Scratch work for creating STAC items for Sentinel-2 L2A 120m Mosaic (https://registry.opendata.aws/sentinel-s2-l2a-mosaic-120/)

sentinel-2-l2a-mosaic-120-2019-12-7-01K.json

{
    "type": "Feature",
    "stac_version": "1.0.0-beta.2",
    "id": "sentinel-2-l2a-mosaic-120-2019-12-7-01K",
    "properties": {
        "start_datetime": "2019-12-07T00:00:00",
        "end_datetime": "2019-12-17T00:00:00",
        "platform": "sentinel-2",
        "gsd": 120,
        "datetime": null
    },
    "geometry": null,
    "links": [],
    "assets": {
        "B02": {
            "href": "s3://sentinel-s2-l2a-mosaic-120/2019/12/7/01K/B02.TIF",
            "type": "image/tiff; application=geotiff; profile=cloud-optimized",
            "title": "B02",
            "eo:bands": [
                {
                    "name": "B02",
                    "common_name": "blue",
                    "center_wavelength": 0.4966,
                    "full_width_half_max": 0.098
                }
            ]
        },
        "B03": {
            "href": "s3://sentinel-s2-l2a-mosaic-120/2019/12/7/01K/B03.TIF",
            "type": "image/tiff; application=geotiff; profile=cloud-optimized",
            "title": "B03",
            "eo:bands": [
                {
                    "name": "B03",
                    "common_name": "green",
                    "center_wavelength": 0.56,
                    "full_width_half_max": 0.045
                }
            ]
        },
        "B04": {
            "href": "s3://sentinel-s2-l2a-mosaic-120/2019/12/7/01K/B04.TIF",
            "type": "image/tiff; application=geotiff; profile=cloud-optimized",
            "title": "B04",
            "eo:bands": [
                {
                    "name": "B04",
                    "common_name": "red",
                    "center_wavelength": 0.6645,
                    "full_width_half_max": 0.038
                }
            ]
        },
        "B08": {
            "href": "s3://sentinel-s2-l2a-mosaic-120/2019/12/7/01K/B08.TIF",
            "type": "image/tiff; application=geotiff; profile=cloud-optimized",
            "title": "B08",
            "eo:bands": [
                {
                    "name": "B08",
                    "common_name": "nir",
                    "center_wavelength": 0.8351,
                    "full_width_half_max": 0.145
                }
            ]
        },
        "B11": {
            "href": "s3://sentinel-s2-l2a-mosaic-120/2019/12/7/01K/B11.TIF",
            "type": "image/tiff; application=geotiff; profile=cloud-optimized",
            "title": "B11",
            "eo:bands": [
                {
                    "name": "B11",
                    "common_name": "swir16",
                    "center_wavelength": 1.6137,
                    "full_width_half_max": 0.143
                }
            ]
        },
        "B12": {
            "href": "s3://sentinel-s2-l2a-mosaic-120/2019/12/7/01K/B12.TIF",
            "type": "image/tiff; application=geotiff; profile=cloud-optimized",
            "title": "B12",
            "eo:bands": [
                {
                    "name": "B12",
                    "common_name": "swir22",
                    "center_wavelength": 2.22024,
                    "full_width_half_max": 0.242
                }
            ]
        }
    },
    "stac_extensions": [
        "eo"
    ],
    "collection": "sentinel-2-l2a-mosaic-120"
}

sentinel-2-l2a-mosaic-120-stac.py

from datetime import datetime, timedelta

import boto3
from pystac import Asset, Extensions, Item, MediaType
from pystac.extensions.eo import Band

s3 = boto3.client('s3')
BUCKET = 'sentinel-s2-l2a-mosaic-120'

BANDS = [{
    "name": "B02",
    "common_name": "blue",
    "center_wavelength": 0.4966,
    "full_width_half_max": 0.098
}, {
    "name": "B03",
    "common_name": "green",
    "center_wavelength": 0.56,
    "full_width_half_max": 0.045
}, {
    "name": "B04",
    "common_name": "red",
    "center_wavelength": 0.6645,
    "full_width_half_max": 0.038
}, {
    "name": "B08",
    "common_name": "nir",
    "center_wavelength": 0.8351,
    "full_width_half_max": 0.145
}, {
    "name": "B11",
    "common_name": "swir16",
    "center_wavelength": 1.6137,
    "full_width_half_max": 0.143
}, {
    "name": "B12",
    "common_name": "swir22",
    "center_wavelength": 2.22024,
    "full_width_half_max": 0.242
}]


def main():
    date_prefixes = list(get_date_prefixes())
    tiles = list(get_tiles(date_prefixes[0]))

    for date_prefix in date_prefixes:
        for tile in tiles:
            # date_prefix and tile have an extra ending /
            # e.g. '2019/12/7/' and '01K/'
            construct_stac(date_prefix.rstrip('/'), tile.rstrip('/'))


def construct_stac(date_prefix, tile):
    year, month, day = date_prefix.split('/')
    year, month, day = int(year), int(month), int(day)

    properties = get_properties(year, month, day)
    geometry = get_geometry(tile)
    bbox = get_bbox(geometry)

    collection_id = 'sentinel-2-l2a-mosaic-120'
    item_id = f'{collection_id}-{year}-{month}-{day}-{tile}'

    item = Item(id=item_id,
                geometry=geometry,
                bbox=bbox,
                datetime=None,
                properties=properties,
                collection=collection_id,
                stac_extensions=[Extensions.EO])

    for band_info in BANDS:
        asset = create_asset(date_prefix, tile, band_info)
        item.add_asset(asset.title, asset)

        eo_band = Band.create(**band_info)
        item.ext.eo.set_bands([eo_band], asset)

    return item


def get_properties(year, month, day):
    start_datetime = datetime(year, month, day)
    end_datetime = start_datetime + timedelta(days=10)

    return {
        'start_datetime': start_datetime,
        'end_datetime': end_datetime,
        'platform': 'sentinel-2',
        'gsd': 120,
    }


def create_asset(date_prefix, tile, band_info):
    band_name = band_info['name']
    href = f's3://{BUCKET}/{date_prefix}/{tile}/{band_name}.TIF'
    return Asset(href=href, title=band_name, media_type=MediaType.COG)


def get_geometry(tile):
    # TODO: load tile geometry from shapefile
    return None


def get_bbox(geometry):
    # TODO: get bounding box from geometry
    return None


def get_date_prefixes():
    for year in [2019]:
        for month in range(1, 13):
            prefix = f'{year}/{month}/'
            resp = s3.list_objects_v2(Bucket=BUCKET,
                                      Delimiter='/',
                                      Prefix=prefix)
            for common_prefix in resp['CommonPrefixes']:
                yield common_prefix['Prefix']


def get_tiles(prefix):
    # Note this assumes all tiles exist for all dates
    resp = s3.list_objects_v2(Bucket=BUCKET, Delimiter='/', Prefix=prefix)
    for common_prefix in resp['CommonPrefixes']:
        yield common_prefix['Prefix'][len(prefix):]


if __name__ == '__main__':
    main()