python-constraint2 2.6.0

python-constraint is a module for efficiently solving CSPs (Constraint Solving Problems) over finite domains.

python-constraint

⚠️ IMPORTANT: this software must be installed with pip install python-constraint2, as the original pip release will not be updated.

A Conda Forge / Mamba version is also available: conda install python-constraint2.

For an overview of recent changes, visit the Changelog.

The complete documentation can be found here.

New: writing constraints in the new string format is preferable over functions and lambdas.

These strings, even as compound statements, are automatically parsed to faster built-in constraints, are more concise, and do not require constraint solving familiarity by the user to be efficient.

For example, problem.addConstraint(["50 <= x * y < 100"]) is parsed to [MinProdConstraint(50, ["x", "y"]), MaxProdConstraint(100, ["x", "y"])].

Similarly, problem.addConstraint(["x / y == z"]) is parsed to [ExactProdConstraint("x", ["z", "y"])].

This feature is in beta and subject to possible change, please provide feedback.

• Introduction

• Examples

• Features

• API documentation

• Download and install

• Testing

• Contributing

• Roadmap

• Contact

Introduction

The python-constraint module offers efficient solvers for Constraint Satisfaction Problems (CSPs) over finite domains in an accessible Python package. CSP is class of problems which may be represented in terms of variables (a, b, …), domains (a in [1, 2, 3], …), and constraints (a < b, …).

Examples

Basics

This interactive Python session demonstrates basic operations:

>>> from constraint import *
>>> problem = Problem()
>>> problem.addVariable("a", [1,2,3])
>>> problem.addVariable("b", [4,5,6])
>>> problem.getSolutions()  # doctest: +NORMALIZE_WHITESPACE
[{'a': 3, 'b': 6}, {'a': 3, 'b': 5}, {'a': 3, 'b': 4},
 {'a': 2, 'b': 6}, {'a': 2, 'b': 5}, {'a': 2, 'b': 4},
 {'a': 1, 'b': 6}, {'a': 1, 'b': 5}, {'a': 1, 'b': 4}]

>>> problem.addConstraint("a*2 == b") # string constraints are preferable over the black-box problem.addConstraint(lambda a, b: a*2 == b, ("a", "b"))
>>> problem.getSolutions()
[{'a': 3, 'b': 6}, {'a': 2, 'b': 4}]

>>> problem = Problem()
>>> problem.addVariables(["a", "b"], [1, 2, 3])
>>> problem.addConstraint(AllDifferentConstraint())
>>> problem.getSolutions()  # doctest: +NORMALIZE_WHITESPACE
[{'a': 3, 'b': 2}, {'a': 3, 'b': 1}, {'a': 2, 'b': 3},
 {'a': 2, 'b': 1}, {'a': 1, 'b': 2}, {'a': 1, 'b': 3}]

Rooks problem

The following example solves the classical Eight Rooks problem:

>>> problem = Problem()
>>> numpieces = 8
>>> cols = range(numpieces)
>>> rows = range(numpieces)
>>> problem.addVariables(cols, rows)
>>> for col1 in cols:
...     for col2 in cols:
...         if col1 < col2:
...             problem.addConstraint(lambda row1, row2: row1 != row2, (col1, col2))
>>> solutions = problem.getSolutions()
>>> solutions   # doctest: +NORMALIZE_WHITESPACE +ELLIPSIS
[{0: 7, 1: 6, 2: 5, 3: 4, 4: 3, 5: 2, 6: 1, 7: 0},
 {0: 7, 1: 6, 2: 5, 3: 4, 4: 3, 5: 2, 6: 0, 7: 1},
 {0: 7, 1: 6, 2: 5, 3: 4, 4: 3, 5: 1, 6: 2, 7: 0},
 {0: 7, 1: 6, 2: 5, 3: 4, 4: 3, 5: 1, 6: 0, 7: 2},
 ...
 {0: 7, 1: 5, 2: 3, 3: 6, 4: 2, 5: 1, 6: 4, 7: 0},
 {0: 7, 1: 5, 2: 3, 3: 6, 4: 1, 5: 2, 6: 0, 7: 4},
 {0: 7, 1: 5, 2: 3, 3: 6, 4: 1, 5: 2, 6: 4, 7: 0},
 {0: 7, 1: 5, 2: 3, 3: 6, 4: 1, 5: 4, 6: 2, 7: 0},
 {0: 7, 1: 5, 2: 3, 3: 6, 4: 1, 5: 4, 6: 0, 7: 2},
 ...]

