Movatterモバイル変換

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 PackageWrite 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

Typingstatic? 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 rif __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 #

Strong Typing● Only do the operations which are supported by the original type. – 1 + '1' → TypeError – 1 == '1' → False● Python is strong typing! 16

ComprehensionCompact 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 ComprehensionsPython 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, muldef 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 Statementclass Example(object): class_attribute = 1 def method(self, …): passexample = 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 LibrariesThe 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

Movatterモバイル変換

Change Language

Programming with Python - Adv.

Embed presentation

Recommended

More Related Content

What's hot

Viewers also liked

Similar to Programming with Python - Adv.

More from Mosky Liu

Recently uploaded

Programming with Python - Adv.