How-To Guides

These guides cover specific tasks and workflows. Each guide assumes you’ve already installed the package (Getting Started).

Common tasks:

• Querying Cloud-Optimised Data
  • Installation
  • Quick Start
  • Exploring Dataset Metadata
  • Querying Data
  • Visualizations
  • Advanced: Custom S3 Storage
  • Example Workflows
  • See Also
• Dataset Configuration
  • Create Dataset Configuration (semi-automatic)
  • Parquet Configuration from NetCDF file
  • Parquet Configuration from CSV file
  • Parquet Configuration with a Custom NetCDF Handler
  • Parquet Configuration from Parquet file
  • Zarr Configuration from NetCDF
  • Run Settings Options
  • AWS OpenData registry
• Adding the dataset to pyproject.toml
• Configuration Validation
  • Understanding Validation
  • Common Validation Errors
  • Required vs Optional Fields
  • Debugging Validation
• Cluster Configuration & Distributed Processing
  • The batch_size Paradox
  • The memory_limit Pitfall
  • nthreads — Thread Count per Worker
  • n_workers — Worker Count
  • worker_vm_types — VM Selection
  • Putting It All Together — Configuration Reference
  • Worked Examples
  • Troubleshooting
  • Additional Resources
• Notebooks
• MCP Server
  • MCP Server Installation
  • Starting the Server
  • Gemini CLI (Linux / Ubuntu)
  • GitHub Copilot CLI (Linux)
  • GitHub Copilot in VS Code (Linux)
  • Claude Desktop Configuration (macOS / Windows)
  • Environment Variables
  • Available MCP Tools
  • Available MCP Resources
  • Example AI Prompts
  • Notebook Builder Workflow
  • Known Code Pitfalls Avoided by the Server
  • Dataset–Notebook Mapping
  • Testing
• Update All Metadata Script
  • Usage
  • Output
  • Notes

I want to…

• Query and analyze cloud-optimized data: Querying Cloud-Optimised Data

• Create a cloud-optimized dataset from NetCDF/CSV files: Dataset Configuration

• Process large datasets in parallel: Cluster Configuration & Distributed Processing

• Write and share a Jupyter notebook for data analysis: Notebooks

• Set up the MCP (Model Context Protocol) server: MCP Server

• Update metadata in the AODN catalog: Update All Metadata Script

Each guide includes examples and troubleshooting tips for common issues.