Routines to calculate empirical cumulative distribution functions from data as described by Langrené and Warin (2020).
The source files are from their StOpt library and are incorporated here as a small standalone library that only has eigen3 as dependence and only needs boost for the testing functions but not to install as a python package.
After downloading/cloning this repository
python setup.py install
should be all that is needed to have your c++ compiler compile the package and install it to your distribution.
The two python commands that get exposed by this library are fastCDF and fastCDFOnSample.
Have a look at NotebookFastCDF.ipynb
We also provide a simple Makefile to compile it the commandline test functions.
Is another 2 dimensional example using one million points.