Add docker scripts to gaussian-splatting

[AmmarkoV]

First of all thanks for providing this amazing work to the Computer Vision Community!

This pull request adds the capability of building a docker container for gaussian-splatting

To install docker with NVIDIA/GPU pass-through support one can follow the following instructions :
https://www.howtogeek.com/devops/how-to-use-an-nvidia-gpu-with-docker-containers/

The docker subfolder is added to the project features 5 scripts/items :

docker/build_and_deploy.sh
Running this script builds and deploys "gaussian-splatting-container" container. It prepares and installs required libraries, starts and then attaches the container making all the required setup ready for use after issuing just one command!

docker/Dockerfile
This is the deployed docker file that builds on nvidia/cuda:12.2.0-devel-ubuntu22.04 and creates a docker container with the required system libraries to run the project

docker/initialize.sh
This is a BASH script called internally during container setup to install CERES, colmap and the python dependencies of the project

docker/run.sh
This is a script that automatically processes and then trains on a video file.

Issuing :
./run.sh VIDEOFILE.MOV kickstarts the procedure storing output in VIDEOFILE.MOV-data/output

docker/setup.sh
Optional script to perform installation of the docker server

[dfldylan]

Thank you for your work

[tfiliano]

Hi there, I have tried to run it and I am getting some errors,

Could you give any tip on how to make it work?