README.md

Using Sample Notebooks

Environment Setup

To install the pre-configured python environment using the conda environment manager, you must first install either Miniconda or Anaconda.

From the command line, create a new conda environment using the provided conda-lock file specific to your platform, named conda-{platform}.lock. The conda-lock file contains the exact versions of all dependencies used to test this code. If you're unsure of your platform, run the following command and note the value listed after "platform:":

conda info

Specifically, the command to run for each supported platform are shown below:

# Create environment on x64 Linux.
conda create --name aurorabp --file conda-linux-64.lock

# Create environment on x64 Mac.
conda create --name aurorabp --file conda-osx-64.lock

# Create environment on ARM-based Mac.
conda create --name aurorabp --file conda-osx-arm64.lock

# Create environment on x64 Windows.
conda create --name aurorabp --file conda-win-64.lock

Alternatively, if you're using a different platform, or have trouble installing using the conda-lock file, as described above, you can install the latest versions of all dependencies using the environment.yml file:

conda env create --name aurorabp --file environment.yml

Running Notebooks

First, you must activate the conda environment created during Environment Setup by running the following command:

conda activate aurorabp

Next, start a Jupyter notebook server by running the following comamnd:

jupyter notebook

Notebooks

The following Jupyter notebook files are provided to give examples of data import, plotting, and prediction:

Example	File	Description
Data Plotting Example	data_plotting_example.ipynb	This example of sample data manipulation includes binning feature data into intervals and plotting histograms as well as feature comparisons.
Waveform Example	waveform_example.ipynb	Load sample measurement data, set up filters, and plot waveforms.
Prediction Example	prediction_example.ipynb	Example BP prediction from sample feature data. This notebook also demonstrates the correct setup of participant-stratified cross-validation.