Use CMR metadata to get S3FS sessions and authenticate queries to CMR using top level APIs

[betolink]

This PR allows earthacces.search_data() and earthaccess.search_dataset() use the auth object to send bearer tokens to CMR, usually this is not required but in the case of restricted datasets or early release data CMR will only return results if we are in the access control lists for a given dataset.

with this PR earthaccess will be able to use the metadata from CMR to grab S3 credentials, perhaps we need to simplify this even more but for now this works.

import earthaccess

earthaccess.login()

results = earthaccess.search_data(
    short_name='SWOT_L2_HR_RIVERSP_1.0',
    cloud_hosted=True,
    bounding_box=(-10, 20, 10, 50),
    temporal=("1999-02", "2019-03"),
    count=10
)

# now earthaccess will use that the S3 credentials if they are present in the metadata if not it will grab the DAAC's
earrthaccess.open(results)

This code was failing because the SWOT mission has its own S3 credentials that are not from PODAAC.

another way of getting the authenticated sessions would be:

if len(results) > 0:
    print(f"Actual granules returned: {len(results)}")    
    fs_s3 = earthaccess.get_s3fs_session(results=results)

And then we could use the storage_options to create authenticated sessions, this is the part that can be simplified:

import fiona
from fiona.session import AWSSession

with fiona.Env(
    session=AWSSession(
        aws_access_key_id=fs_s3.storage_options["key"],
        aws_secret_access_key=fs_s3.storage_options["secret"],
        aws_session_token=fs_s3.storage_options["token"]
    )
):
    # some code  with gpd.read_file() etc

[github-actions]

👈 Launch a binder notebook on this branch for commit a85f531

I will automatically update this comment whenever this PR is modified

👈 Launch a binder notebook on this branch for commit ca6590f

👈 Launch a binder notebook on this branch for commit c4321c0