With this release a complete change of the code base was undertaken.
This includes the following major changes: - The C++ code base and cython interface were substituted by python style code that is compiled just in time using numba - The clustering algorithm was rewritten from scratch and is now 2 - 100 times faster (depending on the data); double hits per event are treated correctly - The API was simplified, different data types (e.g. column values can be given as int, float, uint etc.) are supported! - The unittests were changed - The task of the clusterizer was purified: the clusterizer only clusters hits and does not create histograms anymore
1.1.1
The clusterizer has now variable frame and charge ADC value ranges.
The constructor changes from clusterizer = HitClusterizer(n_columns=1000, n_rows=1000) to clusterizer = HitClusterizer(n_columns=1000, n_rows=1000, n_frames=1, n_charges=1)
1.1.0
This release has bug fixes, enhancements and a higher unit tests coverage.