STAC

This package is an experimental implementation of the SpatioTemporal Asset Catalogs (STAC) client in Julia.

Opening an issue to notify about a missing feature is not helpful for the momement. However, if somebody is interested to make a pull request to implement a missing feature, an issue is a good way to discuss its implementation.

Installation

You need Julia (version 1.6 or later). Inside a Julia terminal, you can download and install STAC issuing these commands:

using Pkg
Pkg.add("STAC")

Example

Accessing a catalog and sub-catalogs are indexed with their identitiers. To find all subcatalog identifiers, one can simply display the catalog structure in a julia session.

using STAC
url = "https://raw.githubusercontent.com/sat-utils/sat-stac/master/test/catalog/catalog.json"
catalog = STAC.Catalog(url)
subcat = catalog["stac-catalog-eo"]
subcat1 = subcat["landsat-8-l1"]
@show subcat1

item = subcat1.items["LC08_L1TP_152038_20200611_20200611_01_RT"]
@show href(item.assets["B4"])

Searching by date range and bounding box:

using STAC, Dates
collections = ["landsat-8-c2-l2"]
time_range = (DateTime(2018,01,01), DateTime(2018,01,02)) # start and end time
lon_range = (2.51357303225, 6.15665815596)  # west, east
lat_range = (49.5294835476, 51.4750237087)  # south, north

catalog = STAC.Catalog("https://planetarycomputer.microsoft.com/api/stac/v1")

search_results = collect(search(catalog, collections, lon_range, lat_range, time_range))

NASA EarthData

Retrieve a list of OPeNDAP URLs from the NASA Common Metadata Repository (CMR) of the collection C1996881146-POCLOUD. A token is obtained from https://urs.earthdata.nasa.gov/home (after registration and login) and clicking on Generate Token:

using STAC, URIs, Dates

token = "put_your_user_token_here"
timerange = (DateTime(2019,1,1),DateTime(2019,12,31))
collection_concept_id = "C1996881146-POCLOUD"
baseurl = "https://cmr.earthdata.nasa.gov/search/granules.stac"

url = string(URI(URI(baseurl), query = Dict(
    "collection_concept_id" => collection_concept_id,
    "temporal" => join(string.(timerange),','),
    "pageSize" => 1000, # default is 100
    "token" => token)))

collection = STAC.FeatureCollection(url)
opendap_url = [href(item.assets["opendap"]) for item in collection]

@show length(opendap_url)
# output 365, one URL per day

To load the dataset, the NetCDF library need to be made aware of your EarthData username and password as explained here.