Intro to Functional Programming

Transcript

1. An intro to Functional
   Programming
   Arusha Coders July 6, 2016
   by Reuben Cummings
   @reubano #arushacoders
   

   View Slide

2. a new way
   of thinking
   @reubano #arushacoders
   

   View Slide

3. pure functions
   

   View Slide

4. pure functions
   function
   input output
   outside
   world
   

   View Slide

5. imperative
   version
   @reubano #arushacoders
   

   View Slide

6. class Rectangle(object):
   def __init__(self, length, width):
   self.length = length
   self.width = width
   @property
   def area(self):
   return self.length * self.width
   def grow(self, amount):
   length = self.length * amount
   self.length = length
   

   View Slide

7. class Rectangle(object):
   def __init__(self, length, width):
   self.length = length
   self.width = width
   @property
   def area(self):
   return self.length * self.width
   def grow(self, amount):
   length = self.length * amount
   self.length = length
   internal state
   

   View Slide

8. class Rectangle(object):
   def __init__(self, length, width):
   self.length = length
   self.width = width
   @property
   def area(self):
   return self.length * self.width
   def grow(self, amount):
   length = self.length * amount
   self.length = length
   internal state
   state mutation
   

   View Slide

9. >>> r = Rectangle(2, 3)
   >>> r.length
   2
   >>> r.area
   6
   >>> r.grow(2)
   >>> r.length
   4
   >>> r.area
   12
   

   View Slide

10. >>> r = Rectangle(2, 3)
    >>> r.length
    2
    >>> r.area
    6
    >>> r.grow(2)
    >>> r.length
    4
    >>> r.area
    12
    same commands,
    different results!?
    

    View Slide

11. >>> r = Rectangle(2, 3)
    >>> r.length
    2
    >>> r.area
    6
    >>> r.grow(2)
    >>> r.length
    4
    >>> r.area
    12
    same commands,
    different results!?
    

    View Slide

12. functional
    version
    @reubano #arushacoders
    

    View Slide

13. def make_rect(length, width):
    return (length, width)
    def grow_rect(rect, amount):
    return (rect[0] * amount, rect[1])
    def get_length (rect):
    return rect[0]
    def get_area (rect):
    return rect[0] * rect[1]
    

    View Slide

14. def make_rect(length, width):
    return (length, width)
    def grow_rect(rect, amount):
    return (rect[0] * amount, rect[1])
    def get_length (rect):
    return rect[0]
    def get_area (rect):
    return rect[0] * rect[1]
    no state!
    

    View Slide

15. >>> r = make_rect(2, 3)
    >>> get_length(r)
    2
    >>> grow_rect(r, 2)
    (4, 3)
    >>> get_area(r)
    6
    >>> get_length(r)
    2
    >>> get_area(r)
    6
    returned a new object
    

    View Slide

16. >>> r = make_rect(2, 3)
    >>> get_length(r)
    2
    >>> grow_rect(r, 2)
    (4, 3)
    >>> get_area(r)
    6
    >>> get_length(r)
    2
    >>> get_area(r)
    6
    attributes didn’t change!
    

    View Slide

17. >>> r = make_rect(2, 3)
    >>> get_length(r)
    2
    >>> grow_rect(r, 2)
    (4, 3)
    >>> get_area(r)
    6
    >>> get_length(r)
    2
    >>> get_area(r)
    6
    attributes didn’t change!
    

    View Slide

18. >>> big_r = grow_rect(r, 2)
    assign output of the grow
    function to a new object
    

    View Slide

19. >>> big_r = grow_rect(r, 2)
    >>> get_area(big_r)
    12
    >>> get_length(big_r)
    4
    attributes of new object
    are as expected
    

    View Slide

20. so what??
    @reubano #arushacoders
    

    View Slide

21. from time import sleep
    class ExpensiveRectangle(Rectangle):
    @property
    def area(self):
    sleep(5)
    return self.length * self.width
    >>> r = ExpensiveRectangle(2, 3)
    >>> r.area
    6
    expensive!!
    >>> r.area
    6
    

    View Slide

22. from functools import lru_cache
    @lru_cache()
    def exp_get_area (rect):
    sleep(5)
    return rect[0] * rect[1]
    >>> r = make_rect(2, 3)
    >>> exp_get_area(r)
    >>> exp_get_area(r)
    6
    6
    expensive!!
    memoization
    

    View Slide

23. lazy evaluation
    

    View Slide

24. eager
    version
    @reubano #arushacoders
    

    View Slide

25. def double(ints):
    doubled = []
    for i in ints:
    doubled.append(i * 2)
    return doubled
    >>> some_ints = range(1, 11)
    >>> double(some_ints)
    [2, 4, 6, 8, 10, 12, 14, 16, 18, 20]
    

    View Slide

26. lazy
    version
    @reubano #arushacoders
    

    View Slide

27. def gen_double(ints):
    for i in ints:
    yield i * 2
    >>> gen_double(some_ints)
    
    >>> gen = gen_double(some_ints)
    >>> next(gen)
    2
    >>> next(gen)
    4
    

    View Slide

28. can we do
    better?
    @reubano #arushacoders
    

    View Slide

29. def gen_double(ints):
    >>> gen_double(some_ints)
    
    >>> gen = gen_double(some_ints)
    >>> next(gen)
    2
    >>> next(gen)
    4
    return (i * 2 for i in ints)
    

    View Slide

30. who
    cares??
    @reubano #arushacoders
    

    View Slide

31. def exp_double(ints):
    doubled = []
    for i in ints:
    sleep(5)
    doubled.append(i * 2)
    return doubled
    def exp_gen_double(ints):
    for i in ints:
    sleep(5)
    yield i * 2
    expensive!!
    expensive!!
    

