Skip to main content

Data structures and utilities for monadic style functional programming.

Project description

PyMonad implements data structures typically available in pure functional or functional first programming languages like Haskell and F#. Included are Monad and Monoid data types with several common monads included - such as Maybe and State - as well as some useful tools such as the @curry decorator for defining curried functions. PyMonad 2.0.x represents and almost complete re-write of the library with a simpler, more consistent interface as well as type annotations to help ensure correct usage.

1 Getting Started

These instructions will get you a copy of the project up and running on your local machine for development and testing purposes.

1.1 Prerequisites

PyMonad requires Python 3.7+. If installing via pip then you will also need Pip and Wheel installed. See those projects for more information on installing them if necessary.

Potential contributors should additionally install pylint and pytype to ensure their code adheres to common style conventions.

1.2 Installing

1.2.1 From the Python Package Index (PyPI) with pip

From a command line run:

pip install PyMonad

1.2.2 Manual Build from PyPI

Download the project files from https://pypi.org/project/PyMonad/#files and from the project directory run:

python setup.py install

If that doesn’t work you may need to run the following instead.

python3 setup.py install

1.2.3 From github

Clone the project repository:

git clone https://github.com/jasondelaat/pymonad.git

Then from the project directory run setup.py as for the manual build instructions above.

1.2.4 Example Usage

The following example imports the tools module and uses the curry function to define a curried addition function.

import pymonad.tools

@pymonad.tools.curry(2) # Pass the expected number of arguments to the curry function.
def add(x, y):
    return x + y

# We can call add with all of it's arguments...
print(add(2, 3)) # Prints '5'

# ...or only some of them.
add2 = add(2)  # Creates a new function expecting a single arguments
print(add2(3)) # Also prints '5'

1.2.5 Next Steps

The PyMonad documentation is a work in progress. For tutorials, how-to, and more head over to the PyMonad Documentation Project. If you’d like to contribute visit the documentation repository here.

1.3 Upgrading from PyMonad 1.3

If you’ve used the 1.x versions of PyMonad you’ll notice that there are a few differences:

1.3.1 Curried functions

Currying functions in PyMonad version 1.x wrapped a function in an instance of the Reader monad. This is no longer the case and currying simply produces a new function as one might expect.

The signature of curry has changed slightly. The new curry takes two arguments: the number of arguments which need to be curried and the function.

from pymonad.tools import curry

def add(x, y):
    return x + y

curried_add = curry(2, add)
# add = curry(2, add) # If you don't need access to the uncurried version.

curry is itself a curried function so it can be used more concisely as a decorator.

from pymonad.tools import curry

@curry(2)
def add(x, y):
    return x + y

1.3.2 Operators

Version 2 of PyMonad discourages the use of operators (>>, \*, and &) used in version 1 so old code which uses them will break. Operators have been removed from the default monad implementation but are still available for users that still wish to use them in the operators package. To use operators:

# Instead of this:
# import pymonad.maybe

# Do this:
import pymonad.operators.maybe

While it’s unlikely operators will be removed entirely, it is strongly suggested that users write code that doesn’t require them.

1.3.3 Renamed Methods

The fmap method has been renamed to simply map and unit is now called insert.

from pymonad.maybe import Maybe

def add2(x):
    return x + 2

m = (Maybe.insert(1)
     .map(add2)
)

print(m) # Just 3

1.3.4 Applicative Syntax

Previously applicative syntax used the & operator or the amap method. amap still exists but there’s now another way to use applicatives: apply().to_arguments()

from pymonad.tools import curry
from pymonad.maybe import Maybe, Just

@curry(2)
def add(x, y):
    return x + y

a = Just(1)
b = Just(2)

c  = Maybe.apply(add).to_arguments(a, b)
print(c) # Just 3

If the function passed to apply accepts multiple arguments then it must be a curried function.

1.3.5 New then method

The then method combines the functionality of both map and bind. It first tries to bind the function passed to it and, if that doesn’t work, tries map instead. It will be slightly less efficient than using map and bind directly but frees users from having to worry about specifically which functions are being used where.

from pymonad.tools import curry
from pymonad.maybe import Maybe, Just, Nothing

@curry(2)
def add(x, y):
    return x + y

@curry(2)
def div(y, x):
    if y == 0:
    return Nothing
    else:
    return Just(x / y)

m = (Maybe.insert(2)
     .then(add(2)) # Uses map
     .then(div(4)) # Uses bind
)

print(m) # Just 1.0

1.3.6 Getting values out of Maybe and Either

Previously, if you need to get a value out of a Maybe or an Either after a series of calculations you would have to access the .value property directly. By the very nature of these two monads, .value may not contain valid data and checking whether the data is valid or not is the problem these monads are supposed to solve. As of PyMonad 2.3.0 there are methods – maybe and either – for properly extracting values from these monads.

