extracting frame from existing JSON-LD document?

Howdy!

I start helping little more with LevelGraph-JSONLD[1] and must right 
away admit that still don't fully understand internals of JSON-LD 
algorithms, including *framing*

At this moment on *put* we use 'Deserialize JSON-LD to RDF' algorithm 
and then store triples, which other extensions can also access (r/w) 
later on, for example LevelGraph-N3[2]

Let's imagine JSON-LD document with many nodes (no blank nodes) and 
various levels of nesting (embedding). I wonder if we could somehow 
extract frame from it which would later enable us to recreate it in the 
same form?

I might not want to save a copy of original document since after 
modifying triples it would become stale. But I would still like to 
recreate it in the same form but with latest data.

I could prepare practical example, but maybe someone can already answer 
this question. Maybe just to clarify it further:

Having only mentioned JSON-LD document

* can one extract from it frame which framing algorithm used to form it?
* do documents exist which have no particular frame that would form them?
* do documents exist where same document could come out as a result of 
applying different frames?

Thanks for any tips!

[1] https://github.com/mcollina/levelgraph-jsonld
[2] https://github.com/mcollina/levelgraph-n3