Gradpy: A tool for automatic differentiation

cs207-FinalProject

Group Name: DualSapiens

Group Number: 13

Members: Jovana Andrejevic, Gopal Kotecha, Jay Li, Ziyi Zhou

This is the main repository for the gradpy package, which includes an autodiff module for automatic differentiation, and a math module for compatibility with special functions. This repository also hosts the therapy_planner package, an application of automatic differentiation for dose delivery optimization relevant to Intensity Modulated Radiation Therapy (IMRT), that automatically ships with gradpy. The latest documentation for gradpy and the featured application therapy_planner is hosted on readthedocs.

Quick gradpy installation guide

1. We suggest working with our packages within a virtual environment. To do so, ensure that virtualenv for Python 3 has been installed.

2. Create a new virtual environment env:

virtualenv env --python=python3

3. Activate the environment:

source env/bin/activate

4. Install gradpy:

pip install gradpy

5. Users can now try the examples shown in the Usage section of the documentation to get started!

Testing gradpy

After installation, users may wish to run tests to validate their installed package is working properly. gradpy comes with a test suite that may be easily run using pytest.

1. Within the virtual environment in which gradpy has been installed, install pytest:

pip install pytest

note: A terminal restart after installing pytest is likely necessary for changes to take effect.

2. Run the gradpy test suite:

pytest --pyargs gradpy

Quick therapy_planner installation guide

therapy_planner may be installed in the same manner as gradpy.

1. Within a virtual environment, install therapy_planner:

pip install therapy_planner

2. Users can now follow the demos in the Featured Application section of the documentation.

Testing therapy_planner

1. Ensure that pytest is installed in the virtual environment.

2. Run the therapy_planner test suite:

pytest --pyargs therapy_planner