zarr-python [WIP] Boundary chunk optimization and typing

This PR addresses #757. is_total_slice takes a slice-like argument and a shape and returns True if the slice-like argument "fills" the shape, False otherwise. When performing writes, the output of is_total_slice determines whether a chunk is completely re-written (fast) or read -> modified -> written (slower). In master, is_total_slice is always called with the array's chunk shape as the second argument, and thus this function always returns False for chunks at the boundary of an array. This PR adds a new function zarr.util.trim_chunks in that essentially computes effective_chunk_shape = (chunk_shape * num_chunks) - array_shape and passes this value to is_total_slice, which allows writing to boundary chunks without a read operation.

Although all the core tests are passing, there's a deadlock in the test_parallel_append test in the test_sync suite. I have no idea what could be causing this.

Additionally, I kind of went crazy with adding type hints. It's possible that I will use this PR to add type hints to the entire library. It's possible that I should split the typing stuff into its own PR, but since none of it affects runtime I figured it couldn't hurt to combine it with some functional changes.

TODO:

[x] Add unit tests and/or doctests in docstrings
[ ] Add docstrings and API docs for any new/modified user-facing classes and functions
[ ] New/modified features documented in docs/tutorial.rst
[ ] Changes documented in docs/release.rst
[ ] GitHub Actions have all passed
[ ] Test coverage is 100% (Codecov passes)