Information Infrastructure II

1
Information Infrastructure II

• I211
• Rajarshi Guha

2
Topics To Be Covered

• Functions
• Writing your own
• The Python library
• Modules

3
What is a Function?

• A piece of code that does a specific thing
• Might be as simple as adding 2 numbers
• Might be as complex as processing an HTTP request
• If your code is doing many things, especially if
  they are repeated
• Make a function out of the repeated tasks

4
Anatomy of a Function
Arguments
Return value
Input
Function
Output
Input
Input

• Could be any type of Python object
• So if you return a list object, it means that
  you can return multiple items
• Document your functions!!!

5
Writing a Python Function

• Decide what arguments it will need
• Decide what it will return
• Note we dont declare the type of a variable
• This is why documentation is very important

def myFunction(x) return x2 print
myFunction(4) print myFunction(hello,
world)
6
Returning from a Function

• Any time you have a return the function will exit
• You can have more than one return statement

• Too many returns can make it difficult to
  understand whats going on

def sign(x) if x lt 0 return -1 if x gt
0 return 1 return 0
7
Returning from a Function

• If you dont write a return the function will
  return None
• But you can explicitly return None if it makes
  sense

• If we didnt find a valid name, theres nothing
  to return
• Returning None lets us know that nothing was found

def validName(x) if x.find(John) gt 0
return Found it return None
8
Scope of Variables

• Scope indicates when and whether a variable is
  visible

myname Rajarshi def validName(myname)
print The name I got was s (myname) print
myname Rajarshi print validName(John) John
print myname Rajarshi
9
Parameter Passing Rules

• When a parameter is sent to a function, the
  function receives a copy
• But a variable holding a list really holds the
  reference to the list
• So if we change the list elements inside a
  function and the changes will be visible outside
  the function

10
Parameter Passing Examples

• You have to be careful when handling lists in
  functions
• Changes made withinthe function will bevisible
  outside it
• This may not be whatyou want

def myFunction(arg1, arg2) arg1 arg1
extra arg2.append(extra) x hello y
a, b myFunction(x,y) print x, y Hello a,
b, extra
11
Laziness Parameter Passing

• We saw that changing a list inside a function,
  makes the changes visible outside the function
• So we can avoid returning changed list, right?
• WRONG
• Doing this can lead to obscure code
• Its OK in some cases

12
So How Should We handle Lists?

• If the list is not going to be modified in a
  function, we dont need to do anything
• If the list is going to be modified, send in a
  copy of the list
• Slicing makes this easy, since a slice returns
  a new list

def myFunction(arg1, arg2) arg1 arg1
extra arg2.append(extra) x hello y
a, b myFunction(x,y) print
x Hello print y a, b
13
Default Parameter Values

• In many cases, we can specify a default value for
  an argument
• Allows us to avoid specifyingit if we want
  thedefault value

def total(x, start 0, end None) if end
is None end len(x) result 0 for I in
range(start, end) result xI
return result x 1,2,3,4,5,6 total(x) 21 tota
l(x, end 2) 3 total(x, start2, end5) 12
14
Default Parameter Values

• Actual parameters are passed in are processed
  leftto right
• All parameterswith defaultsmust come
  afterparameterswithout them

def total(x, start 0, end None) if end
is None end len(x) result 0 for I in
range(start, end) result xI
return result x 1,2,3,4,5,6 total(x) 21 tota
l(x, 2) 3 total(x, 2, 5) 12
15
A Function is an Object

• Well go into more details of objects later on
• But when we say x 2 or y Hello, x and y are
  variables holding objects
• 2 is an integer object
• Hello is a string object
• Functions are also objects
• We can store functions in lists
• We can pass a function to another function
• Lets us write very elegant code

16
Functions as Objects

• We can assign a function to another variable,
  which is then a function

def circumference(radius) return 2 3.14159
radius def func(aFunction, x) return
aFunction(x) circumference(2) 12.56636 circ
circumference circ(2) 12.56636 func(circ,
2) 12.56636 y circ, func ltfunction
circumference at 0x682b0gt, ltfunction func at
0x62530gt

• We can pass a function to another function
• We can store functions in a list

17
Using Functions as Objects

• Find the sum and mean of a sequence of numbers
• Easier than callingeach functionseparately
• Why is the mean reported as 2?

def sum(x) ret 0 for I in x ret I
return ret def mean(x) ret 0 for I
in x ret I return ret / len(x) def
summary(numbers, functionList) values
for func in functionList result
func(numbers) values.append(result)
return values x 1,2,3,4 summary(x, sum,
mean) 10,2
18
Function Attributes

• All functions have an attribute called __name__
• This is like a variable that is always associated
  with a function
• Python makes it available you dont need to
  create it
• But its not really a variable - it cant be
  changed
• Its value is the name the function was defined
  as
• In general, variables with 2 underscores are
  special and well look at them later on

19
Function Attributes
def sum(x) ret 0 for I in x ret I
return ret gtgt sum.__name__ sum gtgt myfunc
sum gtgt myfunc.__name__ sum

• Assigning a function toanother variable doesnot
  change its initial name
• But dont use the __name__ as a variable in a
  function
• It doesnt do what you thinkitll do!

def sum(x) print __name__ return x2 gtgt
sum.__name__ sum gtgt sum(2) __main__ 4
20
Modules

