fix(python): support Polars 2

[robert3005]

Summary

Polars 2.0.0 is out, and the polars>=1.31.0 extra has no upper bound, so users already get it. With Polars 2, every VortexFile.to_polars() scan fails at collect time:

TypeError: DataFrame._from_arrow() got an unexpected keyword argument 'rechunk'

to_polars() called the private pl.DataFrame._from_arrow(batch, rechunk=False), and Polars 2 removed that keyword.

Changes

• vortex-python/python/vortex/file.py: build each batch, and the trailing empty batch, with the public pl.DataFrame(record_batch) constructor. It works on both Polars 1.x and 2.x and is typed as DataFrame. pl.from_arrow would return DataFrame | Series.
• Bump the polars dev dependency to >=2.0.0 and relock (polars / polars-runtime-32 1.34.0 → 2.0.0), so CI tests against Polars 2.
• The user-facing polars>=1.31.0 extra is unchanged.

The serialized expression format that vortex.polars_ parses for predicate pushdown is the same in 2.0. All expression-translation and pushdown tests pass unchanged.

Checks run

• uv run --all-packages pytest vortex-python/test on Polars 2.0.0: 409 passed, 2 skipped, 1 xfailed. Before the fix, 5 to_polars tests failed with the error above.
• uvx ruff check, uvx ruff format --check, uvx ty check vortex-python vortex-python-cuda vortex-ffi/cmake/tests scripts/tests: clean.
• test_polars_.py on Polars 1.31.0 (the extra's floor): same 4 failures as on develop, so nothing new. They are pre-existing: 1.31 serializes scalars as {"dtype": ..., "value": ...}, which polars_to_vortex doesn't parse. This means the >=1.31.0 floor is already too low for predicate pushdown. I've left that for a follow-up rather than widening this PR.

🤖 Generated with Claude Code

https://claude.ai/code/session_01E92wUoS3LaqWVJmmS7vKVq

---

Generated by Claude Code

[codspeed]

Merging this PR will degrade performance by 11.75%

⚠️ Unknown Walltime execution environment detected

Using the Walltime instrument on standard Hosted Runners will lead to inconsistent data.

For the most accurate results, we recommend using CodSpeed Macro Runners: bare-metal machines fine-tuned for performance measurement consistency.

⚠️ 2 benchmarks measured no execution time

Nothing ran under measurement, usually because the compiler removed the code under test. These results are not comparable, so they count as unchanged.

Preventing compiler optimizations

⚠️ Different runtime environments detected

Some benchmarks with significant performance changes were compared across different runtime environments,
which may affect the accuracy of the results.

Open the report in CodSpeed to investigate

❌ 2 regressed benchmarks
✅ 2173 untouched benchmarks
⏩ 534 skipped benchmarks¹

Warning

Please fix the performance issues or acknowledge them on CodSpeed.

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
❌	WallTime	scalar_subtract_neon	10.7 µs	12.2 µs	-12.17%
❌	WallTime	bitpack_blocked_compress_avx2	6.7 µs	7.6 µs	-11.34%
⚠️	Simulation	bench_compare_primitive[(10000, 2)]	64.2 µs	< 1 ns	N/A
⚠️	Simulation	density_sweep_dense_runs[0.9]	< 1 ns	< 1 ns	N/A

Tip

Investigate this regression by commenting @codspeedbot fix this regression on this PR, or directly use the CodSpeed MCP with your agent.

---

_{Comparing rk/polars-2 (94fb980) with develop (881b867)}

Footnotes

1. 534 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