Try Phenopacket curation again

[caufieldjh]

@pnrobinson commented:

It would be really interesting to try to curate a phenopacket from the medical literature using both PhenopacketLab and CurateGPT for say 20 cases and then compare the results with respect to speed and accuracy.

@justaddcoffee tried some case report -> Phenopacket curation here: #13

It may be time for another try with more recent LLMs and more constraints on generated IDs, or even an additional SPIRES refiner step to ground concepts.