We decided on scikit-surgery as the main Project name (pip install skicit-surgery at command line) . Convention dicates that the python package is then called sksurgery (import sksurgery in a python script).
Different packages will be created, to minimise unncessary dependencies on other Python packages. Some inital thoughts were:
algorithms, data structures etc. Only depends on the core scikit packages as defined on https://www.scipy.org/scikits.html (numpy, scipy, matplotlib, pandas, ipython, sympy)
Image & video capture/processing using OpenCV
Visualisation using VTK/OpenGL etc.
GUI elements using qt
Interface to tracker (Atracsys in this case - similar naming convention for other hardware). Likely to have dependency on external C/C++ code
Other than a dependency of scikit-surgery-core, packages can be used standalone. For convenience, we can also make a master scikit-surgery/sksurgery package, that simply installs the latest versions of all the other packages.
For packages which have dependencies on existing C/C++ libraries, there seem to be a range of solutions, depending on the complexity of the code being wrapped: