feature/improved prefix handling

[abeforgit]

Result of a rabbithole that ended nowhere while debugging #261

• Collect seen prefixes on parse
• Allow users to specify extra prefixes ad-hoc

[nvdk]

Result of a rabbithole that ended nowhere while debugging #262

I think you might be linking to the wrong issue/PR here :). I'm not sure what this solves or offers other than a warning when overwriting prefixes with a different value (which may be valuable in itself).

[abeforgit]

Result of a rabbithole that ended nowhere while debugging #262

I think you might be linking to the wrong issue/PR here :). I'm not sure what this solves or offers other than a warning when overwriting prefixes with a different value (which may be valuable in itself).

Ah yes I meant 261, edited.
This doesnt really solve anything, but I made some wrong assumptions yesterday that the problems we were having were related to the way prefixes are handled. This was an area with a mental "todo" already since it's been an awkward implementation from the start.

Basically the "concise term" logic allowed users of the datastore to specify terms in a natural way such as "prefix:predicate", but the prefixes it understood were not related to the ones defined in the document. This fixes that and also allows to specify extra prefixes should that ever be necessary.

So this is more of a "well im here anyway, might as well clean this up" pr

[nvdk]

didn't test the code specifically but didn't encounter any issues and the code looks sound