fix: use unbounded channel for worker write queue

[ltungv]

prevent deadlock by making one of the channels unbounded so that send never blocks on the apply thread.

worker's write channel size grows with the process rate of the apply thread. flushes are also assumed to be relatively cheap. keeping read_rx bounded also prevents the system from being flooded by external requests.

closes #153

local checks

PASS  cargo fmt --all --check  (1s)
PASS  cargo clippy --workspace --all-targets --features testing -- -D warnings  (0s)
PASS  cargo test --workspace --features testing  (56s)
PASS  ruff check python/  (0s)
PASS  mypy --strict beava/
PASS  pytest python (v0 acceptance suite)  (14s)

[Khanathan]

wow, massive PR