README.md

Graph Data Generation and Load into Neo4J

Pre-requisites

• JDK 8

  • Install JDK 1.8 from Oracle
  • Set JAVA_HOME On MacOS, you can do:

     export JAVA_HOME="$(/usr/libexec/java_home -v 1.8)"
    

• Python 3

  • On MacOS install with

     brew install python3
    

• Install Neo4J

  • On MacOS install with

     brew install neo4j
    

  • Setup NEO4J_HOME Point NEO4J_HOME to the root of the NEO4J directory.

     export NEO4J_HOME=/usr/local/Cellar/neo4j/3.2.2/libexec
    

Generate Data

• Clone the repository

   git clone https://github.com/nmdp-bioinformatics/grimm`
   cd grimm
  

• Setup Python3 virtual environment Make sure virtualenv is installed.

  pip3 install virtualenv
  

  Create Virtual Environment

  virtualenv -p python3 venv
  source venv/bin/activate
  

  Install pandas library

  pip3 install pandas
  

• Download and prepare wmda data. Python script downloads reference wmda data and untars it in wmda directory

   cd graph_generator/data
   python wmda_download.py
  

• Generate nodes/edges/toplinks from the reference wmda data. The freqs file is converted to HPF format first.

   cd ..
   python wmda_to_hpf_csv.py
   python generate_neo4j_wmda_hpf.py
  

• Generated nodes and edges files are in the output/csv directory.

   output
   └── csv
       ├── edges.csv
       ├── nodes.csv
       └── top_links.csv
  

• Load the nodes/graph into Neo4J database

   ./bulk_load_neo4j.sh
  

• Login to Neo4j http://localhost:7474/

  • Default username/password is neo4j/ontological