[CB] Refactoring/Cleaning up prepare_prompt/decode

[yannicks1]

[CB] Refactoring/Cleaning up prepare_prompt/decode

As prefills for batch size > 1 have been de-prioritized, this PR cleans up the code considerable and makes it more readable. Code readability becomes increasingly important with the upcoming optimization for homogeneous tkv (e.g. #262 )

[github-actions]

👋 Hi! Thank you for contributing to vLLM support on Spyre.
Just a reminder: Make sure that your code passes all the linting checks, otherwise your PR won't be able to be merged. To do so, first install the linting requirements, then run format.sh and commit the changes. This can be done with uv directly:

uv sync --frozen --group lint --active --inexact

Or this can be done with pip:

uv pip compile --group lint > requirements-lint.txt
pip install -r requirements-lint.txt
bash format.sh

Now you are good to go 🚀

[sducouedic]

maybe? I was confused by "maximal" and "current"

Suggested change

	# Reserve the maximal number of blocks used to serve current sequence
	# Reserve the number of blocks required to serve this new sequence

[sducouedic]

LGTM

Was so nice to read!