python introduction

Introduction to Python

Master in Free Software 2009/2010

Joaquim Rocha <[email protected]>

23, April 2010

Master in Free Software | 20092010 2

What is it?

* Generalpurpose, highlevel programming language

* Created by Guido van Rossum


What is it?

* Named after Monty Python's Flying Circus

* Managed by the Python Software Foundation

* Licensed under the Python Software Foundation license (Free Software and compatible with GPL)


How is it?

* Multiparadigm

* Features garbagecollection

* Focuses on readability and productivity

* Strong introspection capabilities


How is it?

* Batteries included!

* Guido is often called Python's Benevolent Dictator For Life (BDFL) for he makes the final calls in Python's matters


Brief history

* Created in 1991 at the Stichting Mathematisch Centrum (Amsterdam)

* Designed as a scripting language for the Amoeba system

* January 1994: version 1.0


Brief history

* October 2000: version 2.0

* December 2001: version 2.2 (first release under PSF license)

* October 2008: version 2.6

* December 2008: version 3.0


Python Software Foundation License

* Free Software license (recognized by the Open Source Initiative)

* GPL-compatible

* http://python.org/psf/license/

* http://www.python.org/download/releases/2.6.2/license/

http://python.org/psf/license/

http://www.python.org/download/releases/2.6.2/license/


Companies using Python

* Google (and YouTube)

* NASA

* Philips

* Industrial Light & Magic

* EVE Online...


Projects using Python

* Django* Pitivi* Trac* Plone* Frets on Fire* Miro* OCRFeeder...


Some books on Python

* Programming Python (O'Reilly)

* Python in a Nutshell (O'Reilly)

* Dive Into Python (Apress)

* Expert Python Programming (Packt Publishing)


How to install it

* It's already installed in most mainstream distros

* If not, on Debian:

aptget install python


Python Shell

$ pythonPython 2.6.4 (r264:75706, Dec 7 2009, 18:45:15)[GCC 4.4.1] on linux2Type "help", "copyright", "credits" or "license" for more information.>>> import thisThe Zen of Python, by Tim Peters...


Documentation

* Python is really well documented

http://docs.python.org

* On Debian distros, install python2.6doc to view it with Devhelp

http://docs.python.org/


Some helpful functions

* The help command displays the help on a symbol >>> import os >>> help(os.path)

* The dir command displays attributes or available names >>> dir() # displays names in current scope >>> dir(str) # displays attributes of the class and its bases


Python packages and modules

# This is a module, and would be the execution script myapp_run.py myapp/ # A package __init__.py # A special module myappcode.py # A module lib/ # A subpackage __init__.py util_functions.py # Another module

* The __init__.py is often empty but is what allows the package to be imported as a if it was a module.


Example of types

* In Python everything is an object

* str : strings (there is also a string module, that most functions are now available in str)

'This is a string' “This is also a string!” “””And this is a multiline string...”””


Example of types

* int: integer numbers

i = 12 + 4

* float: floating point numbers

f = 0.45


Example of types

* bool: booleans

b = True e = b and c or d

* list: list objects

list_var = ['foo', 45, False]


Example of types

* dict: dictionary objects

d = {'type': 'person', 'age': 24, 10: 'just to show off an int key'}

...


Checking a type

>>> type(some_object) # gives the object's type >>> isinstance('foo bar', str) # returns True


Castings

int('5') # gives the integer 5

str(5) # gives the string '5'

bool('') # gives False


Whitespace indentation

* Python blocks are defined by indentation instead of curly brackets or begin/end statements.

* Failing to properly indent blocks raises indentation exceptions


Whitespace indentation

* If a new block is expected and none is to be defined, the pass keyword can be used

if x is None: pass


Control flow: if

if x < 0: print 'X is negative!' elif x > 0: print 'X is positive!' else: print 'X is zero...'


Control flow: for

for x in range(10): # this is [0..9] if x % 2 == 0: continue print x ** 2


Control flow: while

fruits = ['bananas', 'apples', 'melon', 'grapes', 'oranges'] i = 0 while i < len(fruits): if len(fruits[i]) == 5: break i += 1


Functional programming

* Being multiparadigm, Python also has functions common in functional programming

* It features lambda, map, filter, reduce, ...

numbers = [1, 2, 3, 4] map(lambda x: x ** 2, numbers) # gives [1, 4, 9, 16]

filter(lambda x: x % 2 == 0, numbers) # gives [2, 4]

reduce(lambda x, y: x + y, numbers) # gives 10


Functional programming

* List comprehension

numbers = [1, 2, 3, 4, 5, 6, 7, 8] even_numbers = [n for n in numbers \ if n % 2 == 0]


Error handling: try / except* Equivalent to try, catch statements

import random

n = random.choice(['three', 1]) try: a = 'three ' + n except TypeError: a = 3 + n else: # else body is executed is try's body didn't raise an exception a = 'The answer is ' + a finally: # this is always executed print a


Error handling: try / except

* Getting exception details

try: a = not_defined_var except Exception as e: # the "as" keyword exists since Python 2.6 # for Python 2.5, use “,” instead print type(e)


Error handling: try / except

* Raising exceptions

password = 'qwerty' if password != 'ytrewq': raise Exception('Wrong password!')