Magic squares

This example solves a 4x4 magic square:

>>> problem = Problem()
>>> problem.addVariables(range(0, 16), range(1, 16 + 1))
>>> problem.addConstraint(AllDifferentConstraint(), range(0, 16))
>>> problem.addConstraint(ExactSumConstraint(34), [0, 5, 10, 15])
>>> problem.addConstraint(ExactSumConstraint(34), [3, 6, 9, 12])
>>> for row in range(4):
...     problem.addConstraint(ExactSumConstraint(34), [row * 4 + i for i in range(4)])
>>> for col in range(4):
...     problem.addConstraint(ExactSumConstraint(34), [col + 4 * i for i in range(4)])
>>> solutions = problem.getSolutions()  # doctest: +SKIP

Features

The following solvers are available:

• OptimizedBacktrackingSolver (default)

• BacktrackingSolver

• RecursiveBacktrackingSolver

• MinConflictsSolver

• ParallelSolver

Predefined constraint types currently available (use the parsing for automatic conversion to these types):

• FunctionConstraint

• AllDifferentConstraint

• AllEqualConstraint

• ExactSumConstraint

• MinSumConstraint

• MaxSumConstraint

• MinProdConstraint

• ExactProdConstraint

• MaxProdConstraint

• VariableExactSumConstraint

• VariableMinSumConstraint

• VariableMaxSumConstraint

• VariableMinProdConstraint

• VariableExactProdConstraint

• VariableMaxProdConstraint

• InSetConstraint

• NotInSetConstraint

• SomeInSetConstraint

• SomeNotInSetConstraint

API documentation

Documentation for the module is available at: http://python-constraint.github.io/python-constraint/. It can be built locally by running make clean html from the docs folder. For viewing RST files locally, restview is recommended.

Download and install

$ pip install python-constraint2

Conda / Mamba:

$ conda config --add channels conda-forge
$ conda config --set channel_priority strict
$ conda install python-constraint2
$ mamba install python-constraint2

Testing

Run nox (tests for all supported Python versions in own virtual environment).

To test against your local Python version: make sure you have the development dependencies installed. Run pytest (optionally add --no-cov if you have the C-extensions enabled).

Contributing

Feel free to contribute by submitting pull requests or opening issues. Please refer to the contribution guidelines before doing so.

Roadmap

This GitHub organization and repository is a global effort to help to maintain python-constraint, which was written by Gustavo Niemeyer and originaly located at https://labix.org/python-constraint. For an overview of recent changes, visit the Changelog.

Planned development:

• Support constant modifiers on parsed (variable) constraints instead defaulting to FunctionConstraint, e.g. problem.addConstraint("a+2 == b") or problem.addConstraint("x / y == 100")

• Parse using Abstract Syntax Trees (AST) instead of the current parser to be more robust and support decomposing lambdas

• Rewrite hotspots in C / Pyx instead of pure python mode

• Improvements to make the ParallelSolver competitive (experiments reveal the freethreading mode to be promising)

• Versioned documentation

Contact

• Floris-Jan Willemsen <[e-mail]> (current maintainer)

• Sébastien Celles <[e-mail]> (former maintainer)

• Gustavo Niemeyer <[e-mail]> (initial developer)

But it’s probably better to open an issue.