AWS Open-Data S3 — catalog explorer¶

No network needed: list the registered datasets, inspect a dataset's metadata, resolve variables, and dry-run the key planner. Everything here reads from the bundled S3 catalog — nothing is downloaded.

Setup¶

EarthLens is the unified entry point; Catalog is the S3 catalog loader, and the S3 backend exposes the list of registered datasets directly.

In [ ]:

from earthlens import EarthLens
from earthlens.s3 import Catalog

from earthlens import EarthLens from earthlens.s3 import Catalog

Registered datasets¶

S3.datasets() returns the keys of every AWS Open-Data dataset the backend knows about — the names you can pass as dataset=.

In [ ]:

S3.datasets()

S3.datasets()

Inspect a dataset¶

Load the catalog and resolve the era5 row, then print its bucket, on-disk format, and CRS alongside the variables it exposes.

In [ ]:

cat = Catalog()
era5 = cat.resolve('era5')
print('bucket:', era5.bucket, '| format:', era5.format, '| crs:', era5.crs)
print('default vars:', era5.default_variables)
sorted(era5.variables)

cat = Catalog() era5 = cat.resolve('era5') print('bucket:', era5.bucket, '| format:', era5.format, '| crs:', era5.crs) print('default vars:', era5.default_variables) sorted(era5.variables)

Resolve variables by any name¶

Variables resolve by friendly name, alias, or raw native token — all to one row, so you can request data with whichever spelling you have on hand.

In [ ]:

for key in ['t2m', '2m_temperature', '128_167_2t']:
    print(key, '->', era5.resolve_variable(key).native)

for key in ['t2m', '2m_temperature', '128_167_2t']: print(key, '->', era5.resolve_variable(key).native)

Dry-run the key planner¶

Build a request the normal way and call _search() to plan the S3 keys it would fetch — without downloading anything. Here we ask for a one-day copernicus-dem window over a small AOI.

In [ ]:

src = EarthLens(
    data_source="amazon-s3",
    start='2021-01-01',
    end='2021-01-01',
    aoi=[6.4, 0.4, 6.6, 0.6],
    dataset='copernicus-dem',
)

src = EarthLens( data_source="amazon-s3", start='2021-01-01', end='2021-01-01', aoi=[6.4, 0.4, 6.6, 0.6], dataset='copernicus-dem', )

Iterate the planned hrefs to see exactly which objects the request resolves to.

In [ ]:

for p in src._search():
    print(p.href)

for p in src._search(): print(p.href)