Iterables and Generators

Iterators

Iterators are objects that produce successive items or values from an associated iterable. They:

Hold the state (position) of the iteration
Allow looping just once and must be reinitialized to loop again
Implement the __next__ method that…
- returns the next item in the sequence
- raises the StopIteration exception if there is nothing to return
- can also be invoked using the next(iterable) function

An iterable is an object that can be iterated over.

Must be capable of returning an iterator
Must implement the __iter__ method, callable using the iter function

Simple iterables and iterators examples

Lets define an iterable and an iterator…

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
class Iterable:
    def __iter__(self):
        """
        Called by iter(Iterable())
        """ 
        return Iterator()

class Iterator:
    def __init__(self):
        self.x = -1
        
    def __next__(self):
        """
        Called by next(iterator)
        """ 
        self.x += 1
        return self.x

… and instantiate them.

1
2
3
iterable = Iterable()

print(iterable)

<__main__.Iterable object at 0x7f474c781160>

1
2
3
4
iterator = iter(iterable)
# iter(iterable) ==> iterable.__iter__()

print(iterator)

<__main__.Iterator object at 0x7f474c7814e0>

Let’s call the next() function on the iterator.

1
2
# iterator.__next__() 
print(next(iterator))

An object can also define both next and iter methods.

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
class SimpleIterable:
    def __iter__(self):
        self.x = -1
        return self

#     def __next__(self):
#         self.x += 1
#         return self.x
    
    def __next__(self):
        if self.x <= 3:
            self.x += 1
        else:
            raise StopIteration
        return self.x

1
2
3
iterable = SimpleIterable()
print(type(iterable))
# iterable.x

<class '__main__.SimpleIterable'>

1
2
3
iterator = iter(iterable)
type(iterator)
# iterable.x

__main__.SimpleIterable

1
2
# Call next 5 times
next(iterator)

How to iterate on iterators

Iterators from containers (e.g. lists)

1
2
a = list(range(4))
print(a)

[0, 1, 2, 3]

Calling next() on a list won’t work.

1
next(a)

---------------------------------------------------------------------------

TypeError                                 Traceback (most recent call last)

<ipython-input-10-15841f3f11d4> in <module>
----> 1 next(a)


TypeError: 'list' object is not an iterator

But getting an iterable from a list and iterating on it will.

1
2
3
4
a = iter(a)
# iter(iter) = iter

next(a)

The foreach construct

Built-in in the language with the for … in construct
It allows looping on all elements of an iterable.
Automatically calls the iter(…) function before starting looping

1
2
3
4
print("With iter()")
iterator = iter(iterable)
for item in iterator:
    print(item)

With iter()
0
1
2
3
4

1
2
3
print("Without iter()")
for item in SimpleIterable():
    print(item)

Without iter()
0
1
2
3
4

Generators

Functions containing the keyword yield
yield :
- works similarly to return and returns an object when called…
- … but state of the function is saved
When next() is called again on the generator function, execution resumes where it was left off
Note that generators do not return values when initialized.

Examples

Trivial generator

1
2
3
4
5
6
7
8
def f():
  print("-- start --")
  yield 3

  print("-- middle --")
  yield 4

  print("-- finished --")

1
2
generator = f()
generator

<generator object f at 0x7f47500215e8>

1
next(generator)

-- start --
3

Counter

1
2
3
4
5
6
7
8
9
def counter():
    x : int = 0
        
    while True:
        yield x
        x += 1
        
generator = counter()
generator

<generator object counter at 0x7f4750021930>

1
next(generator)

Generators can also be defined inline!

1
2
3
4
generator = (x for x in range(10))
print("generator type:", type(generator))

list(generator)

generator type: <class 'generator'>

[0, 1, 2, 3, 4, 5, 6, 7, 8, 9]

Mind the difference with list comprehensions!

1
2
not_a_generator = [x for x in range(10)]
print("not_a_generator type:", type(not_a_generator))

not_a_generator type: <class 'list'>

Generators are lazy iterators

They are used to generate values dynamically
- Very useful to cope, for instance, with Out of Memory issues
As iterators, they don’t implement the __len__ method
- i.e., len() function will cause an exception
Generators support bidirectional communication.
- You can pass values to the generator after its initialization
Concurrent and recursive invocations are allowed…
- … even though they are not thread safe out of the box.

