Core Python

Python is an interpreted dynamically typed Language. Python uses indentation to create readable, even beautiful, code. Python comes with so many libraries that you can handle many jobs with no further libraries. Python fits in your head and tries not to surprise you, which means you can write useful code almost immediately.

Python was created in 1990 by Guido van Rossum. While the snake is used as totem for the language and community, the name actually derives from Monty Python and references to Monty Python skits are common in code examples and library names. There are several other popular implementations of Python, including PyPy (JIT compiler), Jython (JVM integration) and IronPython (.NET CLR integration).

Python 2.x vs. Python 3.x

Python comes in two basic flavors these days – Python 2.x (currently 2.7) and Python 3.x (currently 3.3). This is an important difference – some code written for one won't run on the other. However, most code is interchangeable. Here are some of the key differences:

There is a utility called 2to3.py that you can use to convert Python 2.x code to 3.x, while the '-3' command line switch in 2.x enables additional deprecation warnings for cases the automated converter cannot handle. Third party tools like python-modernize and the 'six' support package make it easy to target the large common subset of the two variants for libraries and applications which support both 2.x and 3.x.

Programming as Guido Intended It...

Indentation rules in Python. There are no curly braces, no begin and end keywords, no need for semicolons at the ends of lines - the only thing that organizes code into blocks, functions, or classes is indentation. If something is indented, it forms a block with everything indented at the same level until the end of the file or a line with less indentation.
While there are several options for indentation, the common standard is 4 spaces per level:

def function_block():

    # first block stuff

    # second block within first block:

        # second block stuff

        for x in an_iterator:

            # this is the block for the for loop

            print x

        # back out to this level ends the for loop

    # and this ends the second block...

# more first block stuff

def another_function_block():

Comments and Docstrings

To mark a comment from the current location to the end of the line, use a pound sign, '#'.

# this is a comment on a line by itself

x = 3    # this is a partial line comment after some code

For longer comments and more complete documentation, especially at the beginning of a module or of a function or class, use a triple quoted string. You can use 3 single or 3 double quotes. Triple quoted strings can cover multiple lines and any unassigned string in a Python program is ignored. Such strings are often used for documentation of modules, functions, classes and methods. By convention, the "docstring" is the first statement in its enclosing scope. Following this convention allows automated production of documentation using the pydoc module.

 In general, you use one line comments for commenting code from the point of view of a developer trying to understand the code itself. Docstrings are more properly used to document what the code does, more from the point of view of someone who is going to be using the code

Python is the sort of language that you can just dive into, so let's dive in with this example Python script:

#! /usr/bin/env python

""" An example Python script

    Note that triple quotes allow multiline strings

"""

​

# single line comments are indicated with a "#"

​

import sys         # loads the sys (system) library

​

def main_function(parameter):

    """ This is the docstring for the function """

    print " here is where we do stuff with the parameter"

                print parameter

​

    return a_result     # this could also be multiples

​

if __name__ == "__main__":

    """ this will only be true if the script is called 

        as the main program """

​

    # command line parameters are numbered from 0 

    # sys.argv[0] is the script name

    param = sys.argv[1]    # first param after script name

    # the line below calls the main_function and 

    # puts the result into function_result

    function_result = main_function(param)

Branching

Python has a very straightforward set of if/else statements:

if something_is_true:

    do this

elif something_else_is_true:

    do that

else:

    do the other thing

The expressions that are part of if and elif statements can be comparisons (==, <, >, <=, >=, etc) or they can be any python object. In general, zero and empty sequences are False, and everything else is True. Python does not have a switch statement.

Loops

Python has two loops. The for loop iterates over a sequence, such as a list, a file, or some other series:

for item in ['spam', 'spam', 'spam', 'spam']:

    print item

The code above will print "spam" four times. The while loop executes while a condition is true:

counter = 5

while counter > 0:

     counter -= 1

With each iteration, the counter variable is reduced by one. This code executes until the expression is False, which in this case is when "counter" reaches zero.

