STACError: HREF: <url> does not resolve to a STAC object

[jaiindimple]

Hello there,
I'm trying to load the Capella static STAC catalog using pystac with the following code:

from pystac import Catalog

root_url = "https://capella-open-data.s3.us-west-2.amazonaws.com/stac/catalog.json"
catalog = Catalog.from_file(root_url)
items = list(catalog.get_all_items())
print(f"Total capella scenes: {len(items)}")

However, I'm facing this error during execution:

18 frames
UnicodeEncodeError: 'ascii' codec can't encode character '\xe9' in position 60: ordinal not in range(128)

The above exception was the direct cause of the following exception:

STACError                                 Traceback (most recent call last)
/usr/local/lib/python3.11/dist-packages/pystac/link.py in resolve_stac_object(self, root)
    332                     obj = stac_io.read_stac_object(target_href, root=root)
    333                 except Exception as e:
--> 334                     raise STACError(
    335                         f"HREF: '{target_href}' does not resolve to a STAC object"
    336                     ) from e

STACError: HREF: 'https://capella-open-data.s3.us-west-2.amazonaws.com/stac/capella-open-data-by-capital/capella-open-data-malé/collection.json' does not resolve to a STAC object

The issue seems to be triggered when resolving links with non-ASCII characters (like é). What could be the right way to handle this in pystac.

[jsignell]

Thanks for opening this! I think you are totally right about where the issue is coming from. I can reproduce with just this:

import pystac

collection = pystac.Collection.from_file("https://capella-open-data.s3.us-west-2.amazonaws.com/stac/capella-open-data-by-capital/capella-open-data-malé/collection.json")

It looks like specifically the issue is that we need to do some quoting before passing the url to urllib.request.urlopen- based on https://stackoverflow.com/questions/4389572/how-to-fetch-a-non-ascii-url-with-urlopen. I am taking a look now.

[jsignell]

I also found that this handled by default in requests and urllib3, but if we want to avoid adding extra places within in the code where we rely on optional dependencies, then we need to have some handling within pystac.

[jsignell]

As a workaround you can use a custom stac_io class to handle this or just use RetryStacIO since that uses urllib3 anyways

from pystac import Collection
from pystac.stac_io import RetryStacIO

stac_io = RetryStacIO()
collection = Collection.from_file(
    "https://capella-open-data.s3.us-west-2.amazonaws.com/stac/capella-open-data-by-capital/capella-open-data-malé/collection.json",
    stac_io=stac_io,
)

[jaiindimple]

Thank you for the quick response 🙏🏼