Maxar 'Browse Image' Documentation

[Jack-Hayes]

For now, Maxar ordering relies on clunky G-EGD platform. Providing users with a 'browse' for an input Maxar GeoDataFrame returned from coincident.search will be useful. You can call a GET request to do so if you have the image id via f"https://api.maxar.com/browse-archive/v1/browse/show?image_id={image_id}". The browse TIFs are also accessible. Their API endpoint converts these tifs to JPGs w/o georeferencing on their server... But you can just access the tifs directly as seen below:

%matplotlib inline
import xarray as xr
import rasterio
import numpy as np
import os 

env = rasterio.Env(
    GDAL_DISABLE_READDIR_ON_OPEN="EMPTY_DIR",
    GDAL_HTTP_HEADERS=f'MAXAR-API-KEY:{os.environ["MAXAR_API_KEY"]}'
)

href = 'https://api.maxar.com/discovery/v1/catalogs/imagery/collections/wv01/items/102001008EC5AC00/assets/collections/dg-archive/assets/browse/102001008EC5AC00.browse.tif'

with env:
    da = xr.open_dataarray(href, engine='rasterio', 
                           mask_and_scale=False, # otherwise uint8 -> float32!
                           backend_kwargs=dict(open_kwargs={'overview_level':1})
    )
da

We can also pull inspiration ASP plot and loop in Ben Purinton down the line for more-detailed stereo report on any user-input stereo data hopefully

[scottyhq]

@Jack-Hayes there is already a function for this in the codebase here

coincident/src/coincident/io/xarray.py

Line 63 in 67a837c

def open_maxar_browse(item: pystac.Item, overview_level: int = 0) -> xr.DataArray:

as seen in this notebook.

Let's try to avoid links to private content when possible.

...But looks like the function is from the docs! The API page could certainly be made easier to navigate, and it might be worth adding a series of example jupyter notebooks per-dataset since that is often more interesting to people than API docs :) https://coincident.readthedocs.io/en/latest/api.html