Skip to main content

Chain together lazily computed modifications to iterators

Project description

iterator-chain

Chain together lazily computed modifications to iterators.

One normally needs to do the following

list(map(lambda element: element / 3,
filter(lambda element: element > 32,
map(lambda element: element * 2, [5, 78, 12, 26]))))

Instead do this

iterator_chain.from_iterable([5, 78, 12, 26]) \
.map(lambda element: element * 2) \
.filter(lambda element: element > 32) \
.map(lambda element: element / 3).list()

It allows the developer to read the code in a more natural fashion: from left to right and from top to bottom. The developer no longer needs to "unwrap" the functions to understand the logic.

Install

Include iterator-chain in your requirements.txt file and/or use pip to install it.

$ pip install iterator-chain

API

Start by importing the package.

import iterator_chain

Start the chain

To start the chain, use the from_iterable or from_iterable_parallel function. They take an iterable.

an_iterable = [5, 78, 12, 26]
chain = iterator_chain.from_iterable(an_iterable)
parallel_chain = iterator_chain.from_iterable_parallel(an_iterable)
Function Arguments Description
from_iterable iterable - An iterable to be used in the iterator chain Starts the iterator chain with the supplied iterable. Chaining and terminating methods can now be called on the result.
from_iterable_parallel iterable - An iterable to be used in the iterator chain
chunksize - Keyword. How big of chunks to split the iterator up across the parallel execution units. If unspecified or None, the chunk size will start at 1 and send that many elements to each execution unit. The chunk size will then increment in powers of two and send that many items to each execution unit. This is repeated until the iterator is exhausted. This value is used as the default chunksize for all the following parallel based methods. A specific parallel based method's chunksize can be overrided by supplying the chunksize keyword to that method.
Starts the iterator chain with the supplied iterable. Chaining and terminating methods can now be called on the result. Certain chaining and terminating methods will occur in parallel. Parallel means separate processes to get around Python's GIL.

Continuing the chain

From there, one can call a plethora of additional methods to modify the iterable passed in originally. The methods are outlined below. The methods fall into one of two categories: chaining or terminating.

  • Chaining methods apply some modification to the elements in the iterator, but keeps the chain alive. This allows additional chaining methods to be subsequently called on the result. Because modifications are lazily computed, none of the modifications from chaining methods are applied until after a terminating method is called.
  • Terminating methods also apply some modification, requests some information, or executes something on the elements in the iterator. They stop the chaining by returning an actual value. This value will depend on all the previous chaining methods being executed first.

Chaining methods

Method Arguments Description
map function - A function that takes a single argument Will run the function across all the elements in the iterator.
filter function - A function that takes a single argument Will run the function on every element. function should return a truthy or falsy value. On true, the element will stay; on false, the element will be removed.
skip number - An integer The number number of elements will be skipped over and effectively removed.
distinct Any duplicates will be removed.
limit max_size - An integer The iterator will stop after max_size elements. Any elements afterward are effectively removed.
flatten Any element that is an iterable itself will have its elements iterated over first before continuing with the remaining elements. Strings (str) do not count as an iterable for this method. Dictionaries flatten to its item tuples.
sort key - Keyword. A function of one argument that is used to extract a comparison key from each element
cmp - Keyword. A Python 2.x "cmp" function that takes two arguments
reverse - Keyword. If set to True, the elements will be sorted in the reverse order
Sorts the iterator based on the elements' values. Use key or cmp to make a custom comparison. If key is specified, cmp cannot be used. This method is expensive because it must serialize all the values into a sequence.
reverse Reverses the iterator. The last item will be first, and the first item will be last. This method is expensive because it must serialize all the values into a list.
Parallel Versions
Method Arguments Description
map function - A function that takes a single argument
chunksize - Keyword. Overrides the chunksize supplied to the original from_iterable_parallel
Will run the function across all the elements in the iterator in parallel.
filter function - A function that takes a single argument
chunksize - Keyword. Overrides the chunksize supplied to the original from_iterable_parallel
Will run the function on every element in parallel. function should return a truthy or falsy value. On true, the element will stay; on false, the element will be removed.