Given a Maybe value m, the maybe method takes a default value, which will be returned if m is Nothing, and a function which will be applied to the value inside of a Just.

from pymonad.maybe import Just, Nothing

a = Just(2)
b = Nothing

print(a.maybe(0, lambda x: x)) # 2
print(b.maybe(0, lambda x: x)) # 0

The either method works essentially the same way but takes two functions as arguments. The first is applied if the value is a Left value and the second if it’s a Right.

from pymonad.either import Left, Right

a = Right(2)
b = Left('Invalid')

print(a.either(lambda x: f'Sorry, {x}', lambda x: x)) # 2
print(b.either(lambda x: f'Sorry, {x}', lambda x: x)) # Sorry, Invalid

1.4 Note on efficiency in versions <2.3.5

In pymonad versions 2.3.4 and earlier, an error in the implementation of then meant that some monad types executed then with exponential complexity. As of version 2.3.5 this has been corrected. All monad types now execute then in linear time. A similar problem occured with the map and bind methods for the State monad which have also been fixed in 2.3.5

If you’re using an earlier version of pymonad upgrading to 2.3.5 is highly recommended.

2 Running the tests

2.1 Unit Tests

These tests primarily ensure that the defined monads and monoids obey the required mathematical laws.

On most *nix systems you should be able to run the automated tests by typing the following at the command line.

./run_tests.sh

However, run_tests.sh is just a convenience. If the above doesn’t work the following should:

python3 -m unittest discover test/

2.2 Style Tests

Contributors only need to run pylint and pytype over their code and ensure that there are no glaring style or type errors. PyMonad (mostly) attempts to adhere to the Google Python Style Guide and includes type hinting according to PEP 484.

In general, don’t disable pylint or pytype errors for the whole project, instead disable them via comments in the code. See the existing code for examples of errors which can be disabled.

3 Authors

Jason DeLaat - Primary Author/Maintainer - https://github.com/jasondelaat/pymonad

4 License

This project is licensed under the 3-Clause BSD License. See LICENSE.rst for details.

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

PyMonad-2.4.0.tar.gz (22.9 kB view details)

Uploaded Source

Built Distribution

PyMonad-2.4.0-py3-none-any.whl (29.1 kB view details)

Uploaded Python 3

File details

Details for the file PyMonad-2.4.0.tar.gz.

File metadata

  • Download URL: PyMonad-2.4.0.tar.gz
  • Upload date:
  • Size: 22.9 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/3.4.1 importlib_metadata/4.0.1 pkginfo/1.7.0 requests/2.25.1 requests-toolbelt/0.9.1 tqdm/4.60.0 CPython/3.7.9

File hashes

Hashes for PyMonad-2.4.0.tar.gz
Algorithm Hash digest
SHA256 6b9d98e40bf2e0f1cc1900309365f9ba4997f3432a1d6e71f80f41ace2890204
MD5 7da98cd9d410bc10c52f113ea090d01a
BLAKE2b-256 3b4e68803065743007b7dc000c543f45f6e407f8a1cbc673fdab016df3ca4b44

See more details on using hashes here.

File details

Details for the file PyMonad-2.4.0-py3-none-any.whl.

File metadata

  • Download URL: PyMonad-2.4.0-py3-none-any.whl
  • Upload date:
  • Size: 29.1 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/3.4.1 importlib_metadata/4.0.1 pkginfo/1.7.0 requests/2.25.1 requests-toolbelt/0.9.1 tqdm/4.60.0 CPython/3.7.9

File hashes

Hashes for PyMonad-2.4.0-py3-none-any.whl
Algorithm Hash digest
SHA256 e78d9bf3b0712b165d4f43c29b0c1b23ed3ff11dbae1a6d61402cda7f0827d1a
MD5 61a6aa18cc022456f9d083b00838e4c8
BLAKE2b-256 8d42461af4ed7ea252b46c53b59f965d5f1eda971eef4d0c12f8320286578bf3

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