Lecture 14 - Functional Programming, Python Sorting

Question 1

Define first-class object

Answer 1

An entity within a programming language that can:

• Appear in an expression
• Be assigned to a variable
• Be used as an argument
• Be returned by a function call

Question 2

What do we know about functions in python? (1)

Answer 2

They’re first-class objects

Question 3

What are the features of functional programming? (4)

Answer 3

1. Functions are first-class objects
2. objects are typically immutable
3. a function’s return value depends only on its input parameters
4. Avoiding iteration – use recursion – map/reduce over lists
5. declarative style (what to do), rather than imperative (how to do it)

Question 4

What are the advantages of functional programming? (4)

Answer 4

Modularity

• FP forces use of small, well-defined chunks of code

Composability

• – FP relies on data flowing between functions, so hooking together functions is easy

Ease of testing and debugging

• – when functions don’t depend on global state, every function can be unittested

Formal provability

• lack of assignments and side effects makes it much easier to prove invariants

Question 5

What’s the syntax for generator comprehension?

Answer 5

Same as list comprehension but with parentheses

Question 6

Describe lazy vs eager evaluation. Which is used by most languages?

Answer 6

• Lazy
  
  • Only evaluate an expression when its value is actually needed
  • Allows potentially infinite data structures
• – contrast with “eager” or strict evaluation, which always evaluates an expression even if the value is never actually use

2. Most use eager evaluation

Question 7

What are you doing when you make a generator?

Answer 7

Make the function able to be used as an iterator

Question 8

What happens when yield is executed? (4)

Answer 8

• any function with “yield” is automatically modified to return a generator object which executes the code
• each time yield is executed, its value is returned as the next value of the iterator;
• the function will then resume execution at the next statement which would have executed without the yield
• when the function returns, exception StopIteration is raised automatically

Question 9

What is an infinite generator? Give an example of this.

Answer 9

• An infinite generator function is one that simply never reaches a “return” statement
• A simple one is itertools.count()

Question 10

What does map do? (2)

Answer 10

• Applies a function (1st arg) to the iterable (2nd arg)
• Returns a map object which is an iterator over the results

Question 11

What are the args of filter? What does it return?

Answer 11

• args
  
  • 1st: function that accepts the values from the iterator of the 2nd arg and returns a boolean
  • 2nd: iterable
• Returns the elements from the iterator for which the 1st arg function returned true

Question 12

• What does the reduce function do?
• What are the args

Answer 12

• ● reduce is repeated application of a binary function to accumulate a result
• – reduce(f,[a,b,c,d]) == f(f(f(a,b),c),d)
• – reduce(f,[a,b,c,d],init) == f(f(f(f(init,a),b),c),d)

Question 13

What are the different callables? (6)

Answer 13

• named functions created with def keyword
  ● anonymous functions created with lambda
  ● instances of classes with __call__() method
  ● closures
  – “operations with data attached” (conceptually the inverse of classes)
  ● static methods of instances
  ● generator functions

Question 14

What is currying?

Answer 14

the process of converting a multi-argument function into a chain of single-argument function applications is called “currying”,

Question 15

Describe the idea of a python closure

Answer 15

• A closure is an inner function that remembers and has access to variables in the local scope in which it was created even after the outer function has finished executing”
• In other words: When you have a function def inside another function def, and the outer function def returns the inner function def, if the inner function def uses a value from the outer function’s scope, that value is bound at the time the closure was created (aka when the outer function was called)

Question 16

Describe chain() from itertools (1)

Answer 16

chain together multiple iterators

Question 17

What does tee from itertools do?

Answer 17

split an iterator into two lazily-evaluated copies

Question 18

What does cycle from itertools do?

Answer 18

– cycle(): endlessly cycle through the elements of an iterable

Question 19

What does dropwhile from itertools do?

Answer 19

skip elements of an iterable until the first time a predicate
becomes false

Question 20

What does groupby from itertools do?

Answer 20

create sub-iterators grouped by the value of a keygen function

Question 21

What does partial from functools do?

Answer 21

return an object that behaves like a partially-applied function
● similar to currying

Question 22

What does wraps from functools do?

Answer 22

used to create function decorators like @staticmethod

Question 23

How do you sort a list in python?

Answer 23

Has a .sort() method which sorts it in place.

Question 24

How do you sort an iterable in python (two different argument sets)

Answer 24

result = sorted(iterable)
result = sorted(iterable,key,reverse)
● ‘key’ is a function which returns a sorting key
● ‘reverse’ is a boolean; if True, sort in descending order

Question 25

For closures, when does the variable binding happen?

Answer 25

When the closure is created

Question 26

For closures, exactly which variables are remembered?

Answer 26

only local variables in the enclosing scope of the innner def are bound

Global variables are not remembered.

Question 27

What happens if you change the values of the bound variables in “outer” after “inner” finishes but before “outer” returns?

Answer 27

Doesn’t effect the variables in the closure

Question 28

What are some uses of closures? (3)

Answer 28

• “partial application”,.
  
  • currying (a specific instance of partial application)
• “monkey patching”,

Question 29

What is partial application?

Answer 29

fixing one or more function arguments so that the result can be called by someone expecting to invoke a function with fewer arguments

Question 30

What is currying?

Answer 30

a specific instance of partial application, applied repeatedly to reduce the number of arguments one-by-one until the final result takes only one argument

Question 31

What is monkey patching?

Answer 31

we bind a function pointer so that we can modify the pointer while still being able to call the original function