Dynamic time warping algorithm(s)

Official documentation can be found here: https://cgodin-dev.gitlabpages.inria.fr/dtw

Installing the module from sources

Clone the sources

Start by cloning the sources with:

git clone https://gitlab.inria.fr/cgodin-dev/dtw.git

Conda environment creation (optional)

If you don't have an existing conda environment, create one (named dtw) with:

conda env create -f conda/env/dtw.yaml

You can find the official instructions on how to manually create an environment here.

Install DTW library

Start by activating your (dtw) environment with:

conda activate dtw

Then install the package in develop mode with pip,

(dtw)> python -m pip install -e .

Note: -e is used here to install the code in "develop mode", this way you do not have to re_install the package every time you makes modifications to he sources.

Testing the code

To run an example file, go in data-analysis directory and launch ipython.

cd data-analysis
ipython

Then you can run the arabido-test.py test file with:

ipython> %run arabido-test.py

Publish master branch to ROMI GitHub

The first time:

0. Create a GitHub empty repo
1. Clone this repository and go to the directory
2. Add a 'romi' remote pointing to the empty GitHub repo: git remote add romi https://github.com/romi/dtw.git

To update GitHub master from Inria GitLab

3. Push your modifications to GitHub (from the repository root): git push romi master

Conda packaging

Requirement

Install conda-build, in the base environment, to be able to build conda packages:

conda deactivate
conda install conda-build

⚠️ For macOS, follow these instructions to install the required macOS 10.9 SDK.

Build a conda package

Using the given recipes in dtw/conda/recipe, it is easy to build the conda package:

cd dtw/conda/recipe
conda build .

⚠️ This should be done from the base environment!

Conda useful commands

Purge built packages:

conda build purge

Clean cache & unused packages:

conda clean --all