topozarr 0.0.7

Lightweight multiscale zarr

topozarr - lightweight multiscale zarr pyramids

Python library to create multiscale Zarr pyramids for usage with zarr-layer.

Follows these zarr-conventions:

• multiscales — pyramid structure and resolution levels
• proj: — coordinate reference system (CRS)
• spatial: — affine transform, bounding box, and dimension names

Warning: experimental

Usage

Installation

uv add topozarr
# or
pip install topozarr

Note that zarr is not a core dependency — install it (or icechunk) to write pyramids to storage. The tutorial extra includes everything needed to run the examples below:

uv add 'topozarr[tutorial]'
# or
pip install 'topozarr[tutorial]'

Example

import xarray as xr
import xproj # for crs assignment
from topozarr import create_pyramid

# Load the air_temperature Xarray tutorial dataset
ds = xr.tutorial.open_dataset('air_temperature', chunks="auto")
ds = ds.proj.assign_crs(spatial_ref="EPSG:4326")
print(ds)

pyramid = create_pyramid(
    ds,
    levels=2,
    x_dim="lon",
    y_dim="lat",
    method="mean",  # "mean" (default) | "max" | "min" | "sum"
)
print(pyramid.encoding)
print(pyramid.dt)

levels is the total number of resolution levels, including the original. Level 0 is the original (highest) resolution; each subsequent level is coarsened by a factor of 2 per spatial dimension, so the last level is the coarsest.

Visualization hints

Use layer_hints to embed colormap and color range hints for zarr-layer directly in the pyramid metadata:

from topozarr.metadata import ZarrLayerVarConfig

pyramid = create_pyramid(
    ds,
    levels=2,
    x_dim="lon",
    y_dim="lat",
    layer_hints={"air": ZarrLayerVarConfig(colormap="blues", clim=[230, 310])},
)

These are written into the root zarr-layer metadata key and are optional — omitting layer_hints has no effect on the pyramid structure or encoding.

Chunking

create_pyramid returns a Pyramid with two attributes: pyramid.dt (the DataTree) and pyramid.encoding (recommended chunk and shard sizes per variable per level). Always pass pyramid.encoding as the encoding argument when writing — this is what applies the chunking strategy to the output store.

# Inspect the recommended encoding before writing
print(pyramid.encoding)

The heuristics target ~500KB chunks for web visualization. You can tune shard size with chunks_per_shard, the number of chunks per shard along each spatial dimension (default: 4, giving 4×4 = 16 chunks per shard and ~8MB shards). Valid values are powers of 2: 1, 2, 4, 8, 16, 32. Larger shards reduce task graph overhead when using Dask but increase memory usage.

chunks_per_shard	chunks/shard	approx shard size
1	1	~500KB
4	16	~8MB (default)
8	64	~32MB
16	256	~128MB

Pass chunks_per_shard=None to disable sharding entirely.

Writing

Always pass pyramid.encoding to apply the recommended chunking:

# Write to Zarr
from obstore.store import from_url
from zarr.storage import ObjectStore

store = ObjectStore(from_url(url="<your_bucket_url>", region="<your_region>"))
pyramid.dt.to_zarr(store, mode="w", encoding=pyramid.encoding, zarr_format=3)

# Write to Icechunk
import icechunk

storage = icechunk.s3_storage(bucket="<your_bucket>", prefix="<your_prefix>", from_env=True)
repo = icechunk.Repository.create(storage)
session = repo.writable_session("main")
pyramid.dt.to_zarr(session.store, mode="w", encoding=pyramid.encoding, consolidated=False)
session.commit("write pyramid")

Contributing

Clone the repo and install with the test dependency group:

git clone https://github.com/carbonplan/topozarr
cd topozarr
uv sync --group test

Run tests:

uv run pytest -n auto

Run conformance tests against the GeoZarr spec (requires the conformance group):

uv sync --group conformance
uv run pytest -n auto -m conformance

Lint and format:

uv run pre-commit run --all-files

To regenerate the demo datasets in S3 (requires AWS credentials), install the demo extra and run the build script:

uv sync --extra demo
uv run python scripts/build_demo_data.py --help

License

[!IMPORTANT] This code is licensed under the MIT License - see the LICENSE file for details.

About Us

CarbonPlan is a nonprofit organization that uses data and science for climate action. We aim to improve the transparency and scientific integrity of climate solutions through open data and tools. Find out more at carbonplan.org or get in touch by opening an issue or sending us an email