pySDP

Native Python interface for the RMBL Spatial Data Platform — curated, high-resolution geospatial datasets covering Western Colorado (USA) around Rocky Mountain Biological Laboratory.

Lazy cloud access. Standard scientific-Python types. Feature-parity port of the rSDP R package.

• 🌐 No downloads needed — open 1 GB COGs from S3 and extract point samples without a local copy. GDAL VSICURL handles the range reads.
• 🔄 Composable return types — xarray.Dataset, geopandas.GeoDataFrame, pandas.DataFrame. No custom classes to learn; everything plugs into the PyData ecosystem.
• 📅 Time-series aware — uniform pandas.DatetimeIndex across Daily / Monthly / Yearly products. ds.sel(time="2019"), .resample, .groupby("time.year") all work.
• 🚀 Dask-ready — lazy chunked reads via chunks="auto" (requires pysdp[dask]) let you scale to stacks that don't fit in memory.

📘 Full documentation: https://rmbl-sdp.github.io/pySDP/ 📖 Getting started • User guides • API reference

Install

pip install pysdp                 # core: catalog, raster, extraction, download
pip install "pysdp[dask]"         # + lazy chunked reads via Dask
pip install "pysdp[stac]"         # + pystac-client + odc-stac for STAC
pip install "pysdp[hub]"          # + dask-gateway for JupyterHub clusters
pip install "pysdp[all]"          # everything

Requires Python 3.11+. All core deps (rasterio, rioxarray, xarray, geopandas, pystac, scipy, ...) have wheels on PyPI for Linux, macOS (Intel + Apple Silicon), and Windows — no system GDAL install needed.

30-second tour

import pysdp
import geopandas as gpd

# Discover what's in the catalog
cat = pysdp.get_catalog(domains=["UG"], types=["Vegetation"])
cat[["CatalogID", "Product", "Resolution"]].head()

# Open a raster lazily — the 1 GB landcover COG stays on S3
landcover = pysdp.open_raster("R3D018")

# Sample at field sites (auto-reprojects to the raster's CRS)
sites = gpd.GeoDataFrame(
    {"site": ["Roaring Judy", "Gothic", "Galena Lake"]},
    geometry=gpd.points_from_xy(
        [-106.853186, -106.988934, -107.072569],
        [38.716995, 38.958446, 39.021644],
    ),
    crs="EPSG:4326",
)
samples = pysdp.extract_points(landcover, sites)
samples

See the user guides for deeper walkthroughs — catalog discovery, raster wrangling, field-site sampling, and making publication-quality maps.

Public API

get_catalog() and browse() accept include_deprecated=True to surface deprecated products (which carry a NewVersionID column pointing at the replacement) and with_issue_counts=True to attach the OpenIssues count per dataset.

Full signatures and examples: https://rmbl-sdp.github.io/pySDP/api/.

Coming from rSDP?

pySDP is a direct port with the same catalog, same vocabulary, and feature parity. The rSDP → pySDP mapping in SPEC §5 is the authoritative reference; the short version:

Status & roadmap

v0.6.0 (current): tracks rSDP v0.6. Catalog discovery, lazy raster access, point/polygon extraction, bulk download, weekly drone imagery, irregular time-series (get_dates()), visual catalog browser (browse()), dataset version-control (deprecation + NewVersionID), and the data-products issue tracker (report_issue() / known_issues()). Tests and CI cover Python 3.11 / 3.12 / 3.13 × Linux / macOS / Windows.

Upcoming (see ROADMAP.md for details):

• Phase 7 — JupyterHub + Dask Gateway integration (pysdp.hub), GDAL cloud-tuning defaults
• Phase 8a — Distributed extraction partition helpers + at-scale recipes
• Phase 8b — Polygon zonal stats + benchmark harness

Project documents

• SPEC.md — v0.1 design specification (API, dependencies, phased plan)
• ROADMAP.md — post-v0.1 integrations (Hub, distributed extraction)
• CHANGELOG.md — release history
• docs/contributing.md — development setup + PR workflow

License

MIT — see LICENSE.

Citation

pySDP is part of the infrastructure around the RMBL Spatial Data Platform. If it supports your research, please cite the SDP (citation guidance at https://www.rmbl.org/scientists/resources/spatial-data-platform/). A formal citation for pySDP itself will be added at the 0.1.0 stable release.

Acknowledgements

pySDP is developed by Ian Breckheimer and collaborators at Rocky Mountain Biological Laboratory. Built on the shoulders of giants: xarray, rioxarray, rasterio, geopandas, pystac, xvec, and odc-stac.