.cpp:5026: error: cannot convert ‘PySliceObject*’ to ‘PyObject*’ for argument ‘1’ to ‘int PySlice_GetIndices(PyObject*, Py_ssize_t, Py_ssize_t*, Py_ssize_t*, Py_ssize_t*)’ ./infomap_wrap.cpp: In function ‘void std_deque_Sl_unsigned_SS_int_Sg____setitem____SWIG_0(std::deque<unsigned int, std::allocator >, PySliceObject, const std::deque<unsigned int, std::allocator >&)’: ./infomap_wrap.cpp:5035: error: cannot convert ‘PySliceObject*’ to ‘PyObject*’ for argument ‘1’ to ‘int PySlice_GetIndices(PyObject*, Py_ssize_t, Py_ssize_t*, Py_ssize_t*, Py_ssize_t*)’ ./infomap_wrap.cpp: In function ‘void std_deque_Sl_unsigned_SS_int_Sg____delitem____SWIG_1(std::deque<unsigned int, std::allocator >, PySliceObject)’: ./infomap_wrap.cpp:5044: error: cannot convert ‘PySliceObject*’ to ‘PyObject*’ for argument ‘1’ to ‘int PySlice_GetIndices(PyObject*, Py_ssize_t, Py_ssize_t*, Py_ssize_t*, Py_ssize_t*)’ error: command 'gcc' failed with exit status 1 make[1]: *** [python] Error 1 make[1]: Leaving directory `/usr/local/infomap' My computer is centos, python version is 3.5.1, gcc 4.4.6, how to solve this problem?Thanks very much

Hi,

I've been messing around with examples in Python and managed to cause seg-fault. When running default example for NetworkX (example-networkx.py) but with --overlapping option this happens:

=======================================================
  Infomap v0.19.0 starts at [2017-06-10 16:00:52]
  -> Configuration: overlapping
                    two-level
  -> Use undirected flow and 2nd order Markov dynamics
=======================================================
Building Infomap network from a NetworkX graph...
Find communities with Infomap...
Finalizing network...
Segmentation fault (core dumped)

My GCC version: gcc (Ubuntu 6.2.0-5ubuntu12) 6.2.0 20161005

Cheers!

Good morning. I'm trying to run infomap on the online version. Since I didn't know how to get results with my work data, I have started with the well known and documented example of the "southern women".I intend to find clusters in this example and then move it into my own data, but I haven't managed to do so yet. I have created the example data of the sw.txt file to copy it to https://www.mapequation.org/infomap/ and I run it with all kind of parameters, for example -i bipartite --ftree --clu , but I have not achieved the expected result, which according to several consulted papers should be something similar to the attached sw.jgp file. I'm sure that something is wrong with the infomap execution process, but I can't find it, if you would be so kind to give me some clue of what is wrong I would be very grateful. Thank you very much in advance for your answers.

sw.txt

Replicate in python:

edges = [
    [1,2],
    [2,3],
    [3,1],
    [3,4],
    [4,5],
    [5,6],
    [6,4],
]
G = nx.DiGraph()
G.add_edges_from(edges)
py.find_communities(G, recorded_teleportation=True, num_trials=1) # gives index codelength 0.102
py.find_communities(G, recorded_teleportation=True, num_trials=2) # gives index codelength 0.251

My aim is to run the following command from Matlab. [failed, message] = system([infomapfolder '/Infomap-2.6.0/Infomap --clu -2 -s' num2str(randnum) ' -N' num2str(reps) ' ' pajekfilename ' ' pathstr]);

I am working on MacOS (M1 processor...) and I have been able to successfully install infomap with pip install infomap as signalled by the messages on the terminalInstallling collected packages: infomap and Successfully installed infomap-2.6.0

However, I am unable to import the package and I get this message:

---------------------------------------------------------------------------
ImportError                               Traceback (most recent call last)
Input In [1], in <cell line: 1>()
----> 1 import infomap

File ~/opt/anaconda3/lib/python3.9/site-packages/infomap.py:38, in <module>
     36     from . import _infomap
     37 else:
---> 38     import _infomap
     40 try:
     41     import builtins as __builtin__

ImportError: dlopen(/Users/matilde/opt/anaconda3/lib/python3.9/site-packages/_infomap.cpython-39-darwin.so, 0x0002): Library not loaded: /usr/local/opt/libomp/lib/libomp.dylib
  Referenced from: /Users/matilde/opt/anaconda3/lib/python3.9/site-packages/_infomap.cpython-39-darwin.so
  Reason: tried: '/usr/local/opt/libomp/lib/libomp.dylib' (no such file), '/usr/local/lib/libomp.dylib' (no such file), '/usr/lib/libomp.dylib' (no such file)

If the infomap was correctly installed, I could use pyrunfile from Matlab in order to run infomap.

I am able to import infomap when working on an old Mac. Hence, I strongly suspect the problem is related to M1 and I would welcome any advice on possible ways forward.

I tried as an alternative to compile from source with the following (always considering that I am on macOS Monterey, version 12.2.1)

git clone [email protected]:mapequation/infomap.git
cd infomap
make

However, I get an error here as well as

c++ -Wall -Wextra -pedantic -Wnon-virtual-dtor -std=c++14 -Wshadow -O3 -Xpreprocessor -fopenmp -c src/core/MetaMapEquation.cpp -o build/Infomap/core/MetaMapEquation.o
c++ -Wall -Wextra -pedantic -Wnon-virtual-dtor -std=c++14 -Wshadow -O3 -Xpreprocessor -fopenmp -c src/core/MemMapEquation.cpp -o build/Infomap/core/MemMapEquation.o
c++ -Wall -Wextra -pedantic -Wnon-virtual-dtor -std=c++14 -Wshadow -O3 -Xpreprocessor -fopenmp -c src/core/InfoEdge.cpp -o build/Infomap/core/InfoEdge.o
c++ -Wall -Wextra -pedantic -Wnon-virtual-dtor -std=c++14 -Wshadow -O3 -Xpreprocessor -fopenmp -c src/core/BiasedMapEquation.cpp -o build/Infomap/core/BiasedMapEquation.o
c++ -Wall -Wextra -pedantic -Wnon-virtual-dtor -std=c++14 -Wshadow -O3 -Xpreprocessor -fopenmp -c src/core/StateNetwork.cpp -o build/Infomap/core/StateNetwork.o
c++ -Wall -Wextra -pedantic -Wnon-virtual-dtor -std=c++14 -Wshadow -O3 -Xpreprocessor -fopenmp -c src/core/iterators/InfomapIterator.cpp -o build/Infomap/core/iterators/InfomapIterator.o
c++ -Wall -Wextra -pedantic -Wnon-virtual-dtor -std=c++14 -Wshadow -O3 -Xpreprocessor -fopenmp -c src/core/InfomapBase.cpp -o build/Infomap/core/InfomapBase.o
src/core/InfomapBase.cpp:35:10: fatal error: 'omp.h' file not found
#include <omp.h>
         ^~~~~~~
1 error generated.
make: *** [build/Infomap/core/InfomapBase.o] Error 1

Thank you, Matilde

Running -T1 -L 1 with --cluster-data doesn't run any optimization.

Unless --no-infomap, continue with optimization after initial partition before consolidation of first level.