Link Recomendation in GitHub using a variant of supervised random walks
This is the CS224W Fall 2013 project on link recommendation in GitHub.
To create data -
- cd db
- Run sqlite3 test.db < create.sql
To load data into db
- python load.py -date1:2012-10-04-00 -date2:2012-04-12 -db:db/test.db -stream:online
To generate edge lists
- cd analysis
- ./gengraphs.sh ../db/test.db false
Setting the second argument to true above will also plot the graphs
To get graph properties Prereq - Must have edge lists ready - See how to generate edge lists
- python properties.py ../data/collaborators.out
PREREQS - Installation
This project depends on the following modules. Their installation instructions are also presented below.
-
Install Numpy git clone git://github.com/numpy/numpy.git numpy cd numpy python setup.py install --user
-
Install SciPy wget "http://downloads.sourceforge.net/project/scipy/scipy/0.13.1/scipy-0.13.1.tar.gz?r=&ts=1386460361&use_mirror=iweb"_ cd download python setup.py install --user
-
Install mlabwrap from BitBucket (latest source) wget https://bitbucket.org/nikratio/mlabwrap/get/d61f299e7887.zip unzip d61f299e7887.zip cd nikratio-mlabwrap-d61f299e7887 python setup.py install --user