Courses :: BME205 :: Fall 2011, Section 01

BME205, Fall 2011, Section 01: Intro to Python

INTRO TO PYTHON LECTURE OUTLINE
BME205, Fall 2011
Andrew Uzilov
Contact: auzilov@ucsc.edu


(0) Disclaimers and Background

We will probably run out of time; can continue outside of class (TA office hours?).

Why Python?
    - Easy for beginners, but many advanced features.
    - Consistent syntax (more or less).
    - Good for prototyping (fast to write).
    - Good for debugging (e.g. stack traces).
    - Powerful standard library:
        - unit tests (module "unitest")
        - profiler (module "profile")
        - regular expressions (module "re")
        - iterators (module "itertools")
        - tests of docstring-embedded examples (module "doctest")
        - and many more!
    - Widely adopted, including by bioinformatics community.
    - Thriving development community for language itself and associated tools.

But remember: there is no "one true/best language"; you will have to
learn others.

Several Python implementations: CPython, Jython, IronPython, etc.


(1) Resources for Python programming

- docs.python.org
    - "index" is most useful
    - "Library Reference" is next in usefulness
    - These are the most relevant pages for BME205 Assignment #1:
        Tutorial docs:
http://docs.python.org/tutorial/interpreter.html
            http://docs.python.org/tutorial/introduction.html
            http://docs.python.org/tutorial/controlflow.html
            http://docs.python.org/tutorial/inputoutput.html
        Understanding strings (and more generally, sequences):
            http://docs.python.org/library/stdtypes.html#string-methods
            http://docs.python.org/library/stdtypes.html#string-formatting-operations
            http://docs.python.org/library/stdtypes.html#sequence-types-str-unicode-list-tuple-buffer-xrange
    - Built-in functions: http://docs.python.org/library/functions.html

- Python in a Nutshell book (useful for advanced users only)

- Editors/IDEs
    - WingIDE has a free version: http://wingware.com/downloads/wingide-101/4.0.4-1/binaries
    - Eclipse with PyDev: http://pydev.org
    - I recommend emacs (Aquamacs on Mac OS X)
        - Use python-mode.el mode (NOT python.el): https://launchpad.net/python-mode
        - Auto-completion package available (looks in current buffer to append
          to current dictionary on-the-fly):
              http://cx4a.org/software/auto-complete/
                  or
              http://www.emacswiki.org/emacs/AutoComplete
        - recommended by John St. John: YaSnippet (http://code.google.com/p/yasnippet/);
          can be integrated with AutoComplete

- epydoc for documentation (it's like javadoc).

- Mac OS X only: when installing Python, I recommend using MacPorts instead of
  downloading the DMG from python.org .


(2) Starting up Python

- Python versions: use 2.7.x, do NOT use 3.x yet!
- Python installation on SOE machines - see Evan the TA for this.
- Command-line interpreter
    - useful for playing around with one line of code at a time
    - using it as a calculator
    - help() function, takes strings or references:
        help ('print')
        help (print)  # same, as long as name is bound to a reference
        help ('modules')
        x = 205
        help (x)
        s = 'some arbitrary string'
        help (s.rstrip)  # help on object methods
  - Whitespace matters!  But only in indentation.
  - History (readline support), reverse search.
  - Ctrl+D to exit.


(3) Python types and data model

- Everything is a reference; the object to which we refer has the type.
- Binding: associating a (possibly named) reference to an object.
  Sometimes called "variable assignment", but in Python that's a misnomer.
- Most important types:
    int, float, bool, string, list, tuple, dict, None
- Categories of types
    mutable vs immutable
    sequence vs non-sequence
- Type conversion.
- Useful reference
    - standard type hierarchy: http://docs.python.org/reference/datamodel.html#the-standard-type-hierarchy
    - how to use standard types: http://docs.python.org/library/stdtypes.html


(4) Writing your first Python program (demo0.py)

- Shebang/hash-bang.
- chmod +x to make program executable.
- Two ways to run a Python program in a shell
    - Use shebang to locate interpreter (must have done chmod +x):
        ./progName.py
    - Ignore shebang, specify explicitly which Python interpreter to
      use (also ignores chmod +x setting):
        python progName.py
        /usr/local/bin/python2.7 progName.py


(5) Other Python demo programs (demo*.py).
    See program code to learn.


(6) Modules that I find most useful in bioinformatics work

- Standard Python library modules
    argparse - parsing command-line arguments
    array - memory-efficient arrays
    itertools - iteration and combinatorics
    random - generate random numbers, shuffle lists
    re - regular expressions
    sys - system-level stuff
  If it is not obvious to you why they are useful, ask me.
  The awesomeness of some modules (e.g. "itertools") may not be obvious.

- 3rd party modules
    - pysam for SAM/BAM files (sequencing reads): http://code.google.com/p/pysam/
    - ruffus for pipelines (replaces makefiles): http://ruffus.org.uk/
    - rpy2 for Python/R integration: http://rpy.sourceforge.net/rpy2.html


(7) Etc.

- Functional programming with Python: http://docs.python.org/howto/functional.html
  Has fun tidbits such as "how to multiply together all items in a list".

Attachments