Programming with Python - Adv.

Programming with Python
Adv. Topics

Mosky

1

Mosky:
● The examples and the PDF version are available at:
– j.mp/mosky-programming-with-python.
● It is welcome to give me any advice of this slide or
ask me the answers of the challenges.
– mosky.tw

2

Topics
● Basic Topics ● Adv. Topics
– Python 2 or 3? – Module and Package
– Environment – Typing
– hello.py – Comprehension
– Common Types – Functional Technique
– Flow Control – Object-oriented Prog.
– File I/O – Useful Libraries
– Documentation ● Final Project
– Scope – A Blog System
3

Module and Package
Write you own module and package!

4

Module and Package
● A Python file is just a Python module:
– import module_a # module_a.py

● A folder which has __init__.py is just a Python package:
– import package_x # __init__.py
– import package_x.module_b # package_x/module_b.py
– from . import module_c
# (in package_x.moudle_b) package_x/module_c.py
– $ python -m package_x.module_b

● Do not name your file as any built-in module.
– ex. sys.py

5

Module and Package (cont.)
● The tree:
– .
├── moudle_a.py
└── package_x
├── __init__.py
├── module_b.py
└── module_c.py

6

The import Statement
● A module is only imported at the first import.
● import module
module.val = 'modified'
– The module is affected by this modification.
● from module import val
val = 'modified'
– The module is not affected by this modification.
– It does a shallow copy.

7

Typing
static? dynamic? weak? strong?

8

Static Typing
● Checking types in compile time.
● Usually, it is required to give a type to a variable.
● Python is not static typing.

long x short y

double z

c1 c2 c3 c4 c5 c6 c7 c8

9

Dynamic Typing
● Checking types in run time.
● A variable just points to an object.
● Python is dynamic typing.

x object A

y object b

object c

NOTE: This is an animation and it is not correct in the PDF version.
10

Duck Typing (cont.)
● A style of dynamic typing.
● Happy coding without the template, the generics … etc.
● If it is necessary to check type:
– if hasattr(x, '__iter__'):
● adapt the type inputed
– assert not hasattr(x, '__iter__'), 'x must be iterable'
● notify the programmer
– if isinstance(x, basestring):
● the worst choice

12

Duck Typing (cont.)
#!/usr/bin/env python ● String and integer both
# -*- coding: utf-8 -*-
# file: ex_dyn.py support += operator.
def dynsum(*seq): ● Write the code with
r = seq[0]
for item in seq[1:]:
elasticity.
r += item

return r

if __name__ == '__main__':

print dynsum(1, 2, 3)
print dynsum('x', 'y', 'z')

13

Duck Typing (cont.)
● BUT, it will confuse you when your project is
going to big.
– Name your variables with hint of type.
● item vs. items
● employee vs. employee_name
● args vs. kargs
– Documentation does matter.

14

Weak Typing
● It converts the type if you do an operation which is
not supported by the original type.
● In JavaScript:
– 1 + '1'
→ '11'
– 1 == '1'
→ true
● Python is not weak typing!

15

Strong Typing
● Only do the operations which are supported by the
original type.
– 1 + '1'
→ TypeError
– 1 == '1'
→ False
● Python is strong typing!

16

Comprehension
Compact your statements.

17

List Comprehension
[i for i in range(10)]
[i ** 2 for i in range(10)]
[f(i) for i in range(10)]
[i for i in range(10) if i % 2 == 0]
[i for i in range(10) if not i % 2 == 0]
[i for i in range(10) if g(i)]

18

List Comprehension (cont.)
List comprehension: is equal to:
[ r = []
(i, j) for i in range(3):
for i in range(3) for j in range(3):
for j in range(3) r.append((i, j))
]

19

List Comprehension (cont.)
List comprehension: is equal to:
[ r = []
[ for i in range(3):
(i, j) t = []
for i in range(3) for j in range(3):
] t.append((i, j))
for j in range(3) r.append(t)
]

20

Generator Comprehension
● Generator comprehension:
– The examples:
● (i for i in range(10))
● f(i for i in range(10))
– It is like xrange.
– Lazy evaluation → Save memory.