A better way to iterate over consecutive numbers is use the “range” function that returns an iterator from 0 to n-1:

for i in range(5):

    print i  # Prints numbers from 0 to 4

The “range” function supports other modes that allow to specify the start value and the step value:

range(4, 10)    # Returns values: 4, 5, 6, 7, 8, 9

range(4, 10, 2) # Returns values: 4, 6, 8

Python is different from languages like C or Java in how it thinks about er- rors. Languages like Java are “look before you leap” (LBYL) languages. That is, there is a tendency to check types and values to make sure that they are legal before they are used. Python, on the other hand, thinks of things more in a “easier to ask for forgiveness than permission” (EAFP) style. In other words, Pythonic style would be more likely to go ahead and try the operation and then handle any problems if they occur:

try:

    item = x[0]

except TypeError:

    #this will print only on a TypeError exception 

    print "x isn't a list!"

else:

    # executes if the code in the "try" does NOT 

    # raise an exception

    print "You didn't raise an exception!"

finally:

     #this will always print

     print "processing complete"

In this case, a list or sequence operation is attempted and if it fails because it's the wrong type, the except clause just deals with it. Otherwise the exception will be raised normally. Then, whether an exception happens or not the finally clause will be executed, usually to clean up after the operation in either case.

Variables and Types

Python is a dynamically typed language, but it is also a strongly typed language. So a variable could end up referring to different types of objects, but the object that it's referring to at any given moment is strongly typed. For example:

x = 1       # x points to an integer object

y = 2       # y also points to an integer object

z = x + y   # z points to an integer object – 3

a = y       # a points to the same int object as y

y = "2"     # y now points to a different object, a string

z = x + y   # throws a type mismatch (TypeError) exception since an integer and a string are different types and can't be added.

z = x + a   # z now points to an int (3), since a is pointing to an int 

Python 3.6 allows you to write underscores in numeric intervals to represent them in a more readable fashion:

i = 10_000_000_000

f = 5_123_456.789_123

Duck Typing – If It Quacks Like a ...

While Python objects themselves are strongly typed there is a large amount of flexibility in how they are used. In many languages there is a pattern of checking to be sure an object is of the correct type before attempting an operation. This approach limits flexibility and code reuse – even slightly different objects (say, a tuple vs. a list) will require different explicit checking.

In Python, things are different. Because the exception handling is strong we can just go ahead and try an operation. If the object we are operating on has the methods or data members we need, the operation succeeds. If not, the operation raises an exception. In other words, in the Python world if something walks like a duck and quacks like a duck, we can treat it like a duck. This is called "duck typing".

Python Data Types

Python Built-in Object Types

Python also has built-in object types that are closely related to the data types mentioned above. Once you are familiar with these two sets of tables, you will know how to code almost anything!

Python Operators

The following table lists Python’s common operators:

Sequence Indexes and Slicing

There are several Python types that are all sequential collections of items that you access by using numeric indexes, like lists, tuples, and strings. Accessing a single item from one of these sequences is straightforward – just use the index, or a negative index to count back from the end of the sequences. E.g., my_list[-1] will return the last item in my_list, my_list[-2] will return the second to last, and so on.

Python allows you to return multiple elements of a collection at once:

a, b, c = [1, 2, 3]

Similar syntax works if we need to assign multiple variables at the same time:

a, b, c = lst[1], lst[4], lst[10]

Python 3.5 adds a new twist to this syntax and allows you to unpack a part of an original list into a new list:

a, *b, c = [1, 2, 3, 4, 5]

# a == 1

# b == [2, 3, 4]

# c == 5

Sequence unpacking also works in the “for” loop to iterate over a list of tuples:

for a, b in [(1, 2), (3, 4), (5, 6)]:

    print a, b

This is especially useful when we need to iterate over keys and values of a dictionary:

d = {'key1': 1, 'key2': 2}

for key, value in d.items():

    print(key, value)

There are two ways to define strings in Python: either by using single quote characters or by using double quotes characters:

s1 = “Hello”

s2 = ‘Hello’

Both ways are semantically identical and the difference is a matter of a personal preference.