• Every Python file is a module
• Say we write 2 functions called sum and mean in a
  file called func.py
• func.py is a module called func
• We can access the functions and variables in
  func.py by writing
• func.sum
• func.mean

21
Modules

• Lets us keep things smalland simple
• Good design

22
Many Ways to Import

• We can import modules in a variety of ways
• Some ways make life easier
• Other ways are best avoided
• Try to avoid the last form
• You might overwrite a functionwith the same name

import func x 1,2,3 func.sum(x) OR import
func as f x 1,2,3 f.sum(x) OR from func
import s 1,2,3 sum(x)
23
Import Will Execute Statements

• A function definitionmakes the function
  available for later use
• Individual statements will be executed
• Variables can also be accessed from a module
  just likefunctions

import func func.sum(1,2,3) func.funcVar
Hello from func.py 6 7
24
Module Attributes

• Within a module, we can access the __name__
  attribute
• If the module was imported, this gives the name
  of the module
• If the module was run as a program, its value is
  __main__
• This is a good way to include tests
• If the module is run as a program perform tests
• If it was imported by other code, dont do the
  tests

25
Module Attributes
When imported as a moduleno test is performed
func.py
def sum(x) ret 0 for I in x ret I
return ret def mean(x) ret 0 for I
in x ret I return ret/len(x) if __name__
__main__ x 1,2,3 c1 sum(x)
if c1 6 print sum() works fine
else print sum() failed!
import func func.sum(3,4,5)
When run as a programthe test code is run
Macintosh-3 rguha python func.py sum() works
fine
26
Exploring Python Libraries

• Python comes with a variety of modules
• Send/receive mail
• Handle pictures and audio
• Cryptography, numerical analysis
• File system operations
• HTTP connections
• The library is a collection of modules
• Import the ones you need to do the job
• See the Python Library documentation
• See the quick tour or a slightly more advanced
  description

27
System Library (sys)

• Use it to handle command line args, exiting a
  program, writing to STDOUT/STDERR
• Generally we return 0 whenthe program exits
• sys.path is a variable that lists the places
  where Python will look for modules when asked to
  load them
• The current directory is automatically added
• How many directories are in your sys.path?

import sys for i in sys.argv print
i sys.exit(0)
28
Working with the File System

• File systems are different on Unix and Windows
• Python provides the os module
• Contains many functions that hide the differences
• Always try and use the functions in the os
  module, rather than OS specific functions
• Better chance of your program running on
  different machines

29
Working with the Filesystem

• List contents of a directory
• os.listdir(/tmp)
• Change to a new directory
• os.chdir(/home/rguha/src)
• Get the current directory name
• os.getcwd()
• See the documentation page for all the functions

30
More about Files Directories

• You can do lots of things with filenames and
  directory names
• Create a new path
• Split a path in to components
• Check whether something is a file or a directory
• The rules to do these can be tricky
• Usually differ from OS to OS
• Use functions from the os.path module
• Makes life much easier!

31
More about Files Directories

• Check whether something is a directory or a file
• os.path.isdir(/home/rguha)
• os.path.isfile(/home/rguha/x.txt)
• Strip the path from a filename
• os.path.basename(/home/rguha/x.txt) gives
  x.txt
• Join parts of a path to give a full path
• os.path.join(home, rguha, x.txt) gives
  /home/rguha/x.txt (on Unix)

32
Using Math Functions

• The math module contains a variety of functions
• All math functions willreturn something toyou
• Assign it to a variable if you want to do
  something with the result of the function

import math print math.sqrt(4) 2.0 print
math.cos( math.pi / 2) 6.12323399573677e-17 print
math.log10(100) 2.0
33
Doing Things Randomly

• Sometimes you want to get a random number
• Or select one element from a list randomly
• The random module provides methods to do this

34
Doing Things Randomly

• Its useful to remember that randomnumbers are
  notreally random!
• If you run this program3 times, youll get
  3different outputs
• How do you test such a program?

import random x 1,2,3,4 print
random.choice( x ) print random.randrange(6) pri
nt random.random()
35
Doing Things Randomly

• If you want the same sequence of random
  numbers, setthe seed
• Now, if you run the program 3 times, you get
  the same output
• If you dont, the seedgets set from the current
  time

import random random.seed(1234) x
1,2,3,4 print random.choice( x ) print
random.randrange(6) print random.random()
36
Reading From a CSV File

• Reading lines from a CSV files is quite trivial
• Read the file line by line
• Split on comma
• Do something with it
• But there are differences between CSV files
• Some use , some use
• Some will have extra whitespace

37
Reading From a CSV File

• The csv module allows you to handle CSV files
  easily
• Saves you some time
• Makes your code more general

import csv f open(example.csv, r) r
csv.reader(f, quoting csv.QUOTE_NONE) for row
in r print row
import csv f open(example.csv, r) r
csv.reader(f, quoting csv.QUOTE_NONUMERIC) for
row in r print row
38
What Can We Do/

• Internet related stuff
• Connect to web pages, CGIs
• Send email
• Do various forms of compression
• zip, tar, bzip2
• Manipulate raw image/audio data
• If its not in the standard library, someone has
  probably written it

39
Library Usage Policy

• Good programmers will not reinvent the wheel
• Libraries (modules) save you time and effort
• They dont always do exactly what you need
• But they should not be used to hide lack of
  understanding
• For simple library methods, you should be able to
  re-implement that functionality yourself