Kinetics Toolkit (ktk) is an open-source, pure-python biomechanical library developed by Professor Félix Chénier at Université du Québec à Montréal, Canada. It is a package of integrated classes and functions that aims to facilitate research in biomechanics using python. It does not attempt to provide a complete workflow from raw files to final analysis (although it may in a far future), or a main graphical user interface, or magical blackboxes that process everything automatically.
Kinetics Toolkit is mainly addressed to researchers and students in biomechanics with a little background in programming, who may or may not already have a working workflow and who want to understand and control their data. This is why special attention is made to API documentation and tutorials, and to ensure the interoperability of ktk with other environments (using pandas Dataframes as intermediate data containers).
>>> markers = ktk.kinematics.read_c3d_file('my_file.c3d') >>> ktk.Player(markers)
This library is a work in progress and is still very incomplete. The stable version API is however mostly settled and generally well tested using a comprehensive set of unit tests and doc tests, and currently includes:
timeseries.TimeSeries : a generic class to represent time-varying n-dimensional data and events, with many methods to extract, merge and subset TimeSeries data.
kinematics : a module that loads c3d and n3d files as TimeSeries of 3d marker positions.
player.Player : a class that allows visualizing 3d markers using a simple graphical user interface.
filters : a module that wraps some filters from scipy to be applied directly on TimeSeries.
save and load : two functions to save and load results to/from JSON-based ktk.zip files.
and other helper functions.
Please be warned that this is still mostly experimental software. If you are using ktk or are planning to be, you are warmly invited to contact me, first to say Hello :-), and so that I can warn you before doing major, possibly breaking changes. Also remind that I develop ktk mainly for my lab and I have limited resources for troubleshooting. You can however ask your questions and if I can answer, I’ll do.
You can also switch to the development version. This version is also open source, and is much more featured than the stable version. However, it is developed in parallel with my research projects following the needs of the moment and is therefore well less settled, less tested and less stable.
Some external code has been directly included into ktk’s source code. Here are the credits for these nice people.
Clay Flannigan : icp - Python implementation of m-dimensional Iterative Closest Point method
I also want to credit the people involved in ktk’s dependencies:
Benjamin Michaud : ezc3d - Easy to use C3D reader/writer for C++, Python and Matlab
The dedicated people behind major software and packages used by ktk such as python, numpy, matplotlib, pandas, jupyter, pytest, pdoc3, etc.
- Getting started with TimeSeries
- Saving and loading
- API Reference
- Release notes
- Development version