Strings can be used like lists. You can access random characters by index, get the length of a string using the “len” function, and use slicing:

s = "hello"

s[1]    # 'e'

len(s) # 5

s[1:3] # 'el'

In contrast to other programming languages, Python does allow you to concatenate strings with non-string types:

“The ultimate answer is: ” + 42 # Will cause a runtime error

Instead we need to explicitly convert an object into a string first using the “str” function:

“The ultimate answer is: “ + str(42)

Python also allows you to define a multiline string. To do this, we need to start and end a string literal with either three single quote characters or three double quotes characters:

m1 = “””

Multi

line

“””

m2 = ‘’’

Multi

line

‘’’

Starting from Python 2.6 the idiomatic way to format a string is to use the “format” function:

‘Good morning, {}. How are you today?’.format(‘John’)

This replaced an outdated format operator: “%”.

Python 3.6 has introduced a new syntax for string formatting. Now you can refer variables in the scope if you prepend a string literal with the “f” character:

name = 'John'

f'Good morning, {name}. How are you today?'

Function Definitions

Functions are defined with the def keyword and parenthesis after the function name:

def a_function():

    """ document function here"""

    print "something"

Parameters

Parameters can be passed in several ways. If we have the following function:

def foo(a, b, c):

    print a, b, c

The most common way to pass arguments is by position:

foo(1, 2, 3)

You can also pass arguments by name:

foo(b=2, c=3, a=1)

Or you can combine both methods. In this case, you need to pass the positional arguments first:

foo(1, c=3, b=2) 

You can also define a function with a variable number of positional parameters:

def foo(*args):

    # args is a tuple

    print args

foo(1, 2, 3, 4) # prints (1, 2, 3, 4)

Or a variable number of keyword parameters:

def foo(**args):

    # args is a dictionary:

    print args

foo(1, 2, 3, 4) # prints (1, 2, 3, 4)

If you have a list or a dictionary and you want to pass to a function that accepts variable number of parameters, you can use the unpacking feature:

def foo(*args, **kw):

    print args, kw

lst = [1, 2, 3]

d = {‘a’: 1, ‘b’:2}

foo(*lst, **kw) # prints (1, 2, 3) {'a': 1, 'b': 2}

In this case, elements of collections are “unpacked” and passed as separate arguments.

Python 3.6 has introduced syntax for variable annotations that allow you to associate a parameter with a type or any value:

def foo(a: int, b: str, c: “param”, d: 42):

    print(a, b, c, d)

These annotations do not change the semantics of a function, and are intended to be interpreted by third-party libraries.

Returning Values

You can return any Python object from a function – ints, floats, lists, dictionaries, anything.

def foo():

    return dict(“color”: “blue”)

Thanks to tuple packing and unpacking you can also return more than one item a time. Items separated by commas are automatically 'packed' into a tuple and can be 'unpacked' on the receiving end:

a, b, c = (1, 2, 3)

You define a class with the class keyword:

class MyClass(object):

    def __init__(self, par):

        # initialize some stuff

        self.foo = "bar"

    def a_method(self):

        # do something

    def another_method(self, parameter):

        # do something with parameter

Note: In Python 3.x, you can create classes without inheriting from "object" because that's the default.

Instead of defining getters and setters methods, Python supports defining properties that are defined by methods to get and set a value:

class Foo:

def __init__(self):

self._bar = 1

@property

def bar(self):

return self._bar

@bar.setter

def bar(self, value):

self._bar = value

​

foo = Foo()

foo.bar =5

print foo.bar # 5

All methods and fields are “public” in Python, but it is common to start members’ names from “_” if they are not supposed to be used as part of the “public” interface.

Instantiating Classes

Classes are instantiated using the class name:

​

​

my_class_object = my_class()

​

​

When a class object is instantiated, the class's __init__(self) method is called on the instance, usually doing any set up that is needed: initializing variables and the like.

