setup.sh

#!/bin/bash
set -e

echo "=========================================="
echo "Gillies Workshop Hackathon 2025 Setup"
echo "=========================================="
echo ""

# Color codes for output
RED='\033[0;31m'
GREEN='\033[0;32m'
YELLOW='\033[1;33m'
NC='\033[0m' # No Color

# Function to print colored output
print_status() {
    echo -e "${GREEN}[INFO]${NC} $1"
}

print_warning() {
    echo -e "${YELLOW}[WARN]${NC} $1"
}

print_error() {
    echo -e "${RED}[ERROR]${NC} $1"
}

# Check if running on a CUDA-enabled system
print_status "Checking CUDA availability..."
if ! command -v nvidia-smi &> /dev/null; then
    print_warning "nvidia-smi not found. Make sure you're running on a GPU instance."
else
    nvidia-smi
    print_status "CUDA detected successfully!"
fi

# Install Miniconda if not already installed
if ! command -v conda &> /dev/null; then
    print_status "Miniconda not found. Installing Miniconda..."

    # Check if miniconda3 directory exists but conda is not in PATH
    if [ -d "$HOME/miniconda3" ]; then
        print_warning "Miniconda directory exists but conda not in PATH. Removing old installation..."
        rm -rf "$HOME/miniconda3"
    fi

    # Download Miniconda installer
    wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh -O /tmp/miniconda.sh

    # Install Miniconda
    bash /tmp/miniconda.sh -b -p $HOME/miniconda3

    # Initialize conda
    eval "$($HOME/miniconda3/bin/conda shell.bash hook)"
    conda init bash

    # Clean up
    rm /tmp/miniconda.sh

    print_status "Miniconda installed successfully!"
else
    print_status "Miniconda already installed."
    eval "$(conda shell.bash hook)"
fi

conda tos accept --override-channels --channel https://repo.anaconda.com/pkgs/main
conda tos accept --override-channels --channel https://repo.anaconda.com/pkgs/r


# Create conda environment
ENV_NAME="hackathon"
print_status "Creating conda environment: $ENV_NAME..."

if conda env list | grep -q "^$ENV_NAME "; then
    print_warning "Environment $ENV_NAME already exists. Removing it..."
    conda env remove -n $ENV_NAME -y
fi

# Create environment with Python 3.13 (compatible with RAPIDS)
conda create -n $ENV_NAME python=3.13 -y
print_status "Conda environment created successfully!"

# Activate the environment
print_status "Activating conda environment..."
source activate $ENV_NAME

# Install uv for faster pip installs
print_status "Installing uv (fast Python package installer)..."
curl -LsSf https://astral.sh/uv/install.sh | sh
export PATH="$HOME/.local/bin:$PATH"
print_status "uv installed successfully!"

# Upgrade pip
print_status "Upgrading pip..."
pip install --upgrade pip

# Install RAPIDS packages from NVIDIA PyPI using uv
print_status "Installing RAPIDS packages with uv (this may take several minutes)..."
uv pip install \
    --extra-index-url=https://pypi.nvidia.com \
    "cudf-cu13==25.10.*" \
    "dask-cudf-cu13==25.10.*" \
    "cuml-cu13==25.10.*" \
    "cugraph-cu13==25.10.*" \
    "nx-cugraph-cu13==25.10.*" \
    "cuxfilter-cu13==25.10.*" \
    "cucim-cu13==25.10.*" \
    "pylibraft-cu13==25.10.*" \
    "raft-dask-cu13==25.10.*" \
    "cuvs-cu13==25.10.*"

print_status "RAPIDS packages installed successfully!"

# Install honeybee-ml
print_status "Installing honeybee-ml..."
uv pip install honeybee-ml

# Install JupyterLab and common data science packages
print_status "Installing JupyterLab and essential packages..."
uv pip install jupyterlab ipywidgets matplotlib seaborn plotly scikit-learn pandas numpy

# Install HuggingFace CLI with hf_transfer for faster downloads
print_status "Installing HuggingFace CLI with hf_transfer acceleration..."
uv pip install huggingface_hub[cli,hf_transfer]

# Create data directory
DATA_DIR="$HOME/hackathon-data"
print_status "Creating data directory at $DATA_DIR..."
mkdir -p "$DATA_DIR"

# Download datasets from HuggingFace
print_status "Downloading datasets from HuggingFace (Lab-Rasool/hackathon)..."
print_warning "Note: Dataset downloads are large (~694GB total). This may take a while..."
print_status "Using hf_transfer for accelerated downloads (optimized for high-bandwidth connections)..."

# Enable hf_transfer for faster downloads
export HF_HUB_ENABLE_HF_TRANSFER=1

# Download entire dataset (CSV files + images)
print_status "Downloading training and test datasets..."
huggingface-cli download Lab-Rasool/hackathon \
    --repo-type dataset \
    --local-dir "$DATA_DIR"

print_status "Datasets downloaded to $DATA_DIR"

# Verify datasets
if [ -f "$DATA_DIR/train/train.csv" ] && [ -d "$DATA_DIR/train/images" ]; then
    print_status "Training dataset verified: CSV and images found"
else
    print_warning "Training dataset incomplete. Please check $DATA_DIR/train/"
fi

if [ -f "$DATA_DIR/test/test.csv" ] && [ -d "$DATA_DIR/test/images" ]; then
    print_status "Test dataset verified: CSV and images found"
else
    print_warning "Test dataset incomplete. Please check $DATA_DIR/test/"
fi

echo ""
echo "=========================================="
echo "${GREEN}Setup Complete!${NC}"
echo "=========================================="
echo ""
echo "Environment name: $ENV_NAME"
echo "Data directory: $DATA_DIR"
echo ""
echo "To activate the environment in a new shell, run:"
echo "  conda activate $ENV_NAME"
echo ""
echo "Dataset location: $DATA_DIR"
echo "  - Training data: $DATA_DIR/train/"
echo "  - Test data: $DATA_DIR/test/"
echo ""
echo "Happy coding! 🚀"
echo ""