curryparty 0.2.0

Python playground to learn lambda calculus

Curry party

curryparty is a library created to explore, visualize and teach lambda-calculus concepts.

Install

Run pip install curryparty or uv add curryparty depending on your package manager.

How to use

from curryparty import L, V

# build expressions idiomatically
identity = L("x")._("x").build()

zero = L("f", "x")._("x").build()

omega = L("x")._("x").call("x").build()

# you can create more complex terms:

succ = L("n", "f", "x")._("f").call(
    V("n").call("f").call("x")
)

# If you try to combine them, nothing will happen:
bomb = omega(omega)
one = s(zero)

# You need to beta-reduce them:
bomb.beta()
one_with_first_reduction = one.beta()

# if you want the final form:
term = s(zero)
while term is not None:
    term = term.beta()
    print(term)

# this is equivalent to:
for x in term.reduction_chain():
    print(x)

But the main point of this library is the svg-based display system. If you use a notebook such as jupyternotebook or marimo, you will see something like this:

You can also use term.show_reduction to get an animated version.

Tutorial

lambda.py shows a tutorial in marimo format. Click on the button to try it out:

How it works

Under the wood, all the terms are converted into a list of nodes, that can either be a lambda, an application (with 2 arguments) or a variable (with 0 arguments).

For efficiency, they are converted to a polars dataframe in prefix traversal order. If you want to understand how it works, start by looking at the .nodes attribute of any term.