Setting Up A Python Environment Advanced

Aims:

The python environments then need populating with libraries etc. This is the place to get some standard setups

URLs used:

http://www.thisisthegreenroom.com/2011/installing-python-numpy-scipy-matplotlib-and-ipython-on-lion/

TL;DR

use conventions and regular naming for your virtual environments
- prefixes
  - p2_ for python2
  - p3_ for python3

Useful things:

List the virtual environments

lsvirtualenv

Work within a specific virtual environment

workon <virtualenvironment name> #Use a specific virtual environment

Stop using a specific virtual environment

deactivate #stop using the current virtual environment

List what version of python and which libraries are installed

python -V
help('modules')
pip freeze
#for all currently used modules (useful if looking to re-build a bespoke environment)
import sys as s
print s.modules.keys()

Installing python library schemes

workon <virtualenvironment name> #see list using lsvirtualenv
pip install <library name> #copy and paste library schemes from below
deactivate

I recently moved back to ubuntu (or mint at least) and not everything installed properly. Foe GDAL the following helped:

export CPLUS_INCLUDE_PATH=/usr/include/gdal
export C_INCLUDE_PATH=/usr/include/gdal

Python library schemes

These workin both Python 2 and 3. Specific python 2 dependencies are noted.

For Python2 always include

pip install future

Amongst other things this provides consitency with the print statement for python3 using:

from __future__ import print_function

Pre-requisites

You'll need a fortran compiler for scipy

sudo pacman -S gcc-fortan

Data Science

pip install numpy
pip install scipy
pip install pandas
pip install matplotlib
pip install tornado
pip install pyzmq
pip install ipython[all]

Firing up ipython-notebook

ipython notebook --pylab=inline #remember to check the port number if, like me, you automatically start up a default ipython-notebook instance

Stats

pip install patsy
pip install statsmodels # [not working in python3](http://stackoverflow.com/questions/23343484/python-3-statsmodels)
pip install sympy

Semantic Web

pip install rdflib
pip install SPARQLWrapper
pip install FuXi #not working - google this

Code improvement

pip install pyflakes
pip install pylint
pip install rope

Connecting to datasources

pip install psycopg2 #postgres
pip install urllib3 #import data from URLs

Generically useful libraries

pip install datetime
pip install simplejson
pip install zipfile # not installing in python 2 or 3

Geo

pip install folium
pip install gdal
pip install pyproj #projection and reprojection library
    pip install basemap  # not installing in python 2 or 3 try: [compiling by hand](https://github.com/matplotlib/basemap)
    pip install pyshp
    pip install geojson
pip install pyproj basemap gdal shapely fiona
pip install Descartes geoJSON psycopg2 SQLAlchemy
[Cartopy needs compiling](https://github.com/SciTools/cartopy)

Computer vision

pip install pil # not installing in python 2 or 3 TRY: pip install --allow-external pil --allow-unverified pil pil

Visualisation

This is mainly to play about with bokeh. The dependencies for this are, unfortunately, mixed.

Had trouble installing bokeh. Bottom this out later

Dependencies

pip install flask Redis Requests gevent gevent-websocket
pip install bokeh

DataFormats

pip install netCDF4
pip install pyyaml

Datbase querying

pip install sqlalchemy

Written on January 6, 2014