RustDTW

<h1 align="center"> RustDTW </h1> <p align="center">Python extension backed by a multi-threaded Rust implementation of Dynamic Time Warping (DTW).</p> <div align="center"> <img src="https://img.shields.io/pypi/v/rust-dtw?style=flat-square"/><br/> </div>

⚡️ Quick Install

To install rustDTW, simply:

pip install rust-dtw

❓ What is Dynamic Time Warping

In time series analysis, dynamic time warping (DTW) is one of the algorithms for measuring similarity between two temporal sequences, which may vary in speed 1. This has applications in speech recognition, time series classification and neuroscience.

Example Usage

rustDTW was designed for usage with timeseries data from functional brain regions. However any data represented as a numpy matrix can be provided.

import numpy as np
import rust_dtw
  
rust_dtw.dtw(
  s=np.array([0., 1., 2.]), 
  t=np.array([3., 4., 5.]), 
  window=50, 
  distance_mode="euclidean"
  )
>>> 5.0990195

For more examples please see examples/ or explore the wiki.

Developing

Built With

• PyO3
• Maturin
• rust-numpy
• Rayon

Setting up Dev

To get started with development, simply clone the repository and edit the main library code in src/. Once done, simply build and test the code with ./build.sh.

git clone https://github.com/FL33TW00D/rustDTW.git
cd rust-dtw/
./build.sh

Tests

All tests are implemented using pytest.

poetry run pytest

📈 Performance

<div align="center"> <img src="https://raw.githubusercontent.com/FL33TW00D/rustDTW/master/examples/speed/time.png" width=60%/><br/> </div> The above shows the performance of the rustdtw implementation vs the DTAIDistance OpenMP Python version, showing a ~10x speed improvement.

⚠️ License

rustDTW is free and open-source software licensed under the MIT License.