Sequential Stratified Regeneration: MCMC for Large State Spaces with an Application to Subgraph Counting Estimation

This project contains the framework used in the paper "Sequential Stratified Regeneration: MCMC for Large State Spaces with an Application to Subgraph Counting Estimation" available here.

Datasets

The datasets used in the paper are available in SNAP. Please, apply the script "parseEdgesListToGph.py" to put the graph in Ripple's format.

Getting Started

For Cmake instructions check below

Prerequisites

Besides g++ compiler (version 5.4), we need to install the follow libraries in order to compile Ripple:

libboost-all-dev
libtbb-dev
libgsl-dev
libsparsehash-dev

Installing

Run the follow command in the root directory:

makefile

Running the tests

The applications are available in the direction apps. For example, the graph pattern mining application which uses Ripple to count subgraphs is in the directory apps/ripple. To compile a particular application we need to execute the follow command in such application directory:

./compile.sh

Then, a executable (binary) will be generated and we can run the application and get a description of its parameters by executing the follow command line:

./app -h

Cmake Configuration

#####
# Modify sync.sh to add desired location and then sync
#####
./sync <remote folder name> <1 for full 0 for partial>

#####
# Configure
#####
# On Mac Local
brew install boost 
brew install tbb
brew install gsl
brew install google-sparsehash
# On a linux machine running `GCC 7+` create or update your conda environment using
# First time
conda env create --file ripple.yml
# Checking Environment
conda env update --file ripple.yml  --prune

#####
# Running Live
#####
./compile.sh 
./app -i data/yeastbigcomp-sl.gph -o out.txt -k 6 -t 10 -c ./apps/matrioska/config.txt 

## License

This project is licensed under the GNU LGPL.