A.1. Anaconda3

Python 3 has a lot of performance and expressiveness features that are handy for NLP. And the easiest way to install Python 3 on almost any system is to install Anaconda3 (https://www.anaconda.com/download). This has the added benefit of giving you a package and environment manager that can install a lot of problematic packages (such as matplotlib) on a wide range of problematic OSes (like Windows).

You can install the latest version of Anaconda and its conda package manager programmatically by running the code in the following listing.

Listing A.1. Install Anaconda3

$ OS=MacOSX  # or Linux or Windows
$ BITS=_64  # or '' for 32-bit
$ curl https://repo.anaconda.com/archive/ > tmp.html
$ FILENAME=$(grep -o -E -e "Anaconda3-[.0-9]+-$OS-
     x86$BITS.(sh|exe)" tmp.html | head -n 1)
$ curl "https://repo.anaconda.com/archive/$FILENAME" > install_anaconda
$ chmod +x install_anaconda
$ ./install_anaconda -b -p ~/Anaconda
$ export PATH="$HOME/Anaconda/bin:$PATH"
$ echo 'export PATH="$HOME/Anaconda/bin:$PATH"' >> ~/.bashrc
$ echo 'export PATH="$HOME/Anaconda/bin:$PATH"' >> ~/.bash_profile
$ source ~/.bash_profile
$ rm install_anaconda

Now you can create a virtual environment: not a Python virtualenv but a more complete conda environment that isolates all of Python’s binary dependencies from your OS Python environment. Then you can install the dependencies and source code for NLPIA within that conda environment with listing A.2.