base_image: {{ env["RAY_IMAGE_ML_NIGHTLY_GPU"] | default("anyscale/ray-ml:nightly-py37-gpu") }}
env_vars:
  # Manually set NCCL_SOCKET_IFNAME to "ens" so NCCL training works on
  # anyscale_default_cloud.
  # See https://github.com/pytorch/pytorch/issues/68893 for more details.
  NCCL_SOCKET_IFNAME: ens
debian_packages:
  - curl

python:
  pip_packages:
    - pytest
    - xgboost_ray
    - petastorm
    - modin==0.12.1
  conda_packages: []

post_build_cmds:
  - pip3 uninstall -y ray || true && pip3 install -U {{ env["RAY_WHEELS"] | default("ray") }}
  - pip3 install -U --force-reinstall --no-deps xgboost xgboost_ray petastorm  # Avoid caching
  - {{ env["RAY_WHEELS_SANITY_CHECK"] | default("echo No Ray wheels sanity check") }}
  - sudo mkdir -p /data || true
  - sudo chown ray:1000 /data || true
  - rm -rf /data/classification.parquet || true
  - curl -so create_test_data.py https://raw.githubusercontent.com/ray-project/ray/releases/1.3.0/release/xgboost_tests/create_test_data.py
  - python create_test_data.py /data/classification.parquet --seed 1234 --num-rows 1000000 --num-cols 40 --num-partitions 100 --num-classes 2