Saturday, October 24, 2015

Installing numpy and scipy for OSX - using Homebrew

The place I found for instructions on how best to do this, and it has been edited recently, is

I have copied the script from Jörn's page to here for my personal reference. But all credit is Jörn's 
I did need X11 installed and I thought this was possible via Homebrew but I had a problem so I resorted to downloading it from Quartz page. It took a while. It used to be distributed by Apple but no longer.

Notice that some files are downloaded from pip and not brew.

Also, I installed the brew install Caskroom/cask/mactex so that I could run matpltlib with latex fonts and formulas. This does take a while to download. Not sure why.

Possible errors when installing pip pyquery and lxml
You might find compile errors fatal error: 'libxml/xmlversion.h' file not found
You can google solutions to this and this link has a few approaches to solve the problem:

# install PIL, imagemagick, graphviz and other
# image generating stuff
brew install libtiff libjpeg webp little-cms2
pip install Pillow
brew install imagemagick --with-fftw --with-librsvg --with-x11
brew install graphviz --with-librsvg --with-x11
brew install cairo
brew install py2cairo # this will ask you to download xquartz and install it
brew install qt pyqt

# install virtualenv, nose (unittests & doctests on steroids)
pip install virtualenv
pip install nose

# install numpy and scipy
# there are two ways to install numpy and scipy now: via pip or via brew.
# PICK ONE, i prefer pip for proper virtualenv support and more up-to-date versions.
pip install numpy
pip install scipy
# OR:
# (if you want to run numpy and scipy with openblas also remove comments below:)
#brew install openblas
brew install numpy # --with-openblas
brew install scipy # --with-openblas

# test the numpy & scipy install
python -c 'import numpy ; numpy.test();'
python -c 'import scipy ; scipy.test();'

# some cool python libs (if you don't know them, look them up)
# matplotlib: generate plots
# pandas: time series stuff
# nltk: natural language toolkit
# sympy: symbolic maths in python
# q: fancy debugging output
# snakeviz: cool visualization of profiling output (aka what's taking so long?)
#brew install Caskroom/cask/mactex  # if you want to install matplotlib with tex support and don't have mactex installed already
brew install matplotlib --with-cairo --with-tex  # cairo: png ps pdf svg filetypes, tex: tex fonts & formula in plots
pip install pandas
pip install nltk
pip install sympy
pip install q
pip install snakeviz

# ipython with parallel and notebook support
brew install zmq
pip install ipython[all]

# html stuff (parsing)
pip install html5lib cssselect pyquery lxml BeautifulSoup

# webapps / apis (choose what you like)
pip install Flask Django tornado

# semantic web stuff: rdf & sparql
pip install rdflib SPARQLWrapper

# graphs (graph metrics, social network analysis, layouting)
pip install networkx
brew install graph-tool

# maintenance: updating pip libs
pip install pip-tools  # you'll then have a pip-review command, see Updating section below