21

Other Comprehensions
Python 3 only:
● set comprehension:
– {i for i in range(10)}
● dict comprehension:
– {i:i for i in range(10)}
But we can do so with below statements:
● set comprehension:
– set(i for i in range(10))
● dict comprehension:
– dict((i, i) for i in range(10))

22

Functional Technique
Think in the functional way.

23

The any/all Function
● def all_even(seq):
return all(i % 2 == 0 for i in seq)
● all(type(item) is int for item in inputs)
● any(i < 0 for i in inputs)

24

Challenge 3-3: The Primes (cont.)
– limit: in one line. [2, 3, 5, 7, 11, 13,
● hint: use any or all 17, 19, 23, 29, 31,
37, 41, 43, 47, 53,
59, 61, 67, 71, 73,
79, 83, 89, 97]

25

The zip Function
● matrix = [
[1, 2],
[3, 4],
]
● zip(*matrix)

26

The zip Function (cont.)
● result = [
('I am A.', 'email_A'),
('I am B.', 'email_B'),
]
● emails = zip(*result)[1]

27

First-class Functions

#!/usr/bin/env python ● passing functions as
# -*- coding: utf-8 -*-
# file: ex_do.py
arguments.

from operator import add, mul

def do(action, x, y):
return action(x, y)

if __name__ == '__main__':
print do(add, 10, 20)
print do(mul, 10, 20)

28

The lambda Expression
● lambda [args]: [expression]
● It defines an anonymous function.
● It only allows a single expression.
● f = lambda x: g(x)+h(x)
● do(lambda x, y: (x+y)*(x+y), 10, 20)

29

Use sort with Lambda
● d = dict(a=300, b=200, c=100)
● keys = d.keys()
● keys.sort(key=lambda k: d[k])
● for k in keys:
print k, d[k]

30

Use sort with Lambda (cont.)
● names = ['Andy', 'Bob', 'Cindy']
● scores = [70, 100, 95]
● table = zip(names, scores)
● table.sort(key=lambda pair: pair[1])
● for name, score in table:
print name, score

31

The map Function
● map(lambda x: x**2, range(10))
● map(int, '1 2 3'.split(' '))
● map(ord, 'String')
● map(open, [<paths>])
● map(str.split, open(<path>))

32

The map Function (cont.)
● from operator import mul
● a = (1, 2)
● b = (3, 4)
● sum(map(mul, a, b))

33

The filter Function
● filter(lambda i: i % 2 == 0, range(10))
● filter(str.isdigit, strings)
● filter(lambda s: s.endswith('.py'),
file_names)

34

Comprehension vs. map/filter
● [i ** 2 for i in range(10)]
● map(lambda i: i ** 2, range(10))

● [i ** 2 for i in range(10) if i % 2 == 0]
● map(lambda i: i ** 2, filter(
lambda i: i % 2 == 0,
range(10)
))

35

Comprehension vs. map/filter (cont.)
● [ord(c) for c in 'ABC']
● map(ord, 'ABC')

36

Comprehension vs. map/filter (cont.)
Compare the speeds of them:
1. map/filter (with built-in function)
2. comprehension
3. map/filter

37

The reduce Function
● # from functools import reduce # py3
● from operator import add

● seq = [-1, 0, 1]
● reduce(add, s)

● seq = ['reduce ', 'the ', 'lines.']
● reduce(add, s)

38

The partial Function
● from functools import partial
● from operator import add

● rdsum = partial(reduce, add)
● rdsum([-1, 0, 1])
● rdsum(['reduce ', 'the ', 'lines.'])

