Downloads the demonstrations and setups a volume, so other docker containers can access it.

Getting Started

1. Create an .env File with:

DATA_ROOT=data_wombat
NUM_WATERFALL_DEMOS=100 
NUM_VILLAGE_DEMOS=100 
NUM_CAVE_DEMOS=100 
NUM_ANIMAL_DEMOS=100

where DATA_ROOT=data_wombat or DATA_ROOT=data

• data_wombat: Saves the on mounted shared wombat-server folder
• data: Saves on host server (Bison)

2. Run run.sh
3. The downloaded data is now available in the volume ka-basalt-2022-datadownloader_data. When running the core repo ka-basalt-2022, the project will access the data in that volume.
4. [The following is already done for the ka-basalt-2022 repo] You can add that volume to your project's docker-compose.yaml file:

volumes:
  ka-basalt-2022-datadownloader_data:
    external: true

Changing files after docker image was build

1. Start docker container in interactive mode

docker run -it --name kabasalt_data_downloader -v ka-basalt-2022-datadownloader_data:/home/data kabasalt_data_downloader /bin/bash

2. run download_dataset.py e.g.:

python download_dataset.py --json-file /home/resources/find-cave-Jul-28.json --output-dir /home/${DATA_ROOT}/MineRLBasaltFindCave-v0 --num-demos 100
python download_dataset.py --json-file /home/resources/build-house-Jul-28.json --output-dir /home/${DATA_ROOT}/MineRLBasaltBuildVillageHouse-v0 --num-demos 100
python download_dataset.py --json-file /home/resources/pen-animals-Jul-28.json --output-dir /home/${DATA_ROOT}/MineRLBasaltCreateVillageAnimalPen-v0 --num-demos 100
python download_dataset.py --json-file /home/resources/waterfall-Jul-28.json --output-dir /home/${DATA_ROOT}/MineRLBasaltMakeWaterfall-v0 --num-demos 100

follow download process

docker logs -f kabasalt_data_downloader