Skip to content

Instantly share code, notes, and snippets.

@patriciogonzalezvivo
Created October 7, 2014 23:17
Show Gist options
  • Save patriciogonzalezvivo/77da993b14a48753efda to your computer and use it in GitHub Desktop.
Save patriciogonzalezvivo/77da993b14a48753efda to your computer and use it in GitHub Desktop.
How to install Python correctly on Mac OSX

Install Homebrew

ruby -e "$(curl -fsSL https://raw.github.com/mxcl/homebrew/go)"

Add PATH to ~/.bash_profile and ~/.zshrc

export PATH=/usr/local/bin:$PATH

Install Python

brew install python

Add PATH to ~/.bash_profile and ~/.zshrc

export PATH=/usr/local/share/python:$PATH

Install helpful modules

pip install virtualenv
pip install virtualenvwrapper
pip install numpy
brew install gfortran
pip install scipy
brew install freetype
pip install matplotlib
pip install ipython[all]

QT for Python

For QT integration you need to download QT SDK and then PyQT

brew install pyqt

After installing pyqt, Homebrew will prompt you to add the following to your .bash_profile:

export PYTHONPATH=/usr/local/lib/python:$PYTHONPATH

Keep installing dependences for PyQT

brew install zmq
pip install pyzmq
pip install pygments

Setting up Sublime

  1. Download and Install Sublime

  2. Edit Python paths.

    • Sublime >> Preferences... >> Browse Packages...
    • Click on Python/Python.sublime-build
  3. Install PackageControl. The simplest method of installation is through the Sublime Text console. The console is accessed via the ctrl+shortcut or theView > Show Console menu`. Once open, paste the appropriate Python code for your version of Sublime Text into the console.

import urllib2,os,hashlib; h = '7183a2d3e96f11eeadd761d777e62404' + 'e330c659d4bb41d3bdf022e94cab3cd0'; pf = 'Package Control.sublime-package'; ipp = sublime.installed_packages_path(); os.makedirs( ipp ) if not os.path.exists(ipp) else None; urllib2.install_opener( urllib2.build_opener( urllib2.ProxyHandler()) ); by = urllib2.urlopen( 'http://sublime.wbond.net/' + pf.replace(' ', '%20')).read(); dh = hashlib.sha256(by).hexdigest(); open( os.path.join( ipp, pf), 'wb' ).write(by) if dh == h else None; print('Error validating download (got %s instead of %s), please try manual install' % (dh, h) if dh != h else 'Please restart Sublime Text to finish installation')
  1. Modify the code to point to the right brew paths. Ex:
{
	"env":{
		"PATH":"/usr/bin:/bin:/usr/sbin:/sbin:/opt/X11/bin:/usr/local/bin:/usr/local/share/python",
		"PYTHONPATH":"/usr/local/lib/python:/usr/local/Cellar/python/2.7.2/lib/python2.7/site-packages"
		},
	"cmd": ["/usr/local/bin/python", "-u", "$file"],
	"file_regex": "^[ ]*File \"(...*?)\", line ([0-9]*)",
	"selector": "source.python"
}

Python Autocompletion with SublimeCodeIntel

Once you install Package Control, restart ST2 and bring up the Command Palette (Command+Shift+P on OS X, Control+Shift+P on Linux/Windows). Select "Package Control: Install Package", wait while Package Control fetches the latest package list, then select SublimeCodeIntel when the list appears. The advantage of using this method is that Package Control will automatically keep SublimeCodeIntel up to date with the latest version.

@CameronHudson8
Copy link

After adding a path to ~/.bash_profile, you have to run source ~/.bash_profile in the terminal. Otherwise, the change doesn't register.

Copy link

ghost commented Jun 22, 2018

Hi, I have already Python 2.7 & 3.5 on my Mac. I have been writing python programs using the IDLE.
I now have need to install 'Homebrew' - I didn't previously know about 'Homebrew' as I am new to programming.
When I try to install 'Homebrew' on the command line it fails.
When I run 'brew doctor' I get the following warnings:

Warning: "config" scripts exist outside your system or Homebrew directories.
./configure scripts often look for *-config scripts to determine if
software packages are installed, and what additional flags to use when
compiling and linking.

Having additional scripts in your path can confuse software installed via
Homebrew if the config script overrides a system or Homebrew provided
script of the same name. We found the following "config" scripts:
/Library/Frameworks/Python.framework/Versions/3.5/bin/python3.5m-config
/Library/Frameworks/Python.framework/Versions/3.5/bin/python3.5-config
/Library/Frameworks/Python.framework/Versions/3.5/bin/python3-config
/System/Library/Frameworks/Python.framework/Versions/2.7/bin/python2-config
/System/Library/Frameworks/Python.framework/Versions/2.7/bin/python2.7-config
/System/Library/Frameworks/Python.framework/Versions/2.7/bin/python-config

If I uninstall both versions of Python and then reinstall using 'Homebrew' I will lose months of work, is there another way around this?

Thanks.

@1nk1
Copy link

1nk1 commented Aug 2, 2018

Hello @patriciogonzalezvivo,

because of Homebrew/homebrew-science#1291 (comment) you should replace brew install gfortran with brew install gcc.

~ brew install gfortran
Error: No available formula for gfortran
GNU Fortran is now provided as part of GCC, and can be installed with:
brew install gcc
Happy coding,
Darius

Maybe you try this:
brew cask install gfortran

@HansZhang
Copy link

As of now, this does not work anymore. On macOS with Homebrew you have to set the path as follows:

export PATH="/usr/local/opt/python/libexec/bin:$PATH"

Otherwise the system python will come first. Check brew info python for the caveats.

you saved my day, thanks!

@eo1989
Copy link

eo1989 commented Oct 14, 2019

https://gist.github.com/patriciogonzalezvivo/77da993b14a48753efda#gistcomment-2627029

I also have this problem. Not sure how to get around this without having to uninstall everything and reinstall all the data sci packages. I have a feeling miniconda/anaconda messed this up. Anyway I can get miniconda and homebrew to play nice together?

@lazarillo
Copy link

Summary:

What's written below is a helpful way to install Python (+ others) on a Mac if you are doing analytics/data science.


FWIW, what is written by the OP is the "right way" for non-data scientists. If you are using Python for analytics, I recommend using Anaconda (specifically Miniconda) for as much as you can, instead of Pip.

There is nothing wrong with Pip, and in fact it contains far more Python packages than Conda. However, Conda (a) primarily focuses on numerical packages, (2) it can install things outside of Python (which is a real blessing when the Python package you are installing is "just a wrapper" around lower-level C or even FORTRAN packages, and (3) it does a fairly decent job of trying to find the minimally harmful clashes between versions of packages you have installed. Most data scientists, myself included, get kinda sloppy about this.

The "right way" to avoid clashes is to make a separate virtual environment for each different thing that you're working on, but as data scientists, we're often working with packages only as long as it takes us to realize we didn't like the results... but we might like that package in the future. Hence, clutter.


Inevitably, conda will not have things that you want to install. In that case, fall back on pip, but be sure to go into a conda environment before running pip, so that you install into the same place you have put all of your Python packages installed from Conda.

Unfortunately for us analysts using Macs, we have to rely upon brew, pip, and conda for installs. (And sometimes just raw downloads: I have had trouble trying to use R with conda or virtualenv.) It's a hassle, and it doesn't always work out, but my priority order is conda, then pip, then brew, then raw download. brew should not be needed for any Python packages, nor should a raw download.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment