rgd_client - Resonant GeoDataClient

The rgd_client Python package is a well typed, easy to use, and extendable Python client for Resonant GeoData APIs.

Installation

To install the core client

pip install rgd-client

To use other core modules or plugins, install the corresponding client packages. For example, the imagery client plugin is installed with

pip install rgd-imagery-client

All the functions added via a plugin are namespaced under a name defined by that plugin. For the imagery client plugin, this is imagery, so all of these plugin's features are accessed through client.imagery.*. Examples of this are shown below.

Usage

Search and display results

import json
import matplotlib.pyplot as plt
import numpy as np

from rgd_client import create_rgd_client

def plot_geojson(gjs, *args, **kwargs):
    points = np.array(gjs['coordinates'])
    if points.ndim == 3:
        points = points[0]
    if points.ndim == 1:
        points = points.reshape((1, points.size, ))
    return plt.plot(points[:,0], points[:,1], *args, **kwargs)

client = create_rgd_client(username='username', password='password')
bbox = {
    "type":"Polygon",
    "coordinates":[
        [
            [-105.45091240368326,39.626245373878696],
            [-105.45091240368326,39.929904289147274],
            [-104.88775649170178,39.929904289147274],
            [-104.88775649170178,39.626245373878696],
            [-105.45091240368326,39.626245373878696]
        ]
    ]
}

q = client.rgd.search(query=json.dumps(bbox), predicate='intersects')

for s in q:
    print(s['subentry_name'])

plot_geojson(bbox, 'k--', label='Search Region')

for s in q:
    plot_geojson(s['footprint'], label=s['subentry_name'])

plt.legend()
plt.title(f'Count: {len(q)}')

Inspect raster

Preview thumbnails of the raster

import imageio
from io import BytesIO

raster = client.imagery.get_raster(q[0])
plot_geojson(bbox, 'k--')
plot_geojson(raster['outline'], 'r')
load_image = lambda imbytes: imageio.imread(BytesIO(imbytes))

count = len(raster['parent_raster']['image_set']['images'])
for i in range(count):
    thumb_bytes = client.imagery.download_raster_thumbnail(q[0], band=i)
    thumb = load_image(thumb_bytes)
    plt.subplot(1, count, i+1)
    plt.imshow(thumb)

plt.tight_layout()
plt.show()

Download Raster

Download the entire image set of the raster

import rasterio
from rasterio.plot import show

paths = client.imagery.download_raster(q[0])
rasters = [rasterio.open(im) for im in paths.images]
for i, src in enumerate(rasters):
    plt.subplot(1, len(rasters), i+1)
    ax = plt.gca()
    show(src, ax=ax)
plt.tight_layout()
plt.show()

STAC Item Support

The Python client has a search endpoint specifically for Raster data that returns each record in the search results as a STAC Item.

q = client.imagery.search_raster_stac(query=json.dumps(bbox), predicate='intersects')

print(q[0])  # view result as STAC Item

# Download using the search result
paths = client.imagery.download_raster(q[0])
print(paths)

We can also upload new data in the STAC Item format. Here we simply pass back the same STAC Item JSON which will not actually do anything because RGD recognizes that these files are already present with a Raster.

client.imagery.create_raster_stac(q[0])

Please note that the assets in the STAC Item must already be uploaded to a cloud storage provider with either s3:// or https:// URLs. Further, the images must have the data tag on each asset. e.g.:

{
    ... # other STAC Item fields
    'assets': {
        'image-15030': {
            'href': 'http://storage.googleapis.com/gcp-public-data-sentinel-2/tiles/17/S/MS/S2A_MSIL1C_20210302T161201_N0209_R140_T17SMS_20210302T200521.SAFE/GRANULE/L1C_T17SMS_A029738_20210302T161751/IMG_DATA/T17SMS_20210302T161201_B01.jp2',
            'title': 'GRANULE/L1C_T17SMS_A029738_20210302T161751/IMG_DATA/T17SMS_20210302T161201_B01.jp2',
            'eo:bands': [{'name': 'B1'}],
            'roles': ['data'],
        },
        'image-15041': {
            'href': 'http://storage.googleapis.com/gcp-public-data-sentinel-2/tiles/17/S/MS/S2A_MSIL1C_20210302T161201_N0209_R140_T17SMS_20210302T200521.SAFE/GRANULE/L1C_T17SMS_A029738_20210302T161751/IMG_DATA/T17SMS_20210302T161201_B02.jp2',
            'title': 'GRANULE/L1C_T17SMS_A029738_20210302T161751/IMG_DATA/T17SMS_20210302T161201_B02.jp2',
            'eo:bands': [{'name': 'B1'}],
            'roles': ['data'],
        },
        ...  # ancillary files can lack a role but we like to see `metadata` used.
        'ancillary-30687': {
            'href': 'http://storage.googleapis.com/gcp-public-data-sentinel-2/tiles/17/S/MS/S2A_MSIL1C_20210302T161201_N0209_R140_T17SMS_20210302T200521.SAFE/GRANULE/L1C_T17SMS_A029738_20210302T161751/QI_DATA/MSK_TECQUA_B03.gml',
            'title': 'GRANULE/L1C_T17SMS_A029738_20210302T161751/QI_DATA/MSK_TECQUA_B03.gml',
            'roles': ['metadata'],
        },
    }
}

Plugin Development

For instructions on how to develop a plugin for rgd_client, see PLUGINS.md.