Flesh out narrative key logic

What needs to be done

Generate stable keys for CDA narrative tables and rows in an eICR wherever possible in order to correlate the same narrative sections across two different eICRs that come from the same set.
Currently in cda_narrative_keys.py.

Why it needs to be done

We need this for our core matching logic.

Acceptance Criteria

• Working with existing prototype code, determine the best method to generate a key for narrative sections. See cda_narrative_keys.py.

Technical Notes

• The prototype's existing fallback key logic will be fleshed out in a future ticket for scenarios when neither a stable key nor a narrative key exists for a given part of an eICR.
• Development should be done on the david/prototype-diff branch to ensure new code still properly integrates with all of the prototype code that hasn't been fleshed out yet. Once this feature is complete, then a new branch should be created to merge only the files relevant to this feature into main.