Example notebook¶

This notebook demonstrates some of the more interesting features of stacrs.

stac-geoparquet¶

stac-geoparquet is a relatively new specification that describes how to store STAC items in a geoparquet. Use stacrs to read and write stac-geoparquet.

In [2]:

import os.path

import stacrs
import humanize

items = stacrs.search(
    "https://landsatlook.usgs.gov/stac-server",
    collections="landsat-c2l2-sr",
    intersects={"type": "Point", "coordinates": [-105.119, 40.173]},
    sortby="-properties.datetime",
    max_items=1000,
)
stacrs.write("items.json", items)  # GeoJSON feature collection
stacrs.write("items.ndjson", items)  # newline-delimited JSON
stacrs.write("items.parquet", items)  # stac-geoparquet
stacrs.write(  # compressed stac-geoparquet
    "items-compressed.parquet", items, format="parquet[snappy]"
)

for file_name in (
    "items.json",
    "items.ndjson",
    "items.parquet",
    "items-compressed.parquet",
):
    print(f"{file_name}: {humanize.naturalsize(os.path.getsize(file_name))}")

items = stacrs.read("items-compressed.parquet")
assert len(items["features"]) == 1000

import os.path import stacrs import humanize items = stacrs.search( "https://landsatlook.usgs.gov/stac-server", collections="landsat-c2l2-sr", intersects={"type": "Point", "coordinates": [-105.119, 40.173]}, sortby="-properties.datetime", max_items=1000, ) stacrs.write("items.json", items) # GeoJSON feature collection stacrs.write("items.ndjson", items) # newline-delimited JSON stacrs.write("items.parquet", items) # stac-geoparquet stacrs.write( # compressed stac-geoparquet "items-compressed.parquet", items, format="parquet[snappy]" ) for file_name in ( "items.json", "items.ndjson", "items.parquet", "items-compressed.parquet", ): print(f"{file_name}: {humanize.naturalsize(os.path.getsize(file_name))}") items = stacrs.read("items-compressed.parquet") assert len(items["features"]) == 1000

items.json: 21.4 MB
items.ndjson: 21.4 MB
items.parquet: 488.2 kB
items-compressed.parquet: 178.8 kB

Search¶

As seen above, you can search STAC APIs. Using DuckDB, you can search stac-geoparquet files as well, both local and remote.

In [3]:

import contextily
import stacrs
from geopandas import GeoDataFrame

url = "https://github.com/gadomski/2024-09-FOSS4G-NA-stac-rs/raw/refs/heads/main/benches/1000-sentinel-2-items-stac-rs.parquet"
items = stacrs.search(
    url, intersects={"type": "Point", "coordinates": [-105.1019, 40.1672]}
)
data_frame = GeoDataFrame.from_features(items, crs="EPSG:4326")
axis = data_frame.plot(facecolor="none")
contextily.add_basemap(axis, crs=data_frame.crs.to_string())

import contextily import stacrs from geopandas import GeoDataFrame url = "https://github.com/gadomski/2024-09-FOSS4G-NA-stac-rs/raw/refs/heads/main/benches/1000-sentinel-2-items-stac-rs.parquet" items = stacrs.search( url, intersects={"type": "Point", "coordinates": [-105.1019, 40.1672]} ) data_frame = GeoDataFrame.from_features(items, crs="EPSG:4326") axis = data_frame.plot(facecolor="none") contextily.add_basemap(axis, crs=data_frame.crs.to_string())

Search to¶

If you want to save the results of a STAC API search directly to a file, use search_to. This will generally be more performant than using search then write.

In [4]:

stacrs.search_to("items-compressed.parquet",
    "https://landsatlook.usgs.gov/stac-server",
    collections="landsat-c2l2-sr",
    intersects={"type": "Point", "coordinates": [-105.119, 40.173]},
    sortby="-properties.datetime",
    max_items=1000,
    format="parquet[snappy]",
)

stacrs.search_to("items-compressed.parquet", "https://landsatlook.usgs.gov/stac-server", collections="landsat-c2l2-sr", intersects={"type": "Point", "coordinates": [-105.119, 40.173]}, sortby="-properties.datetime", max_items=1000, format="parquet[snappy]", )

Out[4]:

1000