[feature] The documentRef GraphQL field is populated by the collectors

[nchelluri]

Is your feature request related to a problem? Please describe.
After closing #1833 we now have a DocumentRef field in the SourceInformation type that corresponds to the documentRef field in GraphQL verb nodes. The next (and last) step is to populate this field. As part of this work we also want to undo the overloading of the GraphQL origin field that was performed in the recently created --use-blob-url CLI flag for the File collector (done in #1811) and leave origin as it was prior that work and store the blob key only (and always) in the new documentRef field. I am also suggesting that we switch to using the term "blob key" as that seems to match the terminology used in the method signature here: https://pkg.go.dev/gocloud.dev/blob?utm_source=godoc#Bucket.NewWriter

This work is a follow-up to #1844. It is aimed to ease reingestion of documents.

Describe the solution you'd like
The collectors should populate the SourceInformation.DocumentRef field with the blob key (an example value for the field is sha256:4c47f63847cd2b73d77f9008bb8bfa359e6201205004a2c26048a8711fd1c746).