Terminating methods

Method Arguments Description
list Serializes the iterator chain into a list and returns it.
count Returns the number of elements in the iterator.
first default - Keyword. Any value. Returns just the first item in the iterator. If the iterator is empty, the default is returned.
last default - Keyword. Any value. Returns just the last item in the iterator. If the iterator is empty, the default is returned.
max default - Keyword. Any value. Returns the largest valued element in the iterator. If the iterator is empty, the default is returned.
min default - Keyword. Any value. Returns the smallest valued element in the iterator. If the iterator is empty, the default is returned.
sum default - Keyword. Any value. Sums all the elements in the iterator together. If any of the elements are un-summable, the default is returned.
reduce function - A function that takes two arguments
initial - Keyword. Any value.
Applies the function to two elements in the iterator cumulatively. Subsequent calls to function uses the previous return value from function as the first argument and the next element in the iterator as the second argument. The final value is returned. If initial is present, it is placed before the items of the sequence in the calculation, and serves as a default when the sequence is empty.
for_each function - A function that takes one argument and returns nothing Executes function on every element in the iterator. There is no return value. If you are wanting to return a list of values based on the function, use .map(_function_).list().
all_match function - A function that takes one argument and returns a boolean Returns True only if all the elements return True after applying the function to them. Else returns False.
any_match function - A function that takes one argument and returns a boolean Returns True if just one element return True after applying the function to it. If all elements result in False, False is returned.
none_match function - A function that takes one argument and returns a boolean Returns True only if all the elements return False after applying the function to them. Else returns True.
Parallel Versions
Method Arguments Description
for_each function - A function that takes one argument and returns nothing
chunksize - Keyword. Overrides the chunksize supplied to the original from_iterable_parallel
Executes function on every element in the iterator in parallel. There is no return value. If you are wanting to return a list of values based on the function, use .map(function).list().

Examples

import iterator_chain
an_iterable = [5, 78, 12, 26]
iterator_chain.from_iterable(an_iterable) \  #starts the chain
    .map(lambda element: element * 2) \  #multiplies every element by two: [10, 156, 24, 52]
    .filter(lambda element: element > 32) \  #keeps any element greater than 32: [156, 52]
    .map(lambda element: element / 3) \ #divides every element by three: [52.0, 17.333333333333332]
    .list()  #and finally returns a list of the result for later use in your application: [52.0, 17.333333333333332]

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

iterator-chain-1.1.0.tar.gz (8.5 kB view details)

Uploaded Source

Built Distribution

iterator_chain-1.1.0-py3-none-any.whl (10.3 kB view details)

Uploaded Python 3

File details

Details for the file iterator-chain-1.1.0.tar.gz.

File metadata

  • Download URL: iterator-chain-1.1.0.tar.gz
  • Upload date:
  • Size: 8.5 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/1.13.0 pkginfo/1.5.0.1 requests/2.22.0 setuptools/41.0.1 requests-toolbelt/0.9.1 tqdm/4.32.1 CPython/3.7.3

File hashes

Hashes for iterator-chain-1.1.0.tar.gz
Algorithm Hash digest
SHA256 c1ca0fe2ab568c162098560ccd6a43893eec3048a1b61230ad56dfd6ac03e024
MD5 7e7ede33ba2c0f5ae9d01c0909da2054
BLAKE2b-256 906927e9fc79ec697724de680acae9738badce289078fb15b7cc0afe9bc6b055

See more details on using hashes here.

File details

Details for the file iterator_chain-1.1.0-py3-none-any.whl.

File metadata

  • Download URL: iterator_chain-1.1.0-py3-none-any.whl
  • Upload date:
  • Size: 10.3 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/1.13.0 pkginfo/1.5.0.1 requests/2.22.0 setuptools/41.0.1 requests-toolbelt/0.9.1 tqdm/4.32.1 CPython/3.7.3

File hashes

Hashes for iterator_chain-1.1.0-py3-none-any.whl
Algorithm Hash digest
SHA256 3a8c960c3b4e193138e7da24136408e52d3cdd7b2c202392feb8cd6062287ecc
MD5 5fc62a9611234174bc6a49e6e23286c0
BLAKE2b-256 f238f23ef3da1f6f6b4b34ca00dbdbb0545ec5a67e3c6f4612b9c3a27d81d4cb

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page