Environments for developing and deploying the app and submodules

[dmccloskey]

As a developer, I would like to have a quick start environment for developing the app and/or individual submodules and for deploying the app and selected submodules on either GPU or CPU resources. CPU deployment is much easier (except for defining BLAS and other CPU specific acceleration libraries), but GPU deployment is annoying (i.e., defining the OS specific CUDA libraries and dependencies).

Options for defining the environment

1. OS specific instructions with a conda environment
2. Docker container

Potential solutions

1. Take inspiration from the conda environment for developing using the NVIDIA rapids library (See https://github.com/rapidsai/cuvs?tab=readme-ov-file#installing-cuvs)
2. Take inspiration from the NVIDIA docker images https://hub.docker.com/layers/nvidia/cuda/12.6.2-devel-ubuntu22.04/images/sha256-9002ff3326a6b4e904d6d94b6a1d3b83608df02678923931e38fffee0a1089d9?context=explore)