dhidb

dhidb provides read-only Python access to a global 300 m, 12-year (2014-2025) Dynamic Habitat Indices time series stored as a dense TileDB array on public S3-compatible object storage. It supports point, bounding-box, and polygon queries without first downloading the complete database.

DHIs are widely used in spatial ecology as interpretable predictors of habitat quality and vegetation dynamics, including covariates in species distribution models (SDMs) and biodiversity assessments.

Installation

pip install dhidb

For development:

python -m pip install -e ".[test]"

Requirements and optional dependencies

DHIDB requires Python 3.10 or newer. The runtime dependencies and the compatibility ranges published by the package are:

Optional features are installed with package extras. There is no generic dhidb[dependency] extra; use the feature-specific extra you need:

For local development and testing:

python -m pip install -e ".[test,netcdf,zarr]"

The documentation website has its own Node.js dependencies in website/package.json; they are not part of the Python package extras.

Quick start

from dhidb import DHIProvider

with DHIProvider() as db:
    print(db.years)
    print(db.variables)

    for batch in db.iter_bbox(
        bounds=(12.25, 51.25, 12.50, 51.40),  # Leipzig, Germany
        years=range(2014, 2026),
        variables=["dhi_cum", "dhi_min", "dhi_var", "valid_count"],
        batch_shape={"y": 256, "x": 256},
    ):
        print(batch.sizes)

Documentation

The Docusaurus website source is in website/. Build it locally with:

npm --prefix website install
npm --prefix website run start

Data variables

License

MIT