added docker compose for easier set up

[rowellz]

Made a docker compose environment so users or contributors can more easily get H3DGS running on their own machines. It isn't the most ideal docker configuration, but the pay off is definitely worth it since all you need to run is docker-compose up and the environment will be running in a docker container. I tested this by using my own dataset and shelling into the container with docker exec -it hierarchical_3d_gaussians /bin/bash and successfully ran these commands on my own data:

python ./preprocess/generate_colmap.py --project_dir data/my_data

python ./preprocess/generate_chunks.py --project_dir data/my_data

python ./preprocess/generate_depth.py --project_dir data/my_data

python ./scripts/full_train.py --project_dir data/my_data

[rowellz]

Hi @White-Mask-230 thanks for reviewing/approving this, any idea what else needs to happen before merging this? I would like to add glomap to the Dockerfile to speed up the colmap process & to maybe work on a backwards compatible approach with exhaustive feature matching, but that is probably out of scope for this.

[White-Mask-230]

I don't think it's necessary to add anything more