Set schema_URI in 1.2?

[jlaura]

I am testing out 1.2 right now with the data cube extension. I am not seeing a way to use the v1.0.0 implementation of the data cube extension.

If I do something like this to create an item:

item = pystac.Item(id=obj.productid, 
                   geometry=geometry, 
                   bbox=obj.bbox,  
                   datetime=dt,
                   stac_extensions=extensions,
                   href=os.path.join(obj.href,f'{obj.productid}.json'),
                   collection=collection,
                   properties=properties)

and stac_extensions includes: 'https://stac-extensions.github.io/datacube/v1.0.0/schema.json', I get a validation error for the wrong schema version:

pystac.errors.ExtensionNotImplemented: Could not find extension schema URI https://stac-extensions.github.io/datacube/v2.0.0/schema.json in object.`

Note the v2.0.0, but I am explicitly specifying v1.0.0. in my extensions. If I use DataCubeExtension.ext(item, add_if_missing=True), v1.0.0 is retained and v2.0.0 is added to the list.

Is this user error? I am not seeing a way to set the schema URI here.

[TomAugspurger]

What happens if you monkeypatch

pystac/pystac/extensions/datacube.py

Line 19 in 6e6e987

SCHEMA_URI = "https://stac-extensions.github.io/datacube/v2.0.0/schema.json"

by setting

pystac.extensions.datacube.SCHEMA_URI  = "https://stac-extensions.github.io/datacube/v1.0.0/schema.json"

before you run validate?

I don't know if I would recommend doing that, in case other places in the library are relying on something being compatible with 2.0, but it might be a stopgap.

I'm not sure if pystac has thought about versioning the actual extension objects (e.g. a V1DataCube, V2DataCube, etc.). That sounds like a lot of work.

[jlaura]

@TomAugspurger I'll give monkeypatching in the schema uri a go.

Since the extensions are going to version, doesn't pystac need to be able to support different extension versions if it is going to provide schema validation? I understand the amount of work this is. What are the alternatives?

[TomAugspurger]

I think it’s that, or you pin to an older version, or you update your STAC objects. I’m not sure how the extensions work, but subclassing & setting https://github.com/stac-utils/pystac/blob/6e6e987dcb0e711adcd89432bd495199923cf759/pystac/extensions/datacube.py#L653 <https://github.com/stac-utils/pystac/blob/6e6e987dcb0e711adcd89432bd495199923cf759/pystac/extensions/datacube.py#L653> might be a bit less invasive and do the trick.

…

On Dec 11, 2021, at 7:20 AM, jlaura ***@***.***> wrote: @TomAugspurger <https://github.com/TomAugspurger> I'll give monkeypatching in the schema uri a go. Since the extensions are going to version, doesn't pystac need to be able to support different extension versions if it is going to provide schema validation? I understand the amount of work this is. What are the alternatives? — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#668 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAKAOIXYAJPATPNLPPSZYALUQNF2JANCNFSM5IJQZZXQ>.

[duckontheweb]

Since the extensions are going to version, doesn't pystac need to be able to support different extension versions if it is going to provide schema validation? I understand the amount of work this is. What are the alternatives?

Thanks for raising this @jlaura. We have an open issue (#448) to discuss how we want to handle versioning of STAC Extensions within PySTAC and it would be great to have you and @TomAugspurger weigh in on that if you have an opinion. For now, we handle extension versions in the same way as the stac_version: it is automatically updated to the latest supported version and we attempt to apply any migrations from previous versions when we deserialize the object.

[gadomski]

Closing as wontfix ... #1231 should help use correctly load older and newer versions, and #448 has been closed as complete. If folks need to use older versions of a given extension, they'll need to install an older version of PySTAC.