forbiddenfp

forbiddenfp

Summary

Functional-Programming (FP) in a forbidden way:

You can turn arbitrary function into postfix notation in favor of function chaining.

And the library provides many useful patches for builtin/itertools functions.

Install

Compatible with Python 3.7+

pip install forbiddenfp

Examples

# objects are already patched at import time
import forbiddenfp

"abc".print().len()  # print out "abc", then return 3
"abc".then(lambda s: s * 2).filter(lambda s: s == "b").join()  # "bb"

# A more complex one (examples/word_count.py)
("./lorem_ipsum.txt"
 .with_open(lambda path, f: f.read().also(print(f"Reading {path}")))
 .then(lambda s: s.split(" "))
 .counter()
 .print())

See more ./examples.

_unpack

For every higher-order-function chained, we provide an _unpack version as well. e.g. [(1, 2), (3, 4)].sum(lambda x: x[0] * x[1]) can also be called as [(1, 2), (3, 4)].sum_unpack(lambda x, y: x * y)

Why Functional Programming

Separate out control structs (which are functions provided by library) from business logic (which are lambda functions supplied to fill the control structs).

So we can have a clearer scope, and separate concerns, when we want to change on either side of things.

• if ... and if ... else ... are modeled by Maybe/Either monads.
• while ... and for ... in ... are generator/iterators. Additionally, stop-early behavior is takewhile of the sequence.

Philosophically, think more in def transform(old_state) -> new_state, rather than state = modify(state).

What About Auto-completion?

This library is almost an embedded Domain Specific Language (DSL) in Python.

It is understandable that users aren't familiar with what is available without auto-completion features in IDEs.

So we leverage the stub file to provide type hints for the builtin object type, which IntelliJ/PyCharm/VsCode all support.

In order to find where to stub, open your IDE and Python project/file, type object to get a class, CTRL/CMD + Click and use "Find Definition" feature to get to the stub file of your IDE plugin.

Then copy/paste src/forbiddenfp/__init__.py under class object stub.

Note that for IntelliJ/Pycharm, once you CTRL + Click on object, you would need to click on the little icon below to get to the actual stub file. Append content there.

Warning

This library patches builtin object (and hence ALL classes), with hacks around CPython API (provided by forbiddenfruit),

so consider this NSFW (Not Safe For Work).

Known Issues

None doesn't work well with chained keyword arguments.

import forbiddenfp

None.apply(print)  # works
None.apply(func=print)  # doesn't work