A Python binding to the levmar library.

The levmar is GPL'ed ANSI C implementation of the Levenberg-Marquardt (LM) optimization algorithm. The LM algorithm provides a numerical solution to the problem of minimizing a function over a parameter space of a function. The levmar library provides implementation of both unconstrained and constrained LM algorithms (box, linear equation, and linear inequality constraints).

Building Levmar requires the following software installed:

• Python (>=2.6)
• NumPy (>=1.3)
• [optional] nose (>=0.11)

nose is required to execute tests.

In order to build levmar, simply do:

$ python setup.py build
$ python setup.py install

Then, verify a successful installation:

$ python -c "import levmar; levmar.test()"

If you downloaded Levmar from a GitHub repository, you need to have Cython (>=0.13) installed.

$ cython -v levmar/_levmar.pyx
$ python setup.py build
$ python setup.py install
$ python -c "import levmar; levmar.test()"

If you just want to try Levmar without installing it, build it in-place:

$ (cython -v levmar/_levmar.pyx)
$ python setup.py build_ext --inplace -f
[Set up PYTHONPATH appropriately]
$ python -c "import levmar; levmar.test()"

See docstrings and demo scripts contained in the directory ./examples. Documentation of the levmar library can be found at http://www.ics.forth.gr/~lourakis/levmar/.

Takeshi Kanmae <tkanmae@gmail.com>

The MIT license applies to all the files except those in ./levmar-2.6. All of the software in ./levmar-2.6 and only the software therein is copyrighted by Manolis Lourakis and is licensed under the terms and conditions of the GNU General Public License (GPL). See the file LICENSE.txt.

• levmar: http://www.ics.forth.gr/~lourakis/levmar/
• Python: http://www.python.org/
• NumPy: http://www.scipy.org/
• nose: http://somethingaboutorange.com/mrl/projects/nose
• Cython: http://www.cython.org/