map-binning 0.3.1

A map binning tool for spatial resampling

Map Binning Tool

A Python package for spatial resampling and binning of geospatial data, specifically designed for oceanographic datasets. This tool enables efficient downsampling of high-resolution gridded data onto coarser grids while preserving spatial accuracy through intelligent neighborhood averaging.

Overview

The Map Binning Tool provides a robust solution for spatial data aggregation, particularly useful for:

• Downsampling high-resolution oceanographic data (e.g., sea level anomaly, ocean currents)
• Creating consistent multi-resolution datasets
• Reducing computational load while maintaining spatial representativeness
• Processing time-series of gridded data efficiently

The package uses k-d tree algorithms for fast spatial queries and supports both in-memory processing and persistent caching of spatial indices for repeated operations.

Key Features

• Efficient Spatial Binning: Uses scipy's cKDTree for fast nearest-neighbor searches
• Flexible Grid Support: Works with any xarray-compatible gridded dataset
• Automatic Radius Calculation: Intelligently determines search radius based on target grid spacing
• Persistent Caching: Save and reuse spatial indices using pickle serialization
• Time Series Support: Handles datasets with temporal dimensions
• Memory Efficient: Processes large datasets without excessive memory usage
• Oceanographic Focus: Optimized for CMEMS and similar oceanographic data formats

Installation

From PyPI (Recommended)

pip install map-binning

With optional dependencies for development

pip install map-binning[dev]

Developer Installation

Using conda environment

# Create and activate conda environment
conda env create -f environment.yml
conda activate map-binning

# Install the package in development mode
pip install -e .

From source

git clone <repository-url>
cd map_binning
pip install -e .

Quick Start

Basic Usage

import xarray as xr
from map_binning import Binning

# Load your datasets
ds_high = xr.open_dataset('high_resolution_data.nc')
ds_low = xr.open_dataset('low_resolution_grid.nc')

# Initialize the binning tool
binning = Binning(
    ds_high=ds_high,
    ds_low=ds_low,
    var_name='sla',  # variable in the dataset to bin (e.g., sea level anomaly)
    xdim_name='longitude',  # longitude dimension name
    ydim_name='latitude',   # latitude dimension name
    search_radius=0.1  # optional: search radius in degrees
)

# Perform binning
result = binning.mean_binning()

Advanced Usage with Caching

# Create binning index and save it for reuse
result = binning.mean_binning(
    precomputed_binning_index=False,
    pickle_filename="my_binning_index.pkl",
    pickle_location="./cache"
)

# Reuse the saved index for subsequent operations
result = binning.mean_binning(
    precomputed_binning_index=True,
    pickle_filename="my_binning_index.pkl",
    pickle_location="./cache"
)

API Reference

Binning Class

Constructor Parameters

• ds_high (xr.Dataset): High-resolution source dataset
• ds_low (xr.Dataset): Low-resolution target grid dataset
• var_name (str): Name of the variable to bin
• xdim_name (str, optional): Longitude dimension name (default: 'lon')
• ydim_name (str, optional): Latitude dimension name (default: 'lat')
• search_radius (float, optional): Search radius in degrees (auto-calculated if None)

Methods

create_binning_index() Creates a spatial mapping between high and low resolution grids.

mean_binning(precomputed_binning_index=False, pickle_filename=None, pickle_location=None) Performs spatial binning using mean aggregation.

Parameters:

• precomputed_binning_index (bool): Use pre-saved spatial index
• pickle_filename (str): Filename for saving/loading spatial index
• pickle_location (str): Directory path for pickle files

Returns: xr.DataArray with binned data on the target grid

Project Structure

map_binning/
├── map_binning/           # Main package directory
│   ├── __init__.py        # Package initialization
│   ├── binning.py         # Core binning algorithms
│   ├── index_store.py     # Pickle serialization utilities
│   └── main.py            # Command-line interface
├── notebooks/             # Jupyter notebooks for examples
│   └── cmems_nrt_coastal_bin.ipynb
├── tests/                 # Unit tests
│   ├── __init__.py
│   └── test_main.py
├── pickle_folder/         # Default location for cached indices
├── pyproject.toml         # Project configuration
├── environment.yml        # Conda environment specification
├── .env.template          # Environment variables template
└── README.md              # This file

Configuration for CMEMS data download

Environment Variables

Copy .env.template to .env and configure:

# Copernicus Marine Service credentials (if using CMEMS data)
COPERNICUSMARINE_SERVICE_USERNAME=<your_username>
COPERNICUSMARINE_SERVICE_PASSWORD=<your_password>

Dependencies

Core dependencies:

• numpy: Numerical computing
• scipy: Scientific computing (k-d tree algorithms)
• xarray: Labeled multi-dimensional arrays
• netcdf4: NetCDF file I/O
• python-dotenv: Environment variable management

Development dependencies:

• pytest: Unit testing framework
• black: Code formatting
• flake8: Code linting
• mypy: Static type checking

Examples

Working with CMEMS Data

import xarray as xr
from map_binning import Binning

# Example with Copernicus Marine data
ds_high = xr.open_dataset('cmems_high_res_sla.nc')
ds_low = xr.open_dataset('cmems_low_res_grid.nc')

# Initialize for sea level anomaly processing
sla_binning = Binning(
    ds_high=ds_high,
    ds_low=ds_low,
    var_name='sla',
    xdim_name='longitude',
    ydim_name='latitude'
)

# Process and cache the result
binned_sla = sla_binning.mean_binning(
    pickle_filename="cmems_sla_index.pkl",
    pickle_location="./cache"
)

# Save the result
binned_sla.to_netcdf('binned_sla_data.nc')

Time Series Processing

The tool automatically handles time dimensions:

# Works seamlessly with time-varying datasets
# Input: (time, lat, lon) -> Output: (time, lat_low, lon_low)
result = binning.mean_binning()

Performance Considerations

• Memory Usage: The tool processes data in chunks and uses efficient numpy operations
• Spatial Index Caching: Save computed spatial indices to avoid recalculation
• Grid Resolution: Performance scales with the product of grid sizes
• Search Radius: Smaller radii improve performance but may miss relevant data points

Contributing

We welcome contributions! Please follow these steps:

1. Fork the repository
2. Create a feature branch (git checkout -b feature/amazing-feature)
3. Make your changes and add tests
4. Run the test suite (pytest)
5. Format your code (black map_binning/)
6. Submit a pull request

Development Setup

# Clone and setup development environment
git clone <repository-url>
cd map-binning-project
conda env create -f environment.yml
conda activate map-binning
pip install -e .[dev]

# Run tests
pytest

# Format code
black map_binning/

# Type checking
mypy map_binning/

Testing

# Run all tests
pytest

# Run with coverage
pytest --cov=map_binning

# Run specific test file
pytest tests/test_main.py

License

This project is licensed under the MIT License - see the LICENSE file for details.

Citation

If you use this tool in your research, please cite:

@software{map_binning_2024,
  author = {Chia-Wei Hsu},
  title = {Map Binning Tool: Spatial Resampling for Oceanographic Data},
  url = {https://github.com/chiaweh2/map_binning},
  version = {0.1.0},
  year = {2024}
}

Support

• Issues: Please report bugs and feature requests via GitHub Issues
• Documentation: Additional examples available in the notebooks/ directory
• Contact: Chia-Wei Hsu ([e-mail])

Acknowledgments

• Built with support for Copernicus Marine Environment Monitoring Service (CMEMS) data
• Utilizes scipy's efficient spatial algorithms
• Designed for the oceanographic research community