Improve first-time observation load for large datasets

[JWCook]

For a large number of observations (say, >10K records), a number of problems can come up during the first-time data load. Some general tasks, roughly in order of complexity:

• Ensure app is responsive before initial observation load finishes
• Make incremental batches of observations available for tagging as they become available
• Add option to only download n most recent observations (or from a given start date)
  • May break some assumptions elsewhere that all user observations are in local db
• Fix progress bar (somewhat complex issue with parallel batch loader reporting to main UI thread)
• Bootstrap observation data load with GBIF research-grade export and/or iNat CSV export
  • Show a notification when attempting to download >10K records and recommend starting with a data export

User with >100K records to stress test with: https://www.inaturalist.org/observations?user_id=martin_der_prinz
ref: https://forum.inaturalist.org/t/naturtag-organize-your-photo-collection-with-inat-metadata/33959/22

Will split this into separate issues after more research/testing.