Reconsider import as too many RIS versions fail

[LukasWallrich]

MEDLINE RIS files are not imported correctly by synthesisr - see example below. @TNRiley will add a note to that effect - but given that this occurs repeatedly, we might want to think about a more stable solution. Given that Zotero is OS, maybe we can port their import code into an R package if that does not exist yet?

medline_prob.ris.zip

[TNRiley]

Dug into this further. The issue is with OVID exports, which can actually search across Medline, Embase, CAB, CINAHL, and others. It appears that the synthesis "best_guess" option is not identifying these .ris as OVID, and therefore not mapping the fields correctly. May be easier to see if Eliza or Martin could lend a hand?

[LukasWallrich]

The synthesisr function already lives in our package (as we had to fix a bug in it) - so we can improve how best_guess works. Can you already identify something that would indicate OVID files accurately?

[LukasWallrich]

The issue is likely in the detect_lookup function - but I don't have time right now to dig into it