uvlpy

A light package for constraint solving by levereging uv's dependency resolver

Example

Example of some basic code to show the features would be:

from uvlpy.uvlpy import System

sys = System()
a, b, c = sys.make_vars([0, 1], [2, 3, 4], list(range(0,9)))
a.val_constr(1, b == 2)
c.constr(b > 1)
sys.constr(a > 0, c < 9, b.to_constr())
print(sys.execute(a, b, c))

Where we:

• Create 3 variables with given possible values
• Constrain a = 1 to require b = 2
• Constrain all values of c to require b > 1
• Constrain the result to require a positive a, c < 9, and to include b
• After execution the output is e.g. [1, 2, 8], a and b can only be 1 and 2, but c can be any of its possible values (uv tends to prefer higher)

You can see the sudoku_example.py for a sudoku solver, which uses lists directly instead of destructuring them into variables like here.

Install

Simply pip install uvlpy, or with uv, uv pip install uvlpy.

Otherwise, you can build it locally with uv build and use the wheel as you see fit.

Note that uv being installed and visible on PATH is required for this library to function.

Why

Can be done.

uvlpy uses uv's solver, which should be reliable. The performance is not great, but sufficient.

However, note that the library creates a lot temporary files and packages to utilize uv, and as such, should be used with care and within reason.