pydtw - lightweight constrained Dynamic Time Warping

I have written some Python bindings for the Euclidean- and Manhattan-flavoured constrained Dynamic Time Warping similarity measure in C++. The library supports warping path extraction:

warping path

Furthermore, the fast windowed min-max-filter by Daniel Lemire and LB_Keogh lower bounds are provided:

warping path

You can check out the source at Github. Happy time series mining!