This project uses spacy and a directed community algorithm to parse a text and return corresponding communities of words. Those communities are created thanks to the directed Louvain algorithm developed by Anthony Perez and Nicolas Dugué.
Spacy:
Spacy is used to parse the text you give as an argument base on the
pipeline of your choice.
Directed louvain:
To build communities of words we choose the use the directed louvain algorithm
base on the louvain algorithm.
This algorithm is based on the modularity formula and will provide accurate
communities in a short execution time.
This project was built on the directed louvain to determine if the direction has an
impact on the output result.
Dependancy:
You will need to install spacy and the pipeline of your choice
to use this project.
You can find more information on this link: https://spacy.io/usage
Compilation:
To use this project you should create a folder (for instance build) inside the
DirectedLouvain/python-binding folder. Inside this folder you will compile the code to produce a python library,
to do so you should execute the command:
cmake ../python-binding
then:
make
You should get a file with the extension .so .
Drag this file into the root of the project and you should be all set.
Usage:
You can use this project as stand-alone or as a module.
In both case, you can set the path to the text you want to parse and
the spacy pipeline to use.