Pampy: Pattern Matching for Python

Pampy is pretty small (150 lines), reasonably fast, and often makes your code more readable and hence easier to reason about. There is also a JavaScript version, called Pampy.js.

You can write many patterns

Patterns are evaluated in the order they appear.

You can write Fibonacci

The operator _ means "any other case I didn't think of".

from pampy import match, _

def fibonacci(n):
    return match(n,
        1, 1,
        2, 1,
        _, lambda x: fibonacci(x-1) + fibonacci(x-2)
    )

You can write a Lisp calculator in 5 lines

from pampy import match, REST, _

def lisp(exp):
    return match(exp,
        int,                lambda x: x,
        callable,           lambda x: x,
        (callable, REST),   lambda f, rest: f(*map(lisp, rest)),
        tuple,              lambda t: list(map(lisp, t)),
    )

plus = lambda a, b: a + b
minus = lambda a, b: a - b
from functools import reduce

lisp((plus, 1, 2))                 	# => 3
lisp((plus, 1, (minus, 4, 2)))     	# => 3
lisp((reduce, plus, (range, 10)))       # => 45

You can match so many things!

match(x,
    3,              "this matches the number 3",

    int,            "matches any integer",

    (str, int),     lambda a, b: "a tuple (a, b) you can use in a function",

    [1, 2, _],      "any list of 3 elements that begins with [1, 2]",

    {'x': _},       "any dict with a key 'x' and any value associated",

    _,              "anything else"
)

You can match [HEAD, TAIL]

from pampy import match, HEAD, TAIL, _

x = [1, 2, 3]

match(x, [1, TAIL],     lambda t: t)            # => [2, 3]

match(x, [HEAD, TAIL],  lambda h, t: (h, t))    # => (1, [2, 3])

TAIL and REST actually mean the same thing.

You can nest lists and tuples

from pampy import match, _

x = [1, [2, 3], 4]

match(x, [1, [_, 3], _], lambda a, b: [1, [a, 3], b])           # => [1, [2, 3], 4]

You can nest dicts. And you can use _ as key!

pet = { 'type': 'dog', 'details': { 'age': 3 } }

match(pet, { 'details': { 'age': _ } }, lambda age: age)        # => 3

match(pet, { _ : { 'age': _ } },        lambda a, b: (a, b))    # => ('details', 3)

It feels like putting multiple _ inside dicts shouldn't work. Isn't ordering in dicts not guaranteed ? But it does because in Python 3.7, dict maintains insertion key order by default

You can match class hierarchies

class Pet:          pass
class Dog(Pet):     pass
class Cat(Pet):     pass
class Hamster(Pet): pass

def what_is(x):
    return match(x,
        Dog, 		'dog',
        Cat, 		'cat',
        Pet, 		'any other pet',
          _, 		'this is not a pet at all',
    )

what_is(Cat())      # => 'cat'
what_is(Dog())      # => 'dog'
what_is(Hamster())  # => 'any other pet'
what_is(Pet())      # => 'any other pet'
what_is(42)         # => 'this is not a pet at all'

All the things you can match

As Pattern you can use any Python type, any class, or any Python value.

The operator _ and built-in types like int or str, extract variables that are passed to functions.

Types and Classes are matched via instanceof(value, pattern).

Iterable Patterns match recursively through all their elements. The same goes for dictionaries.

Using strict=False

By default match() is strict. If no pattern matches, it raises a MatchError.

You can prevent it using strict=False. In this case match just returns False if nothing matches.

>>> match([1, 2], [1, 2, 3], "whatever")
MatchError: '_' not provided. This case is not handled: [1, 2]

>>> match([1, 2], [1, 2, 3], "whatever", strict=False)
False

Using Regular Expressions

Pampy supports Python's Regex. You can pass a compiled regex as pattern, and Pampy is going to run patter.search(), and then pass to the action function the result of .groups().

def what_is(pet):
    return match(pet,
        re.compile('(\w+)-(\w+)-cat$'),     lambda name, my: 'cat '+name,
        re.compile('(\w+)-(\w+)-dog$'),     lambda name, my: 'dog '+name,
        _,                                  "something else"
    )

what_is('fuffy-my-dog')     # => 'dog fuffy'
what_is('puffy-her-dog')    # => 'dog puffy'
what_is('carla-your-cat')   # => 'cat carla'
what_is('roger-my-hamster') # => 'something else'

Using Dataclasses

Pampy supports Python 3.7 dataclasses. You can pass the operator _ as arguments and it will match those fields.

@dataclass
class Pet:
    name: str
    age: int

pet = Pet('rover', 7)

match(pet, Pet('rover', _), lambda age: age)                    # => 7
match(pet, Pet(_, 7), lambda name: name)                        # => 'rover'
match(pet, Pet(_, _), lambda name, age: (name, age))            # => ('rover', 7)

Install

Currently it works only in Python >= 3.6 Because dict matching can work only in the latest Pythons.

I'm currently working on a backport with some minor syntax changes for Python2.

To install it:

$ pip install pampy

or $ pip3 install pampy