A Kalman Filter library written in Python. This package should be viewed as a modular framework consisting of 5 core modules:

1. matrix - a simple linear algebra module.
2. kfilter - linear Kalman Filters that can be iterated in one or two directions.
3. detector - a detector simulation module that implements different geometries that can respond to particles.
4. track - different types of physical tracks that can propagate inside the detector.
5. fitter - a module facilitating interaction between kfilter and detector

Aside from the core modules, tests exist that present the functionalities of matrix, kfilter, track and fitter.

The whole package can be conveniently imported by using:

from JKalFilter import *

This framework was only tested on Python 2.7.8. It will not work with Python 3 unaltered, however the required changes should not be significant.

The whole package can be installed using the command:

python setup.py install

from the main directory.

All tests are located in test/ subdirectories, and currently cannot be run from the command line by executing python test_<name>.py, one must execute them with the full package name from beyond the toplevel directory:

python -m JKalFilter.test.test_matrix

If anyone knows how to fix this, please open and issue and tell me.

If the test requires command line arguments the user will be informed of that.

1. test_matrix.py contains unit tests for all functions contained in the matrix module.
2. test_filter.py contains two tests of LKFilter used for filtering noise from a first and second degree polynomial. Number of iterations and function can be selected by passing appropriate command line arguments.
3. test_track.py contains a test of propagating tracks inside a detector. Needs matplotlib.
4. test_fitting.py contains a test of the fitting capabilities of the FitManager. Needs matplotlib.

The folder libtest contains a simple example of using ctypes to extend Python using C. Build with make and run test_libtest.py to see a comparison of run times.

The test suites test_filter, test_matrix and test_libtest can all be run automatically with py.test.

Documentation is available at github pages.