Workflow for Regular Metadata

[uxairibrar]

Closes #23

• Updated the function to extract only regular textual metadata (skipping geospatial and temporal fields) for manual entry.
• Linked new Publication records to a HarvestingEvent with user and timestamp details.
• Compiled metadata counts and sent an email notification to the initiating user with job details (including collection name and journal URL).
• Added testcase.

[nuest]

Just a little oversight, I think, to remove the temporal field extraction, and a little shift in approaching the solution: it is more important not to fail if geometries are missing, but if they are there, we surely would like them.

Therefore I am not sure how important the harvest_regular... as a seperate function really is.

Do you think you can just put it all in one function that creates sensible output if there are no geometries?

[nuest]

Everything is resolved from my side, please take a look at the failing unit test (it looks rather simple to fix?) and I'll merge right away if that is resolved.

[nuest]

Please update the deduplication to not query all identifiers from the database every time we insert a new record, but instead rely on the database index to be fast.