Add Unit Tests and Codecov Integration for Backwards Compatibility

[thomaskleiven]

Hi,

I've noticed that the library is updated frequently, which is great! However, without unit tests, it can be challenging to ensure that these updates don't break existing functionality. Adding unit tests would help maintain backward compatibility and improve the overall stability of the library.

Benefits:

• Ensure new changes don't introduce bugs
• Increase reliability and stability
• Make it easier for new contributors to understand and contribute to the codebase

Request:

• Integrating a test framework like Google Test
• Add Codecov integration to track test coverage and maintain quality over time

I'm happy to help set up the testing framework, Github Actions or contribute some initial tests if needed.

Thanks for considering this!

[cyrusbehr]

This is a good idea. I've added you as a contributor to the project so you can create a branch and commit directly in the repo and open a PR.

I like the idea of unit tests, as well as having them run in a Github actions. That being said, you'll need to check if GPU enabled runners are provided with the free plan. If so, we can run the unit tests as part of the CI pipeline. If not, just compiling the code will suffice.

[thomaskleiven]

I suggest we start by setting up a centralized build process and then run unit tests. Since GitHub Actions don't support GPU-enabled runners, how about using a self-hosted runner? We can build and run the project on a machine with CUDA, OpenCV with CUDA support, and TensorRT.

I can set up a Jetson Orin NX 16GB I have lying around to start with, in case that would be an option. We can set up the project within a container to ensure a consistent environment, using CUDA 12.2, cuDNN 8.9, TensorRT 8.6, and OpenCV 4.8.0 with CUDA support. What do you think?

[cyrusbehr]

I like the idea