Scans a directory for Cloud-Optimized GeoTIFFs and builds a
STAC catalog JSON file. Each COG becomes one item with
cd:variable and cd:period properties parsed from the filename.
The resulting catalog is compatible with cd_catalog().
Usage
cd_stac_catalog(
cog_dir,
output_path = "catalog.json",
catalog_id = "era5-land",
title = "ERA5-Land Climate Data",
description = NULL,
base_url = "https://stac-era5-land.s3.us-west-2.amazonaws.com"
)Arguments
- cog_dir
Character. Directory containing COG files (.tif).
- output_path
Character. Path to write the catalog JSON. Default
"catalog.json".- catalog_id
Character. STAC catalog ID. Default
"era5-land".- title
Character. Human-readable catalog title.
- description
Character. Optional catalog description.
- base_url
Character. Base URL where COGs will be served. Asset hrefs are built as
{base_url}/{filename}.
Details
Each item's data asset carries file:checksum (a sha256 multihash,
"1220" + hex digest) and file:size from the
STAC file extension v2.1.0,
taken from the local file, so build the catalog after the COGs are final.
Run provenance written into a COG by cd_cog_write() (CD_VERSION,
CD_SHA, CD_RUN_TIME, CD_RUN_ID tags) is copied into the item in the
STAC processing extension
v1.2.0: processing:version, processing:software ({"cd": "<sha>"}),
processing:datetime (the tag's string, unchanged) and, for a GitHub
Actions run, a processing-execution link to the run. An item from an
untagged COG carries none of them and does not declare the extension.
Because the run time is part of the bytes, a checksum identifies a build,
not values: rebuilding unchanged data gives a new checksum.