    View Slide

32. >>> exp_double(some_ints)
    >>> gen = gen_exp_double(some_ints)
    >>> next(gen)
    >>> next(gen)
    [2, 4, 6, 8, 10, 12, 14, 16, 18, 20]
    2
    4
    >>> next(gen)
    6
    

    View Slide

33. but wait,
    there’s
    more!
    @reubano #arushacoders
    

    View Slide

34. >>> from itertools import count
    >>>
    >>> inf_ints = count(1)
    >>> double(inf_ints)
    infinite iterator!!!
    

    View Slide

35. >>> from itertools import count
    >>>
    >>> inf_ints = count(1)
    >>> double(inf_ints)
    ...
    >>> gen = gen_double(inf_ints)
    >>> next(gen)
    42220156
    >>> next(gen)
    42220158
    infinite iterator!!!
    

    View Slide

36. stream processing
    

    View Slide

37. >>> stream = range(1, 10)
    >>> doubled = [2 * x for x in stream]
    >>> doubled
    [2, 4, 6, 8, 10, 12, 14, 16, 18]
    >>> old = [x for x in doubled if x > 10]
    >>> old
    [12, 14, 16, 18]
    >>> divided = [x / 3 for x in old]
    >>> divided
    [4.0, 4.6667, 5.3333, 6.0]
    transform stream
    at each stage
    

    View Slide

38. >>> stream = range(1, 10)
    >>> doubled = [2 * x for x in stream]
    >>> doubled
    [2, 4, 6, 8, 10, 12, 14, 16, 18]
    >>> old = [x for x in doubled if x > 10]
    >>> old
    [12, 14, 16, 18]
    >>> divided = [x / 3 for x in old]
    >>> divided
    [4.0, 4.6667, 5.3333, 6.0]
    user pure functions to
    return new streams
    

    View Slide

39. mini riko
    demo
    @reubano #arushacoders
    

    View Slide

40. extracting web data
    

    View Slide

41. RSS feeds
    

    View Slide

42. RSS feeds
    

    View Slide

43. >>> from riko.modules import (
    ... fetch, truncate, sort,
    ... xpathfetchpage as xfetchpage)
    >>>
    >>> url = 'news.ycombinator.com/rss'
    >>> stream = fetch.pipe(
    ... conf={'url': url})
    >>> stream_l = list(stream)
    >>> len(stream_l)
    30
    >>> stream_l[0]
    {'comments': 'http...', 'title':
    'Think...', 'published': 'Tue, 5 Ju...'}
    convert generator
    into a list
    

    View Slide

44. >>> truncated = truncate.pipe(
    ... stream_l, conf={'count': 5})
    >>> truncated_l = list(truncated)
    >>> len(truncated_l)
    5
    >>> sorted = sort.pipe(
    ... truncated_l, conf={
    ... 'rule': {
    ... 'sort_key': 'title'}})
    >>>
    >>> sorted_l = list(sorted)
    >>> sorted_l[0]
    {'title': 'Building a BitTorre...', ...}
    

    View Slide

45. >>> xpath = '/html/body/center/table...'
    >>> xconf = {
    ... 'url': {'subkey': 'comments'},
    ... 'xpath': xpath}
    >>> comments = [
    ... xfetchpage.pipe(s, conf=xconf)
    ... for s in sorted_l]
    >>> item = next(comments[0])
    >>> item.keys()
    ['p', 'div', 'class', 'content']
    >>> item['content']
    "Great work! Quite annoying actually..."
    

    View Slide

46. pipeline
    version
    @reubano #arushacoders
    

    View Slide

47. from riko.collections import sync
    Pipe = sync.SyncPipe
    sconf = {'rule': {'sort_key': 'title'}}
    comments = (
    Pipe('fetch', conf={'url': url})
    .truncate(conf={'count': '5'})
    .sort(conf=sconf)
    .xpathfetchpage(conf=xconf)
    .output)
    item = next(comments)
    

    View Slide

48. head to head
    functional imperative
    cacheable ✔
    testable ✔ ?
    composable ✔
    simple ✔
    stateful ✔
    @reubano #arushacoders
    

    View Slide

49. Resources
    https:/
    /docs.python.org/3.5/library/itertools.html
    https:/
    /docs.python.org/3.5/library/functools.html
    https:/
    /lodash.com/
    http:/
    /coffeescript.org/
    https:/
    /www.youtube.com/watch?v=7aYS9PcAITQ
    https:/
    /github.com/nerevu/riko
    @reubano #arushacoders
    

    View Slide

50. Resources
    https:/
    /github.com/reubano/meza
    http:/
    /ocw.mit.edu/courses/electrical-engineering-
    and-computer-science/6-001-structure-and-
    interpretation-of-computer-programs-
    spring-2005/
    https:/
    /www.infoq.com/presentations/Simple-
    Made-Easy
    @reubano #arushacoders
    

    View Slide

51. Thanks!
    Reuben Cummings
    @reubano
    @reubano #arushacoders
    

    View Slide