Dynamic value generation example

1
2
3
4
5
6
7
8
9
from datetime import datetime
print("MS output format:", datetime.now().microsecond)

def very_unsafe_prng(max_value):    
    while True:
        yield datetime.now().microsecond % max_value
        
generator = very_unsafe_prng(10)
generator

MS output format: 445995

<generator object very_unsafe_prng at 0x7f4750021750>

1
2
# No len!
len(generator)

---------------------------------------------------------------------------

TypeError                                 Traceback (most recent call last)

<ipython-input-22-95dd6a62a607> in <module>
      1 # No len!
----> 2 len(generator)


TypeError: object of type 'generator' has no len()

1
next(generator)

Bidirectional communication

Bidirectional communication allows to send values to the generator. Relies on three methods:

.send(…): sends the value to the generator and, like next(), returns the next value
.throw(…): throws the passed exception after resuming the generator that will handle it
.close(): stops the generator. Equivalent to .throw(GeneratorExit())
yield can be used in expressions to assign values to generator’s variables
- Values will be assigned when the generator resumes from yield

Example

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
from random import choice

# Define allowed values ([1, 7])
values = list(range(1, 8))

# Define a generator
def seven_and_half():
    values_sum = 0
    results = []

    player = 1
    
    # Keep going until generator is closed
    while True:
        # Pick a card (pseudo) randomly
        value = choice(values)

        # Accumulate the values
        values_sum += value
        
        # 
        try:
            response = yield value, values_sum
        except GeneratorExit:
            results.append((player, values_sum),)
            
            for player, score in results:
                print("Player {} scored {}".format(player, score))
            break
        
        if response is False or response is None:            
            results.append((player, values_sum),)
            values_sum = 0
            player += 1
            
    print("Exiting")

1
2
# Init the generator
generator = seven_and_half()

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
keep_playing = None

# Keep in mind that
# generator.send(None) === next(generator)
    
while True:  
    if keep_playing is False:
        break
        
    print("Picking a card.")
    value, values_sum = generator.send(keep_playing)
        
    print("Picked {}. Total: {}".format(value, values_sum))
    
    if values_sum > 7:
        print("You lost!")
        keep_playing = False    
    else:
        keep_playing : bool = (input("Keep picking? ") in ["y", "Y", True] )
    # print(output)

Picking a card.
Picked 5. Total: 5
Keep picking? n

1
generator.close()

Player 1 scored 5
Exiting

Exercise

Intro

CSV (Comma-Separated Values) files are text files where each row is a data record and columns are separated by commas (or some other character).

The first row is (usually) the header (i.e., the name of the corresponding column).

A CSV file looks like:

id,name,surname
0,Mickey,Mouse

Request

You have to process a CSV file. Lets assume it is too large to fit in RAM.

You should process it in small pieces, e.g., by reading each line sequentially using a generator.

Specifically, after reading the header, for each line of the file create a dictionary with the column-value associations.

A few tips

Pathlib module offers classes representing filesystem paths with semantics appropriate for different operating systems
Use the zip(*iterables) builtin function. From the docs:
- Builds an iterator that aggregates elements from each of the iterables
- That is, it returns an iterator of tuples, i.e., The i-th element of the tuple contains the i-th element from each of the argument sequences or iterables.
- The iterator stops when the shortest input iterable is exhausted
Use the with statement. From the docs: the with statement is used to wrap the execution of a block with methods defined by a context manager. This allows common try…except…finally usage patterns to be encapsulated for convenient reuse.
As an example CSV, download as raw the Rolling Stone Magazine’s list of “The 500 Greatest Albums of All Time.” from GitHub.

Solution

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
from pathlib2 import Path


def dataset_reader(file):
    # Lets use pathlib instead of using the open() function,
    # with open(file, "r+") as f:

    # Creating a Path instance.
    file = Path(file)

    # Not actually needed, just showing some functionality
    if not file.absolute():
        file = file.resolve()
        
    print("Opening", file.name, "in folder", file.parent)
    
    if not file.exists():
        raise FileNotFoundError("File doesn't exist!")
        
    
    with file.open("r+", encoding="ISO-8859-15") as f:
        header = f.readline()
        columns = header.strip().split(',')
        
        print("Found columns:", columns)
        for line in f:
            values = line.strip().split(',')
#             print(values)
            
            try: 
                yield dict(zip(columns, values))
            except GeneratorExit:
                print("Closing the generator!")
                break

1
2
3
file = "albumlist.csv"

generator = dataset_reader(file)

1
next(generator)

Opening albumlist.csv in folder .
Found columns: ['Number', 'Year', 'Album', 'Artist', 'Genre', 'Subgenre']

{'Number': '1',
 'Year': '1967',
 'Album': "Sgt. Pepper's Lonely Hearts Club Band",
 'Artist': 'The Beatles',
 'Genre': 'Rock',
 'Subgenre': '"Rock & Roll'}

References

Last updated on 25 Jul 2024