geebeam Documentation

geebeam is a Python package for building geospatial training datasets from Google Earth Engine.

Contents

• Getting Started
  • What is geebeam?
  • Prerequisites
  • Installation
  • Authentication
  • Step 1: Define the image you want
  • Step 2: Run the pipeline
  • Step 3: Inspect the output
  • Step 4. Adding a second image
  • What’s next
  • Get Help / Contribute
• Output Types
  • GeoTIFF (default)
  • WebDataset
  • TensorFlow Datasets
  • TFRecord
• Scaling up with Dataflow
  • Some DataFlow Gotchas
• EE’s 50 MB per-request limit
• API Reference
  • geebeam
    • Submodules
      • geebeam.pipeline
        • Functions
        • Module Contents
          • run_pipeline()
          • sample_and_run_pipeline()
          • grid_and_run_pipeline()
      • geebeam.sampler
        • Functions
        • Module Contents
          • sample_region_random()
          • sample_region_grid()
          • split_sets()

Note

This project is under active development.