feat!: acquire data with wags-tails, overhaul static file handling

[jsstevenson]

• use wags-tails to acquire lrg_refseqgene and mane_summary files from NCBI. Get rid of the DataDownload class.
• don't acquire files until classes are constructed, instead of when CST is imported (close Don't download resource files on import #284). Optionally suppress remote check for latest version.
• provide check_status() method to preview whether data dependencies can be acquired without errors (close Provide resource availability check methods #286). Includes current UTA_DB_URL in logging (close Make default UTA DB URL value programmatically accessible #287)
• raise FileNotFoundError if user-provided file paths don't exist. raise ValueError if they aren't files
• move around some modules/classes and rename the data module to resources
• Use importlib.resources to acquire the transcript mappings file rather than relying on relative file structure (hypothetically more reliable for reasons I don't totally understand)
• Update SeqRepo snapshot references to latest version (close Update SeqRepo collection to 2024-02-20 #270)

As far as I can tell, this shouldn't impact anything in production, because the environment variables are the first thing cst checks -- these changes basically only affect what happens if those fail or aren't provided.

[jsstevenson]

probably a git crime to do all of these at once but they felt relevant

[jsstevenson]

this method is long -- could be refactored into separate methods for each resource check

[korikuzma]

nice 🚀

[korikuzma]

Rather than linking to main page for rtd, it might be good to point to the api reference for the source

[jsstevenson]

I think this is sufficient to close #204 for now. If we want to circle back and make it actually lazy -- initialize a fetch on instance creation and then await it when actually calling methods -- we can reopen.

[korikuzma]

Sweet