If the class __init__( ) method accepts a parameter, it can be passed in:

my_class_object = my_class(param)

Inheritance and Mixins

Python supports multiple inheritance. This does provide you with more ways to shoot yourself in the foot, but a common pattern for multiple inheritance is to use "mixin" classes.

Abstract Base Classes, Metaclasses

Abstract base classes are defined in PEP 3119. You can create abstract base classes via the abc module, which was added in Python 2.6.

A metaclass is a class for creating classes. You can see examples of this in Python built-ins, such as int, str or type. All of these are metaclasses. You can create a class using a specific metaclass via __metaclass__. If that is not specified, then type will be used.

Python comes with a concept known as comprehensions. There are 3 types: list comprehensions, dict comprehensions and set comprehensions.

Following is an example of a list comprehension:

new_list = [x for x in range(5)]

This will create a list from 0-5. It is the equivalent of the following for loop:

new_list = []

for x in range(5):

   new_list.append(x)

A dictionary comprehension is similar. It looks like this:

new_dict = {key: str(key) for key in range(5)}

# new_dict == {0: ‘0’, 1: ‘1’, 2: ‘2’, 3: ‘3’, 4: ‘4’}

A set comprehension will create a Python set, which means you will end up with an unordered collection with no duplicates. The syntax for a set comprehension is as follows:

new_set = {x for x in 'mississippi'}

What Does 'Pythonic' Mean?

'Pythonic' is the term that Pythonistas use when they are talking about code that uses the language well and the way that its creators intended. Being Pythonic is a very good thing. Using Java-esque camel cased variable names is not Pythonic, but using it for class names is. Writing for loops in the style of C/C++ is considered un-Pythonic. On the other hand, using Python data structures intelligently and following the Python style guide makes your code Pythonic.

PEP(Python Enhancement Proposal)-20 is the Zen of Python. Written by long time Python developer Tim Peters, the Zen is acknowledged as the core philosophy of Python. In fact, it is always accessible in any Python environment by using import this:

The Zen of Python, by Tim Peters
Beautiful is better than ugly.
Explicit is better than implicit.
Simple is better than complex.
Complex is better than complicated.
Flat is better than nested.
Sparse is better than dense.
Readability counts.
Special cases aren't special enough to break the rules.
Although practicality beats purity.
Errors should never pass silently.
Unless explicitly silenced.
In the face of ambiguity, refuse the temptation to guess.
There should be one--and preferably only one--obvious way to do it.
Although that way may not be obvious at first unless you're Dutch.
Now is better than never.
Although never is often better than *right* now.
If the implementation is hard to explain, it's a bad idea.
If the implementation is easy to explain, it may be a good idea.
Namespaces are one honking great idea -- let's do more of those!

PEP-8 – the Python Style Guide

Python has its own style guide known as PEP8 that outlines various guidelines that are good to follow. In fact, you must follow them if you plan to contribute to Python Core. PEP 8 specifies such things as indentation amount, maximum line length, docstrings, whitespace, naming conventions, etc.

Python's Default Shell

Python is one of several languages that has an interactive shell which is a read-eval-print-loop (REPL). The shell can be enormously helpful for experimenting with new libraries or unfamiliar features and for accessing documentation.

Importing and Using Modules and Libraries

Using external modules and libraries is as simple as using the import keyword at the top of your code.

Of the three styles of import the second (import lib) has the advantage that it is always clear what library an imported element comes from and the chances for namespace collision and pollution are low. If you are only using one or two components of a library the first style (from lib import x) makes typing the element name a bit easier. The last style (from lib import *) is NOT for production code – namespace collisions are very likely and you can break module reloading. There is one major exception to this rule that you will see in many examples and that concerns the include Tkinter GUI toolkit. Most Tkinter tutorials import it as follow: from Tkinter import *. The reason is that Tkinter has been designed so that it is unlikely to cause namespace collisions.

