Tune s3fs cache settings for optimal performance

[chuckwondo]

Based upon benchmarking tests, I've set the default s3fs cache settings to one of the better performing combinations. There were a number of combinations that performed very similarly, so the choice was a bit of a toss up, as many of them provide roughly a 10% speed improvement (on average) over the speed of downloading files. The speedup is not particularly notable, but that's not surprising, given that the granule files are not cloud-optimized.

However, this sets us up for possible further improvements, even with the non-cloud-optimized files, if we discover a more efficient means of writing our subsetting logic that requires fewer s3 read requests.

Fixes #77
Fixes #69

[wildintellect]

Seems good for me, but you've got an error in the test due to a missing import for typing related to fsspec. Probably want to resolve that before merging.

Run make typecheck
+ [[ true == \t\r\u\e ]]
+ mypy src tests typings
src/gedi_subset/subset.py:23: error: Skipping analyzing "fsspec": module is installed, but missing library stubs or py.typed marker  [import-untyped]
src/gedi_subset/subset.py:23: note: See https://mypy.readthedocs.io/en/stable/running_mypy.html#missing-imports
Found 1 error in 1 file (checked 16 source files)
make: *** [Makefile:56: typecheck] Error 1
Error: Process completed with exit code 2.

[chuckwondo]

@wildintellect, I fixed the build failure. Did you want to take another pass over the changes before I merge?