pyadbcxy 0.0.4

An automatic differentiation package

Welcome to PyADBCXY

PyADBCXY is an automatic differentiation software package that can calculate the values and derivatives of complex functions while maintaining high levels of accuracy.

Contributors

Group Name: cs107-BCXY
Group Number: 12
Group Members:

• Charlie Harrington, Harvard University, [e-mail]
• Bowen Zhu, Harvard University, [e-mail]
• Yaxin Lei, Harvard University, [e-mail]
• Xiang Bai, Harvard University, [e-mail]

Installation

Install the package with pip:

pip install pyadbcxy

Then you can run a simple example such as

import pyadbcxy as ad
x = ad.Variable(3)            # instantiate x variable
y = ad.Variable(4., 5.)       # instantiate y variable
f = lambda x, y: x + y        # define function of interest
fmode = ad.Forward(f, (x, y)) # instantiate forward mode
fmode.calculate()             # evaluate function at x and y
print(fmode.value)            # print value of f at x and y
print(fmode.derivative)       # print derivative of f and x and y

For further details on package usage as well as the math behind it, please see the documentation.

For developers

To install the package for further development, clone the repository to your machine:

git clone https://github.com/cs107-BCXY/cs107-FinalProject.git

Then, move into the repository and install the package dependencies

cd cs107-FinalProject
pip install -r requirements.txt

Testing

To run the tests, you can execute

python -m unittest discover -s tests -p 'test_*.py'

Alternatively, you can run the test driver script. The default testing framework is unittest and can be run with:

bash tests/run_tests.sh

If you'd like to execute the tests using pytest, simply specify in the command:

bash tests/run_tests.sh pytest

Finally, if you'd like to see the code coverage report, you can do so with the coverage keyword:

bash tests/run_tests.sh coverage

License

This project is licensed under the MIT License - see the LICENSE for more details.