SuDoKu solver and generator

This program is a command-line utility to resolve and generate SuDoKu grids.

Installation

Python >= 2.6 or >= 3.2 is required.

Run in a virtualenv (or as root for a systemwide install):

$ make install

Usage

To obtain detailed usage information, run:

$ sudoku --help

When resolving a SuDoKu, the grid can be specified as an argument, on the standard input, or in a file with the --input option. Known cells in the grid are represented by the corresponding figure, empty cells by one these characters: underscore, dash, space, dot, zero, and line breaks are ignored. If there are several solutions to the grid, all of them are shown.

When generating a SuDoKu, no arguments are necessary. It is possible to generate several grids at a time with the --count option.

If the --estimate option is set, an estimation of the difficulty of the grid is printed. The estimation is a (float, int) pair and the figures are the resolution time for a human and for a computer, respectively. Difficulty for a human starts at 1 and is generally stays below 6, although no upper limit is enforced. Difficulty for a computer is the number of resolution steps required by the algorithm used in the program.

The output format in which the grids are displayed can be selected with the --format option.

The --debug option is only available when using the Python implementation or when the C implementation is compiled in debug mode.

Demo

Run:

$ make demo

Tests

The following tests can be run:

$ make test         # Unit tests and comparison of the two implementations
$ make test-cover   # Code coverage of tests on the Python implementation
$ make test-memory  # Check for reference leaks in the C implementatino

If the test fail, run make clean and retry.

Programmer API

Internally, a Python module called sudoku is defined. Both a pure Python and an equivalent C implementation are available for this module. It contains two objects: a SuDoKu class and a Contradiction exception. The sudoku module will automatically use the best implementation available. Public methods of the SuDoKu class are documented in sudoku/pysudoku.py.

Copyright

Copyright (c) 2008-2014 Aymeric Augustin. All rights reserved.