A.6. Windows
The command-line tools for package management, such as cygwin on Windows, aren’t that great. But if you install GitGUI on a Windows machine, that gets you a bash prompt and a workable terminal that you can use to run your Python REPL console:
- Download and install the git installer (https://git-scm.com/download/win).
- Download and install the GitHub Desktop (https://desktop.github.com).
The git installer comes with a version of the bash shell that should work well within Windows, but the git-gui that it installs isn’t very user friendly, especially for beginners. Unless you’re using git from the command line (a bash shell within Windows), you should use GitHub Desktop for all your git push/pull/merge needs on Windows. We had problems throughout the editing of this book when git-gui did unexpected things that overwrote commits by others whenever there was a version conflict, even in files that weren’t involved in the conflict. So that’s why we ask you to install GitHub Desktop (http://desktop.github.com) on top of raw git and git-bash. GitHub Desktop gives you a more user-friendly git experience, letting you know when you need to pull and push or merge some changes.[3]
Big thanks to Benjamin Berg and Darren Meiss at Manning for figuring this out, and for all the hard work they put into making this book presentable.
Once you have a shell running in a Windows terminal, you can install Anaconda and use the conda package manager to install the nlpia package just like the rest of us, using the instruction in the github repository README (http://github.com/totalgood/nlpia).
A.6.1. Get Virtual
If you get frustrated with Windows, you can always install VirtualBox or Docker and create a virtual machine with an Ubuntu OS. That’s the subject of a whole book (or at least a chapter), and there are better people at that than we are:
- Jason Brownlee (https://machinelearningmastery.com/linux-virtual-machine-machine-learning-development-python-3)
- Jeroen Janssens (http://datasciencetoolbox.org)
- Vik Paruchuri (www.dataquest.io/blog/docker-data-science)
- Jamie Hall (http://blog.kaggle.com/2016/02/05/how-to-get-started-with-data-science-in-containers)
Another way to get Linux into your Windows world is with Microsoft’s Ubuntu shell app. I’ve not used it, so I can’t vouch for its compatibility with the Python packages you’ll need to install. If you try it, share what you learn with us at the nlpia repository with a feature or pull request on the documentation (https://github.com/totalgood/nlpia/issues). The Manning NLPIA forum (https://forums.manning.com/forums/natural-language-processing-in-action) is also a great place to share your knowledge and get assistance.