Problem
Atm we only use submission_id internally but does not provide a way for users to do operations like patching metadata, or querying which file was added by which submission because once a file is ingested we don't have a hand the properties of the ingested file.
Ideally we should have some sort of "object_id" which is the identifier of the file object within the system and that the user can use for doing operations on it.
We would need some of the following:
- A unique identifier for each ingest request such that users can map back their file ingestions.
- Persistent request status or failure details so that users can retrieve the result of a completed request.
- Provenance showing which requests created or later modified a document.
- Consistent support for archives containing multiple files, e.g., a response request would return all of object-ids/submission-ids from the files within the archive.
Some thoughts about what we discussed @lewismc. If we agree, we can break this up into separate tasks and build support for this.
Problem
Atm we only use
submission_idinternally but does not provide a way for users to do operations like patching metadata, or querying which file was added by which submission because once a file is ingested we don't have a hand the properties of the ingested file.Ideally we should have some sort of "object_id" which is the identifier of the file object within the system and that the user can use for doing operations on it.
We would need some of the following:
Some thoughts about what we discussed @lewismc. If we agree, we can break this up into separate tasks and build support for this.