mph-deep-learning-pipeline

A data pipeline that represents point clouds as multi-parameter persistent homology landscapes for deep learning models

Python3 virtual environment

Before using the code it is best to setup and start a Python virtual environment in order to avoid potential package clashes using the requirements file:

# Navigate into the repository root directory

# Create a virtual environment
python3 -m venv <env-name>

# Activate virtual environment
source <env-name>/bin/activate

# Install dependencies for code
pip3 install -r requirements.txt

# When finished with virtual environment
deactivate

Installations

Xming

If the user is a Windows Subsystem for Linux (WSL) user they will need to install Xming in order to have a viewer to visualise the persistence diagrams generates by Rivet.

Rivet

Full instructions for installation of the Rivet appication is provided by the following documentation when building the C++ application from the master branch.

Multi-parameter persistence landscapes

The pipeline will be extensively using the code base from Multiparameter_Persistence_Landscapes in order to generate the multi-parameter persistent landscapes. No installations required from this submodule, simply the preceeding installations are required at this point for this submodule to be used.

Run pipeline

The follwowing commands to make the pipeline BASH script executable to run the end to end pipeline with the desired configurations

# Navigate into the repository root directory

cd src/

chmod 744 pipeline.sh

./pipeline.sh