fix(python): fix broken pylance benchmark scripts

[mocobeta]

The benchmarks in python/benchmarks seem broken or obsolete. This PR fixes various minor issues in the benchmark scripts.

• Add necessary dependencies to benchmarks group.
• Fix a typo in a test method signature in test_index.py.
• Add mandatory num_sub_vectors param when creating IVF_PQ index in test_index.py.
• Fix an assertion in test_file.py. The right side of the assertion should be the number of sampled rows, not NUM_ROWS.

Run benchmarks on main branch

❯ python -V
Python 3.13.13

❯ uv run pytest python/benchmarks/ -m "not slow" --ignore python/benchmarks/test_random_access.py

...
==================================================================================================== short test summary info =====================================================================================================
FAILED python/benchmarks/test_file.py::test_sample_integer[2_0] - assert 10 == 10000000
FAILED python/benchmarks/test_file.py::test_sample_integer[2_1] - assert 10 == 10000000
FAILED python/benchmarks/test_file.py::test_sample_integer[2_2] - assert 10 == 10000000
FAILED python/benchmarks/test_file.py::test_sample_nested_integer - assert 10 == 10000000
FAILED python/benchmarks/test_index.py::test_create_ivf_pq_torch_cpu - ModuleNotFoundError: torch.device requires 'torch' module to be installed
FAILED python/benchmarks/test_index.py::test_create_ivf_pq_torch_cpu_one_pass - ModuleNotFoundError: torch.device requires 'torch' module to be installed
FAILED python/benchmarks/test_index.py::test_optimize_index[12000-5-256] - ValueError: num_partitions and num_sub_vectors are required for IVF_PQ
FAILED python/benchmarks/test_index.py::test_optimize_index[12000-5-512] - ValueError: num_partitions and num_sub_vectors are required for IVF_PQ
ERROR python/benchmarks/test_index.py::test_transform_vectors_no_precomputed_parts
ERROR python/benchmarks/test_search.py::test_geo_rtree[no_index] - ModuleNotFoundError: No module named 'geoarrow'
ERROR python/benchmarks/test_search.py::test_geo_rtree[with_index] - ModuleNotFoundError: No module named 'geoarrow'
==================================================================== 8 failed, 280 passed, 2 skipped, 2 deselected, 6 warnings, 3 errors in 431.23s (0:07:11) ====================================================================

Run benchmarks on this branch

❯ python -V
Python 3.13.13

❯ uv run pytest python/benchmarks/ -m "not slow" --ignore python/benchmarks/test_random_access.py

...
Legend:
  Outliers: 1 Standard Deviation from Mean; 1.5 IQR (InterQuartile Range) from 1st Quartile and 3rd Quartile.
  OPS: Operations Per Second, computed as 1 / Mean
======================================================================================== 291 passed, 2 skipped, 2 deselected, 5 warnings in 731.11s (0:12:11) =========================================================================================

Summary by CodeRabbit

Summary by CodeRabbit

• Chores

  • Added benchmark-related dependencies for PyTorch and GeoArrow Rust support.

• Bug Fixes

  • Corrected sampling benchmark assertions to verify the actual sampled row count.
  • Updated index optimization benchmarks to pass the intended vector configuration.
  • Fixed benchmark test fixture wiring for vector transformation benchmark setup.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: ASSERTIVE

Plan: Pro Plus

Run ID: df112dee-f06a-429f-8b44-a2ed1d009885

📥 Commits

Reviewing files that changed from the base of the PR and between 65582a0 and dbdc9a6.

📒 Files selected for processing (1)

• python/python/benchmarks/test_index.py

---

📝 Walkthrough

Walkthrough

Benchmark dependencies are expanded, sampling benchmarks assert sampled row counts, IVF-PQ optimization specifies sub-vector configuration, and vector transformation uses the dataset fixture.

Changes

Benchmark updates

Layer / File(s)	Summary
Sampling result assertions python/python/benchmarks/test_file.py	Sampling tests now compare result row counts with the number of sampled indices.
Index benchmark setup and environment python/python/benchmarks/test_index.py, python/pyproject.toml	The IVF-PQ benchmark adds an explicit sub-vector parameter, the transformation benchmark uses test_dataset, and benchmark dependencies include PyTorch and GeoArrow Rust packages.

Estimated code review effort: 2 (Simple) | ~10 minutes

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 0.00% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title clearly matches the main change: fixing broken benchmark scripts in the Python package.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

_{Comment @coderabbitai help to get the list of available commands.}