Python for Climate Science

1. Introduction

• Python tutorials on the web
  • Geo232 web pages and lab manual
  • CDAT pages: Introduction to Python
  • Konrad Hinsen's Python Pages
• General and more advanced Python references
  • python.org., providing link to many Python resources
  • Paul Dubois' indispensible documentation of the Numeric and Masked Array modules
• Books on Python

2. Use of the basic constructs:Programming Python with Style and Flair

• Modules and the Import statement
  • Fundamental role of modules
  • Every Python script is a module
  • Store your functions in a module! It's not like Matlab.
  • from ... import *. import ... as ...
  • Name-spaces. Nested modules
  • Example: A table of physical constants
  • Getting information about modules and objects: dir( ) and help( )
• Lists,oh, the things you can do with lists!
  • Building lists using append()
  • Building lists using functions
  • Useful list operations (methods)
  • Accessing an element or range from a list
• Loops: for,while,continue,break
  • Looping over lists
  • The range( ) and xrange( ) statements
• Define lots of functions
  • Function definition and the argument list
  • A warning on side-affecting arguments
• Program Development Techniques
  • Use modules to organize objects and functions
  • Use Idle, with a scratchpad for new functions/classes
  • Use the interpreter to test functions
• Ways to run Python scripts
  • From inside interpreter
    • Import a module (not recommended)
    • Import a module then invoke functions
    • Using the execute statement
  • From the command line
    • Use the Python statement
    • The pound-bang hack
• A warning on reference vs. identity
• Strings, oh the things you can do with strings!
  • String operations and methods
  • Search,replace,split, etc.
  • Using format strings
    • Getting a fixed length field ('%0<n>i'%N)
  • The string module. string.atoi( ), etc.
  • Newline, <cr> and tab codes
• Objects and their uses
  • Fundamental importance of the index methods __getitem( )__ and __setitem( )__
• Input and output
  • Input from the terminal: raw_input( )
  • Files
    • read( ),readline( ),readlines( )
    • write( )
    • seek( )
  • Example: Processing a space-delimited ascii data file
  • Example: Converting a Unix text file to Mac format crfix1.py
  • Interaction with c files: the struct module
    • Handling endian-ness
  • A note on Fortran files
• Dictionaries and their uses
  • Tip: Dictionaries are nice to use with __getitem( )__ and __setitem( )
• Exception handling
  • try/except blocks
  • raise (Also use raise to bail out of a loop or script)
  • use of assert statement in debugging

3. Python as a scripting language

• Your friend, the os module
• Generic directory and path commands
• glob.glob( )
• Executing arbitrary system commands os.system( )
• Example: Pruning restart files.
• Example: Converting a directory of history files to netCDF

4. Python for simulation

• A few examples in unadorned Python
  • Chaos in the Logistic map: The Logistic Map Explorer class
  • A 1-dimensional Runge-Kutta solver, with applications
  • Orbits of the planets
• Introduction to Numeric arrays
  • Creating arrays
  • Array arithmetic
  • Gotcha #1: Reference vs copy
  • Gotcha#2: Promotion from float to double
  • Array cross sections
• Example: The 1D diffusion equation
  • The model--Explicit stepping version
  • The model--Implicit version
  • A more interesting boundary condition: Radiative cooling from the top
  • A simple sea-ice model
  • Hey there Snoopy, don't be loopy! A short polemic on program design. Reasons for writing code in a "non-loopy" way. (Efficiency, but it also makes it easier to do automatic parallelization, since high-level arithmetic on objects is a signal of parallelism. Also encourages dealing with higher-levels of abstractions, making code more readable and maintainable.)
• Use of Masked arrays
• Extending Python with c and SWIG
  • Introduction to SWIG
  • A note on building Swigged modules under Mac OSX (Darwin)
  • Finding the right level for building objects
  • Global variables
  • Interacting with c arrays
  • Interacting with c structures
  • Example 1:
  • Example 2: A simple float array extension
  • Example 3: The diffusion equation
• Working with Fortran
  • General remarks on c and Fortran interaction
  • Example: Writing a c stub for a Fortran routine
  • Dealing with Fortran arrays
  • Dealing with Fortran common blocks
• Manipulating Numeric arrays with Fortran or c

5. Python for Data Analysis

• The cdms module
  • Overview
  • Variables,file variables, and Masked Variables
  • Reading and writing netCDF files
    • Read data from a netCDF file, and find information about what's in the file
    • Read data with a scale and add_offset
    • Write a variable to a netCDF file
    • Modifying a file by writing to a file variable
  • About axes
    • Getting information about axes
    • Specifying a variable's axes
    • Creating a new axis
  • Averaging over an axis
• Plotting data using vcs (needs Python 2.2)
  • Making a basic contour plot.
    • Customization: Continents, contour values,etc
    • Choosing a plot method (isofill, etc)
    • Map projections
  • Line plots
  • Saving plots to a graphics file
• Plotting data using ncl scripts