feat: add progress bars for ETL

[jsstevenson]

I don't think we have tqdm as an issue here, but this does it.

In doing so, refactors GFF processing to use gffpandas instead of gffutils (in order to be able to get a total number for tqdm). As a happy side effect, this significantly cuts down Ensembl and (I think) NCBI loading time (using Pandas, unfortunately -- with a little effort we could probably spin up something better, maybe next Tech the Halls). It did entail a fair amount of internal refactoring for those sources.

[jsstevenson]

I'm converting this to draft. My brain on this issue has gone like this

1. We should drop GFF Utils for gffpandas
2. We should just use a Polars-based GFF reader instead
3. Wait, GFFs are just tab-separated files, we probably don't need a special reader if all we're doing is looping through them from the top

We should handle the GFF reader issue and return to this afterwards

[github-actions]

This PR is stale because it has been open 7 day(s) with no activity. Please review this PR.

[korikuzma]

@jsstevenson do you intend to come back to this at some point?