C1 Coursework

This is my repository for the C1 coursework.

The report for this project is in pdf format and is located in the report directory.

dual_autodiff

The dual_autodiff package is a Python package designed to perform automatic differentiation using dual numbers.

Features

• Dual class: Core attributes and methods. It is used to create dual numbers and implement arithmetic operations on them as well as some common mathematical functions.

Installation

Ensure you have Python 3.9 or higher. You will need Python 3.10 or 3.11 to install the Cython wheels. You can install the Python package from within the root directory of the repository by running:

pip install -e .

This will also install the required dependencies (numpy).

To install the Cythonized version of the package, navigate to the Cython project folder dual_autodiff_x and run:

pip install -e .

Alternatively, you can install directly from the wheels, located in the dual_autodiff_x/wheelhouse directory.

Usage

A quick example of how to use the package is shown below.

import dual_autodiff as df 
x=df.Dual(2,1)
print(x.sin())
y=df.Dual(3,4)
print(x+y)

For a more detailed introduction, see the tutorial notebook located in the docs/Tutorial directory. Please also refer to the documentation for more information.

Documentation

To build the documentation, navigate to the docs directory and run:

make html

The documentation will be located in the docs/_build/html directory. Open the index.html file in a web browser to view the documentation.

Testing

To run the tests for the Python package, run the following command from the root directory:

pytest -s tests/*

To check the Cython package is correctly installed, you can open the cython_test.ipynb notebook located in the root directory and run the cells in there.

How to run the Jupyter Notebooks

You will need to install the required packages to run the Jupyter notebooks in this repository. A simple way to guarantee that the notebooks will run is to create a new conda environment using the environment.yml file provided.

Clone this GitLab repository to your local machine.

git clone https://gitlab.developers.cam.ac.uk/phy/data-intensive-science-mphil/assessments/c1_coursework/fm565.git

Create a conda environment by running:

conda env create -f environment.yml

in the root directory of this repository.

This will create a new conda environment called C1Coursework. This will install the necessary packages for this project, listed in the requirements.txt file.

Activate the environment by running:

conda activate C1Coursework

This may automatically create a Jupyter Kernel for the new environment. If not, you can create a kernel manually e.g.

python -m ipykernel install --user --name C1Coursework --display-name "C1Coursework (Python 3.11)"

You should now be able to run the notebooks in this repository.

To deactivate the conda environment, run conda deactivate.

Declaration of Use of Autogeneration Tools

Microsoft Copilot was used in the following cases:

• In generating some of the test cases in the test suite where they became repetitive. For example, where I have written tests for the addition of dual numbers, Copilot was used to autocomplete suggestions for subtraction, multiplication, and division of dual numbers. These were then manually checked for errors.
• In generating some of the alternative wordings in the docstrings for the Dual class, which were then modified before being used.
• To modify the way in which dual numbers are displayed, Copilot suggested the use of the __repr__ method which I then implemented manually.
• To implement integers and floats as dual numbers, Copilot suggested the __radd__ etc. methods which I then implemented manually. It also suggested the use of 'isinstance' to check if the input is a int or float, which I then implemented manually in each case.
• To separate the Dual class member functions in autodiff_tools into a separate file, Copilot suggested importing each function into the main file. I then implemented this manually for the Dual class.

A declaration of the use of generative tools in writing the report is given in the report itself.