39

The partial Function (cont.)
● from functools import partial
● from fractions import gcd as _gcd

● _gcd(6, 14)
● gcd = partial(reduce, _gcd)
● gcd([6, 14, 26])

40

Closure
● from math import log

● def mklog(n):
return lambda x: log(x, n)

● log10 = mklog(10)
● log10(100) # n = 10

41

Closure (cont.)
● setattr(DictLike, attrname,
# it is a colsure
(lambda x:
property(
lambda self: self.__getitem__(x),
lambda self, v: self.__setitem__(x, v),
lambda self: self.__delitem__(x)
)
)(attrname)
)

42

The yield Statement
● def mkgen(n):
for i in range(n):
yield i ** 2

● It is a generator.

● gen = mkgen(10)
● for i in gen:
print i

43

Decorator
● def deco(f):
def f_wrapper(*args, **kargs):
print 'DEBUG: ', args, kargs
return f(*args, **kargs)
return f_wrapper

● @deco # is equal to add = deco(add)
def add(x, y):
return x+y

● add(1, 2)

44

Object-oriented Programming
is also available.

45

The class Statement
class Example(object):
class_attribute = 1
def method(self, …):
pass
example = Example()
print example

46

The Class in Python (cont.)
● Everything in Python is object.
– Class is an object, too.
● All class inherit the object → new-style classes
– Use new-style classes. It provides more features.
– Python 3: auto inherit the object.
● Supports multiple inheritance.
– Searching attributes/methods is like BFS.

47

Bound and Unbound Method
● unbound method
– def m(self, ...)
– C.m(c, ...)
● bound method (instance method)
– c.m(...)

48

Class Method and Static Method
● class method
– @classmethod
def m(cls, …)
– C.m()
– c.m()
● static method
– @staticmethod
def m(...)
– C.m()
– c.m()

49

The Data Model of Python
● Special methods
– __init__
– __str__
– __repr__
– __getitem__ → x[key]
– __setitem__ → x[key] = value
– __delitem__ → del x[key]
…
● ref:
docs.python.org/2/reference/datamodel.html

50

Protocol
● It like interface, but it is only described in doc.
● The examples:
– iterator protocol
● object which supports __iter__ and next
– readable
● object which supports read
– ...

51

The employee class
● see examples/ex_empolyee.py .

52

Do Math with Classes
● see examples/ex_do_math_with_classes/ .

53

Challenge 6: Give a Raise
● Give your employee a cindy = Empolyee(...)
raise. cindy.add_salary(1000)
– without limit print cindy.salary
– limit: prevent
modifying salary by
attribute.
● hint: use property

54

Useful Libraries
import antigravity

55

Useful Libraries
The built-in libraies:
– random – pickle
– datetime – json
– re – pprint
– hashlib/hmac – gzip
– decimal – timeit
– glob – logging
– collections – unitest
– subprocess – doctest
– multiprocessing – pdb
– gc ...

56

Useful Libraries (cont.)
The third-party libraries on PyPI:
– Requests – Use it instead of the poor built-in urllib.
– lxml – Do you need to parse HTML? Use it!
– PyYAML – YAML is a the best of data serialization standards.
– PIL – Python Image Library
– NumPy and SciPy – are for mathematics, science, and engineering.
– SymPy – is for symbolic mathematic
– Bottle, Flask or Django – are the web frameworks.
– Sphinx – helps you to create documentation.
...

57

Challenge 7: Iterable Interger
● Make integer iterable. x = IterableInt(10)
– without limit for b in x:
– limit 1: don't use string print b
– limit 2: use collection
● hint: use property 0
1
0
1

58

Final Project : A Blog System

59

Final Project : A Blog System
● The cookbook:
– Flask – A micro web framework.
– Use pickle to store the posts.
– Optional:
● A database instead of pickle. (ex. MySQL, PostgreSQL, ...)
● A cache layer. (ex. memcached, redis, ...)
● A message queue for async jobs. (ex. RabbitMQ, ...)

60

It is the end.
Contact me? http://mosky.tw/

61

Programming with Python - Adv.

Recommended

More Related Content

What's hot (19)

Viewers also liked (9)

Similar to Programming with Python - Adv. (20)

More from Mosky Liu (8)

Recently uploaded (20)

Programming with Python - Adv.