Creating Modules and Libraries

Creating modules and libraries in Python is straightforward. Let’s say we have the following file:

# utils.py

def foo(a, b):

    return a + b

To allow Python to import it we need to put into a directory with the following structure:

├── my_library

│ ├── __init__.py

│ ├── utils.py

If “my_library” is in the current working directory of an interpreter we can import ”utils” in the following way:

from my_library import utils

utils.foo(4, 2)

Notice that to turn “my_library” into a Python library we need to put an empty “__init__.py” file in it. Otherwise Python won’t import modules from a directory.

Be aware that when a module is imported, its code is executed. Therefore, if you want a module to contain an entry point to your application, it is recommended to add the following “if” statement to it:

def foo(a, b):

    return a + b

if __name__ == “__main__”:

    # Application logic goes here

With this you can import functions from a module and safely execute it with a Python interpreter.

Python comes with a standard library of modules that can do much of what you need to get done. The standard library is quite extensive – it would take weeks to become familiar with everything in it.

Whenever you feel the need to go looking for an additional external library, you should first look carefully in the standard library – more often than not, a perfectly good implementation of what you need is already there.

Getting Other libraries

If you find yourself needing additional functionality, you should go take a look in the Python Package Index (PyPI). There you will find thousands of packages that cover a vast array of topics.

To install the packages, you can use pip or easy_install, both of which you'll need to download from PyPI. For full instructions on bootstrapping with these tools, see http://www.pip-installer.org/en/latest/installing.html Sometimes those utilities won't work and you'll have to use the package's included setup.py to do the installation, which normally goes something like this:

python setup.py install

You will see a lot of information output to your screen when you execute the above. In some cases, the module has C headers and will require a C/C++ compiler installed on your machine to complete installation correctly.

Numpy and Scipy

Numpy and scipy are extensive mathematical libraries written to make operating on large data collections easier. As Python's presence in scientific communities has grown, so has the popularity of numpy and scipy. Currently there are conferences devoted to them and to scientific computing. For graphing, you might want to try matplotlib.

IPython – the shell and more The default Python shell has some annoying limitations – it's inconvenient to access the host operating system, there is no good way to save and recover sessions, and it's not easy to export the commands of a session to an ordinary script file. This is particularly irksome for scientists and researchers who may want to spend extensive time exploring their data using an interactive shell.

To address these issues IPython answers these and other problems.

To get IPython, go to http://ipython.org/ and download the version best suited to your operating system.

Web libraries

One of the main uses for Python these days is for web programming. There are several popular web frameworks as described below, as well as other libraries for dealing with web content.

Django
Arguably the most popular web framework, django has taken the Python world by storm in the past few years. It has its own ORM, which makes it very easy to interact with databases.

Pyramid
A Python framework originally based on Pylons, but is now a rebranding of repoze.bfg. Pyramid supports single file applications, decorator-base config, URL generation, etc.

Flask
Flask is also a micro web framework for Python, but it is based on Werkzeug and Jinja2.

Requests
Requests is an HTTP library that provides a more Pythonic API to HTTP requests. In other words, it makes it easier to download files and work with HTTP requests than the standard library.

Beautifulsoup
A great HTML parser that allows the developer to nagivate, search and modify a parse tree as well as dissecting and extracting data from a web page.

Other Libraries

Python has many other libraries ranging over all kinds of topics.
Here is a sampling:

Twisted - networking

Natural Language Tool Kit (NLTK) - language processing

Pygame - games in Python

SQLAlchemy - database toolkit

Python documentation
Python 3 - http://docs.python.org/3/

Python 2 - http://docs.python.org/2.7/

Tutorials

Official - http://docs.python.org/2/tutorial/
Learn Python - http://www.learnpython.org/
Learn Python the Hard Way -
http://learnpythonthehardway.org/book/
Python Anywhere - https://www.pythonanywhere.com/

Books

Python 3 Object Oriented Programming by Dusty Phillips
Python Cookbook (2nd Edition) (Python 2.x) by Alex Martelli, Anna Ravenscroft, David Ascher
Python Cookbook (3rd Edition) (Python 3.x) by David Beazley, Brian K. Jones
Python Standard Library by Example by by Doug Hellmann
Python in Practice by Mark Summerfield
Dive Into Python by Mark Pilgrim