Load all catalog input data via dlt

In preparation for porting transformations to be done in a database via dbt, we can load input data into the database via dlt (from where it can, for now, be read back out and transformed via Pandas).

Inputs to include:

- Full source lists from catalog YAML files (assemblies/organisms/outbreaks)
- NCBI genomes
- SRA metadata? (Not used by either catalog)
- UCSC assemblies
- Gene model URLs
- Datacache directories
- Organism images