Narrative PDFs are being referenced incorrectly in the DMPHub JSON record

[briri]

For an example, see DMP#doi.org/10.48321/D1BF4BE9E7 the last entry is the most relevant.

In reality, the PostNarrative Lambda should do the following:

• replace the existing entry
• delete the old narrative PDF from S3 (unless it is referenced by a prior version)
• update the GetDmp Lambda so that it only ever returns the last entry (in the event that there are multiples again)

[briri]

Updated the landing page to always take the last entry in PR #157 as a temporary fix which makes this work less of a priority but something we should still address