How to expose and query multiple files (e.g. an S3 folder) as a single Asset?

[marcofranke]

As far as I understand, in the Eclipse Dataspace Connector an Asset describes a data object, such as a document, an AAS (Asset Administration Shell), or a log message.
As a Data Provider, I can create such an Asset, which can then be discovered by a Consumer via catalog queries and eventually exchanged.
My question is: how does this concept work if I don’t want to expose a single file as an Asset, but rather an entire folder in an S3 bucket that contains multiple files?
I am particularly interested in understanding:

How can the Consumer query the current content (e.g. the list of files) within that folder?

When a data transfer (pull) is initiated, how can the Consumer know what exactly can be retrieved before starting the transmission?

Any hints or examples on how to model or implement this scenario (e.g., structured metadata, artifacts grouping, or dynamic asset contents) would be highly appreciated.

[arnoweiss]

1. Granting access to multiple files is documented here [1]
2. Metadata/Querying: Metadata is usually modelled as properties on a Dataset. This is configured via the properties array during the Asset definition on the Provider-side. Each Dataspace needs conventions on the semantics of such metadata properties [3]. If that's agreed, Consumers can use the QuerySpec object in conjunction with the CatalogRequest [2].

[1] https://github.com/eclipse-edc/Technology-Aws/tree/main/extensions/data-plane/data-plane-aws-s3#properties
[2] https://github.com/eclipse-tractusx/tractusx-edc/blob/main/docs/usage/management-api-walkthrough/04_catalog.md#catalog-request
[3] https://github.com/eclipse-tractusx/tractusx-edc/blob/main/docs/usage/management-api-walkthrough/01_assets.md#creating-an-asset