Imports

* Import a module:

import datetime d = datetime.date(2010, 4, 22)

* Import a single class from a module:

from datetime import date d = date(2010, 4, 22)


Imports

* Import more than one class from a module:

from datetime import date, time d = date(2010, 4, 22) t = time(15, 45)

* Import a single module from another module:

from os import path notes_exist = path.exists('notes.txt')


Imports

* Import everything from a module:

from datetime import * t = time(15, 45)

* Rename imported items:

from datetime import date from mymodule import date as mydate # otherwise it would have overridden the previous date


Functions

def cube(number): return number ** 3

# optional arguments def power_from_square(base, exponent = 2): # every argument after an optional argument # must be also optional return base ** exponent


Functions

# the exponent argument is optional power(2) # returns 4 power(2, 3) # returns 8

# named arguments can be used power(base=8, exponent=4)

# and can be even interchanged power(exponent=2, base=4)


Object Oriented Programming

class Car:

# this is a class variable min_doors = 3

# contructor def __init__(self, model): # this is an instance variable self.model = model # this is a private variable self._color = 'white'


Object Oriented Programming

# private method def _is_black(self): if self._color == 'black' or self._color == '#000': return True return False

# public method def set_color(self, color): if color in ['black', 'white', 'red', 'blue']: self._color = color

def get_extense_name(self): return 'A very fine %s %s' % (self.model, self._color)


Inheritance

# this is how we say Car extends the Vehicle class class Car(Vehicle):

def __init__(self, model): # we need to instantiate the superclass Vehicle.__init__(self, tyres = 4) self.model = model …


Executing code

* Here is how is Python's main:

if __name__ == '__main__': a = 1 print a

* Execute it with:

$ python print_one.py


Properties

* Properties abstract actions on an attribute, as get, set, del or doc

# When you set/get person.name, it will do a customized set/get class Person: def __init__(self, name): self.name = name def _set_name(self, name): self._name = str(name) def _get_name(self): return 'My name is ' + self._name name = property(_get_name, _set_name)


Distributing your application

* Distutils allows to easily packages your application, install it, etc.

* Our app structure:

SupApp/ setup.py supaapp data/ supaapp.png src/ supaApp/

supaapp.py lib/ util.py


Distributing your application* The magic script

from distutils.core import setup setup(name = 'SupaApp', version = '0.5beta' description = 'A supadupa app!' author = 'Cookie Monster' author_email = '[email protected]' license = 'GPL v3', packages = ['SupaApp', 'SupaApp.lib'], package_dir = {'': 'src'}, scripts = ['supaapp'], data_files = [('share/icons/hicolor/scalable/apps', ['data/supaapp.png'])] )



* The MANIFEST.in generates the MANIFEST file which describes what's to be included in the package:

include supaapp recursiveinclude data *



* How to generate a source package

$ python setup.py sdist

* How to install an app

$ python setup.py install


Docstrings

* Are nowadays written in ReStructuredText

* Documentation of a module:

"""This is a nice module. It shows you how to document modules """ class Person: """This class represents a person. A person can have a name and ... """ def __init__(self, name): "Instantiates a Person object with the given name" self.name = name


Some useful modules

import sys# Available packages/modules must be under the folders# sys.path hasprint sys.path# shows the arguments given after the python commandprint sys.argv

# get png image namesimport ospng_names = [os.path.splitext(f)[0] for f in \ os.listdir('/home/user/images') if f.endswith('png')]


Some useful modules

* ElementTree is an easy to use XML library

from xml.etree import ElementTree as ETtree = ET.parse("index.html")print tree.findall('body/p')


Decorators

* Decorators are functions that control other functions

import sysimport os

def verbose(function):

def wrapper(*args, **kwargs): print 'Executing function: ', function.__name__ print 'Arguments: ', args print 'Keyword Arguments: %s\n' % kwargs return function(*args, **kwargs) return wrapper


Decorators

@verbose # this is how you apply a decoratordef list_dir(directory): for f in os.listdir(directory): print f

if __name__ == '__main__': # list the directory we specify as the first program's argument list_dir(sys.argv[1])


Decorators

* Testing the decorator

$ python decorator_example.py /home/user/Documents Executing function: list_dir Arguments: ('/home/user/Documents/',) Keyword Arguments: {}

notes.txt python_presentation.odp book_review.odt


PDB: The Python Debugger

* Calling pdb.set_trace function will stop the execution and let you use commands similar to GDB

def weird_function(): import pdb; pdb.set_trace()

# code that is to be analyzed...

python introduction

Documents

python apress

python python2

programming python oreilly

psf license

free software license

nutshell oreilly

scripting language

highlevelprogramming