Skip to content

Latest commit

 

History

History
56 lines (42 loc) · 2.51 KB

README.md

File metadata and controls

56 lines (42 loc) · 2.51 KB
Raw

Data Science VM

Vagrant + VirtualBox scripts for setting up an Ubuntu (32 bit) Data Science server. If you want to make changes or improvements, don't be shy; send me a note or a pull request.

This system takes a download-it-all approach. As a result, it can take quite a whilte to download everything and set the server up. Do adjust the Vagrantfile to focus only on the tools you need.

Installing Locally with VirtualBox

  • Install VirtualBox
  • Install Vagrant
  • Set up the Ubuntu Precise32 Vagrant Box vagrant box add precise32 http://files.vagrantup.com/precise32.box
  • Run vagrant up This should launch the box and download/install a suite of data science tools.

Installing on Amazon AWS

Vagrant can also launch, provision, and manage Amazon AWS instances. For more, read the vagrant-aws page.

  • vagrant plugin install vagrant-aws
  • Create a new AWS access key or note the location of your existing key
  • Create a 'dummy' vagrant box that will stand in for AWS vagrant box add awsubuntu https://github.com/mitchellh/vagrant-aws/raw/master/dummy.box
  • Change the base box in Vagrantfile.aws to 'awsubuntu'
  • Add your access ID, key name, and key file location to Vagrantfile.aws
  • Copy Vagrantfile.aws over Vagrantfile
  • Add port 22, 8888, and 8787 to the default security group on your AWS account, or create a new security group with SSH access and add it to the Vagrant file
  • To launch your VM, run vagrant up --provider=aws

Tools Included

  • Python NLTK
  • Python Stanford Core NLP Library, with Web API
  • iPython Notebook
  • R Studio
  • Python Scikit-learn
  • Vim
  • Screen

Accessing R Studio

Port forwarding is automatically set up on port 8787. If you're running Vagrant locally, load localhost:8787. If remotely, load http://SERVER_ADDRESS:8787 The default username and password match the default password of Vagrant.

iPython Notebook

The proxy is automatically set up on port 8888. run ipython notebook and access localhost:8888

Accessing Web Services

If you launch other servers, you need to set up port forwarding. Read this guide to [port forwarding on Vagrant] (http://docs-v1.vagrantup.com/v1/docs/getting-started/ports.html)

Running the Stanford CoreNLP Python Module

python corenlp-python/corenlp/corenlp.py