ci: actually fix benchmarking

[cpcloud]

• ci: ensure that benchmarks do not reference data that subject to disappear
• ci: actually fix benchmarking

[github-actions]

ACTION NEEDED

Ibis follows the Conventional Commits specification for release automation.

The PR title and description are used as the merge commit message.

Please update your PR title and description to match the specification.

[cpcloud]

Passing locally, merging:

…/ibis on  fix-bench is 📦 v8.0.0 via 🐍 v3.12.2 via ❄️   impure (ibis-3.12.2-env)
❯ pytest --benchmark-disable ibis/tests/benchmarks/test_benchmarks.py
======================================================================================= test session starts ========================================================================================
platform linux -- Python 3.12.2, pytest-8.1.1, pluggy-1.4.0
benchmark: 4.0.0 (defaults: timer=time.perf_counter disable_gc=False min_rounds=5 min_time=0.000005 max_time=1.0 calibration_precision=10 warmup=False warmup_iterations=100000)
Using --randomly-seed=1277268930
rootdir: /home/cloud/src/ibis
configfile: pyproject.toml
plugins: xdist-3.5.0, pytest_httpserver-1.0.10, cov-4.1.0, benchmark-4.0.0, randomly-3.15.0, mock-3.12.0, repeat-0.9.3, snapshot-0.9.0, clarity-1.0.1, hypothesis-6.99.2
collected 121 items

ibis/tests/benchmarks/test_benchmarks.py ............x..x......x...........................x..............................x......x......x.............x...........                           [100%]

============================================================================ 113 passed, 8 xfailed in 69.62s (0:01:09) =============================================================================