plum-dispatch 2.6.1

Multiple dispatch in Python

Plum: Multiple Dispatch in Python

Everybody likes multiple dispatch, just like everybody likes plums.

The design philosophy of Plum is to provide an implementation of multiple dispatch that is Pythonic, yet close to how Julia does it. See here for a comparison between Plum, multipledispatch, and multimethod.

Note: Plum 2 is now powered by Beartype! If you notice any issues with the new release, please open an issue.

Installation

Plum requires Python 3.10 or higher.

pip install plum-dispatch

Documentation

See here.

What's This?

Plum brings your type annotations to life:

from numbers import Number

from plum import dispatch


@dispatch
def f(x: str):
    return "This is a string!"


@dispatch
def f(x: int):
    return "This is an integer!"


@dispatch
def f(x: Number):
    return "This is a number, but I don't know which type."

>>> f("1")
'This is a string!'

>>> f(1)
'This is an integer!'

>>> f(1.0)
'This is a number, but I don't know which type.'

>>> f(object())
NotFoundLookupError: `f(<object object at 0x7fd3b01cd330>)` could not be resolved.

Closest candidates are the following:
    f(x: str)
        <function f at 0x7fd400644ee0> @ /<ipython-input-2-c9f6cdbea9f3>:6
    f(x: int)
        <function f at 0x7fd3a0235ca0> @ /<ipython-input-2-c9f6cdbea9f3>:11
    f(x: numbers.Number)
        <function f at 0x7fd3a0235d30> @ /<ipython-input-2-c9f6cdbea9f3>:16

[!IMPORTANT] Dispatch, as implemented by Plum, is based on the positional arguments to a function. Keyword arguments are not used in the decision making for which method to call. In particular, this means that positional arguments without a default value must always be given as positional arguments!

Example:

from plum import dispatch

@dispatch
def f(x: int):
   return x

>>> f(1)        # OK
1

>> try: f(x=1)  # Not OK
... except Exception as e: print(f"{type(e).__name__}: {e}")
NotFoundLookupError: `f()` could not be resolved...

This also works for multiple arguments, enabling some neat design patterns:

from numbers import Number, Real, Rational

from plum import dispatch


@dispatch
def multiply(x: Number, y: Number):
    return "Performing fallback implementation of multiplication..."


@dispatch
def multiply(x: Real, y: Real):
    return "Performing specialised implementation for reals..."


@dispatch
def multiply(x: Rational, y: Rational):
    return "Performing specialised implementation for rationals..."

>>> multiply(1, 1)
'Performing specialised implementation for rationals...'

>>> multiply(1.0, 1.0)
'Performing specialised implementation for reals...'

>>> multiply(1j, 1j)
'Performing fallback implementation of multiplication...'

>>> multiply(1, 1.0)  # For mixed types, it automatically chooses the right optimisation!
'Performing specialised implementation for reals...'

Projects Using Plum

The following projects are using Plum to do multiple dispatch! Would you like to add your project here? Please feel free to open a PR to add it to the list!

• Coordinax implements coordinates in JAX.
• fasttransform provides the main building block of data pipelines in fastai.
• GPAR is an implementation of the Gaussian Process Autoregressive Model.
• GPCM is an implementation of various Gaussian Process Convolution Models.
• Galax does galactic and gravitational dynamics.
• Geometric Kernels implements kernels on non-Euclidean spaces, such as Riemannian manifolds, graphs, and meshes.
• LAB uses Plum to provide backend-agnostic linear algebra (something that works with PyTorch/TF/JAX/etc).
• MLKernels implements standard kernels.
• MMEval is a unified evaluation library for multiple machine learning libraries.
• Matrix extends LAB and implements structured matrix types, such as low-rank matrices and Kronecker products.
• NetKet, a library for machine learning with JAX/Flax targeted at quantum physics, uses Plum extensively to pick the right, efficient implementation for a large combination of objects that interact.
• NeuralProcesses is a framework for composing Neural Processes.
• OILMM is an implementation of the Orthogonal Linear Mixing Model.
• PySAGES is a suite for advanced general ensemble simulations.
• Quax implements multiple dispatch over abstract array types in JAX.
• Unxt implements unitful quantities in JAX.
• Varz uses Plum to provide backend-agnostic tools for non-linear optimisation.

See the docs for a comparison of Plum to other implementations of multiple dispatch.