Overture — multi-theme download (live)¶

A single request can pull several themes / types at once; one file is written per feature type. Here we grab places and building footprints for the same block, targeting the release Overture publishes right now.

Setup¶

Pull in the pieces we need: the EarthLens entry point to run the download, releases to see which Overture release the download will target, and geopandas to inspect the written GeoParquet files.

In [ ]:

import geopandas as gpd

from earthlens.core import EarthLens
from earthlens.overture import releases

import geopandas as gpd from earthlens.core import EarthLens from earthlens.overture import releases

Area of interest¶

Define a small Times Square block as a [south, north] / [west, east] bounding box, and the directory the output files will be written to.

In [ ]:

LAT_LIM = [40.757, 40.759]  # [south, north]
LON_LIM = [-73.987, -73.984]  # [west, east] — a Times Square block
OUT = "_overture_out"

LAT_LIM = [40.757, 40.759] # [south, north] LON_LIM = [-73.987, -73.984] # [west, east] — a Times Square block OUT = "_overture_out"

Which release will this read?¶

Overture publishes a dated release roughly monthly and keeps only the newest one or two on S3, pruning the rest — so a release id is a moving target, and an id that worked last quarter may no longer exist. Leaving release=None (the default) targets whatever is published now, which is what we want here.

You can pin a release when you need a byte-reproducible download, but the pin holds only while that release is still on S3; once it is pruned the fetch fails with No files found that match the pattern. Ask Overture what it is serving today:

In [ ]:

release = releases.latest_release()
release

release = releases.latest_release() release

Download both themes¶

Request places and buildings in one call. Overture writes one GeoParquet file per feature type, so download() returns a list of paths — one per type.

In [ ]:

engine = EarthLens(
    data_source="overture",
    variables={"places": [], "buildings": []},
    aoi=[LON_LIM[0], LAT_LIM[0], LON_LIM[1], LAT_LIM[1]],
    path=OUT,
)
paths = engine.download()
[p.name for p in paths]

engine = EarthLens( data_source="overture", variables={"places": [], "buildings": []}, aoi=[LON_LIM[0], LAT_LIM[0], LON_LIM[1], LAT_LIM[1]], path=OUT, ) paths = engine.download() [p.name for p in paths]

Inspect the results¶

Read each written file back with geopandas and report its feature count and the set of per-row license ids (buildings are OSM-derived ODbL-1.0; places carry permissive licenses).

In [ ]:

for p in paths:
    gdf = gpd.read_parquet(p)
    print(f'{p.name}: {len(gdf)} features, licenses={sorted(gdf.license_id.unique())}')

for p in paths: gdf = gpd.read_parquet(p) print(f'{p.name}: {len(gdf)} features, licenses={sorted(gdf.license_id.unique())}')