[Tests] Add long context batch tests

[rafvasq]

Description

• Cleans up create_text_prompt
• Adds create_seq_prompt
• Adds six test cases:
  • 32 prompts with 512 tokens each
  • 16 prompts with 1.5k tokens each
  • 8 prompts with 3k tokens each
  • 4 prompts with 5k tokens each
  • 2 prompts with 9k tokens each
  • 1 prompt with 17k tokens

[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 🚀

[rafvasq]

Test passes on CPU:

tests/e2e/test_spyre_cb.py::test_long_context_batches[TP(1)-eager-ibm-ai-platform/micro-g3.3-8b-instruct-1b] PASSED [ 50%]
tests/e2e/test_spyre_cb.py::test_long_context_batches[TP(1)-eager-ibm-ai-platform/micro-g3.3-8b-instruct-1b-FP8] SKIPPED [100%]

[yannicks1]

LGTM. Final remark: will this (in its current form) run with GHA? if so, I believe we should remove it. On cpu there is nothing new here, and it is expected to work. IMO this is a test for Spyre only and should not be run on cpu for every PR...

[yannicks1]

lgtm