ftoolz

Collection of higher-order and utility functions built on top of cytoolz.

Module overview

Ftoolz are split into few generic modules.

adt package

Package that provides implementation for various Abstract Data Types (ADTs).

ADT	Description
MutIter(s0)	mutable iterator that can be both consumed and appended to, optionally initialized with init state s0

functoolz package

Package that provides higher-order functions commonly associated with Functor, Applicative and Monad.

Moreover, there are general package-level functions implemented in __init__.py.

Function	Description
attempt(ex, f, args, kwargs)	equivalent to try_except with g = lambda _: None
chain(*fs)	reversed function composition chain(f, g) = g o f
silence(errors)	decorator that silences (selected/all) errors raised by decorated function
try_apply(f, args, kwargs)	equivalent to attempt with ex = Exception
try_except(ex, f, g, args, kwargs)	f(args, kwargs) and on exception(s) ex fallback to g(args, kwargs)

The package content is organized into modules by individual type class:

1. iter.py for class Iterable. Warn some functions might not be pure because input iterable is consumed.
2. opt.py for class Optional
3. seq.py for class Seq (Sequece). Methods typically return tuple instances to preserve immutability.

Module function overview

def / .py	iter	opt	seq
apply	x	x	x
apply2	-	x	-
applyN	-	+	-
flatmap	x	x	x
flatten	x	x	x
fmap	x	x	x
fmap2	x	x	x
fmap3	-	x	-
fmapN	-	+	-
fproduct	x	x	x
generate	x	-	x
lift	x	x	x
product	x	x	x
unit	x	*	x
zip_map	+	-	+

• x - implemented, statically type checked
• + - implemented, possible runtime type errors
• * - not implemented, supported natively
• - - not implemented

traverse package

Each module contains traversable-related functions for traversables Iterable and Seq. Individual modules are named and reserved for single functor that wraps elements of the traversable sequence.

List of currently implemented functions in modules (functors):

def / .py	opt
sequence_iter	x
sequence_seq	x
traverse_iter	x
traverse_iter	x

unsafe package

Package unsafe is dedicated to all impure functions that are prone to or will directly use side-effects.

reflection

This module contains functions for class discovery, inspection and modifications.

Table of contents

Function	Description
abstract(type)	True iff type has abstract methods
implementations(type, package)	discover all non-protected implementations of type under package
import_all(package)	recursively import everything under package
protected(type)	True iff type defines attribute __protected__ = True
subclasses(type, package)	discover all subclasses of type under package

dicttoolz

This module contains functions that work with Map (Mapping) instances.

Table of contents

Function	Description
map_val(dict, key, f)	apply value transformation f on value in dict under key
swap(dict, key1, key2)	swap arbitrary values for key1 and key2 in given mapping
swap_values(dict, key1, key2)	same as swap but preserving concrete value type V

itertoolz

This module contains functions that work with Iterable instances.

Table of contents

Function	Description
associate(key_fn, iterable)	associate elements of iterable to keys selected by key_fn
associate_to(key_fn, value_fn, iterable)	associate values obtained from iterable by value_fn to keys
selected by key_fn
collect(iterable)	materialize iterable into a sequence if it's not one already
empty(iterable)	check if iterable is empty, returns flag and unchanged iterable
enumerate_with_final(iterable)	same as iter_with_final but adds index as third part
filter_not_none(iterable)	filter out None elements from iterable
find(predicate, iterable)	find first element of iterable satisfying predicate
first(sequence)	return first element of a sequence or None
fold_right(op, iterable, z)	fold iterable by applying binary operator op from the right
head_tail(iterable)	split iterable into head element and tail iterable
head_tail_list(iterable)	same as head_tail but materialized tail into list
iter_with_final(iterable)	creates iterable of tuples of original element and final flag
last(sequence)	return last element of a sequence or None
make_str(iterable, key_fn, separator)	create string of tokens from iterable selected by key_fn with separator
order_by(iterable, by, key_fn)	order iterable w.r.t. order given by keys sequence by (given key-getter key_fn) and fill in missing values as None
positions(sequence)	collect positions of non-unique items in original sequence
split_by(predicate, iterable)	split elements of iterable by predicate to positives and negatives
take(n, iterable)	take first n elements of an iterable
take_first(iterable)	take first element of an iterable or fail
try_take_first(iterable)	same as take_first but returns None
try_take_last(iterable)	take last element of an iterable or None
unique_list(iterable)	return distinct elements of an iterable as Seq
unique_sorted(iterable)	return distinct elements of an iterable in natural order as Seq

predicates

This module contains common Predicates, i.e. functions from generic or concrete A to bool.

Table of contents

Predicate	Description
some(optional)	True iff optional is not None
none(optional)	True iff optional is None
even(integer)	True iff integer is even
odd(integer)	True iff integer is odd
vall(*args)	True iff no argument has boolean value False
vany(*args)	True iff boolean value of at least one argument is True

typing

Typing contains helpful type aliases and other type-related definitions.

cytoolz

Cytoolz is a cython implementation of a python library supporting functional style called toolz.

We highly recommend reading the API docs and using it in your project.

Ftoolz does not fork but rather extends cytoolz and provides typed stubs for it's API. Please note that the typed stubs do not cover all the functions from cytoolz.

Also some valid cases might not be covered due to Python's restricted typing capabilities.

Setup development environment

It is highly recommended to use virtual environment to develop and test ftoolz. For making things easy there are two make targets to setup ftoolz:

• make setup-dev which creates new virtual environment in ./venv
• make setup that just installs dependencies for development

Of course one can use his/her own favourite tool to create and manage python venv.

To activate the prepared venv run:

source venv/bin/activate

and for deactivation simply deactivate.

Running checks and tests

Type checking

Type checking is done using mypy (for configuration see mypy.ini) and can be executed by:

make type-check

Code style checking

Ftoolz uses Flake8 for enforcing PEP 8 and other code smells.

make flake8-check

Linting

Linting is configured in .pylintrc and can be run via:

make lint

Tests

Unit and doc tests with coverage can be run by

make tests

One can also run all checks and tests at once via

make release-check

Note: Make sure you run these commands in an activate venv or a container.

Distribution

Project uses setuptools for distribution. Check settings in setup.py.