test(spark): characterize partition transform resolution

[jackylee-ch]

Rationale for this change

resolveTransforms turns Spark partition transforms into the directory keys and column indices a writer uses, and had no tests. It is a pure static function, so the behaviour is cheap to pin.

What changes are included in this PR?

Test only. Covers the directory keys derived for identity, years/months/days/hours and bucket; the type guards (temporal transforms reject a non-temporal column, hours also rejects a date); the bucket count and multi-column index resolution; and the rejections for an unknown transform name, a missing numBuckets argument, a zero-reference transform and a column outside the schema.

Also covers the Java serialization round trip of ResolvedTransform — the reason resolution happens eagerly on the driver, since Spark's Transform objects cannot cross to executors.

Verified non-vacuous: disabling the temporal type guard fails the guard cases.

• ./gradlew :vortex-spark_2.13:test --tests '…PartitionedVortexDataWriterTest' — 11 pass
• same on :vortex-spark_2.12:test — 11 pass
• spotlessCheck (both variants) and javadoc — clean

What APIs are changed? Are there any user-facing changes?

None.

AI assistance

Prepared with agentic AI assistance. I read the resolution code and its callers, confirmed each assertion against the current implementation, and verified the guard tests fail when the guard is removed.

[codspeed-hq]

Merging this PR will degrade performance by 10.85%

❌ 2 regressed benchmarks
✅ 1942 untouched benchmarks
⏩ 85 skipped benchmarks¹

Warning

Please fix the performance issues or acknowledge them on CodSpeed.

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
❌	Simulation	decompress[u64, (1000, 16)]	64 µs	72.1 µs	-11.19%
❌	Simulation	take[small_m/shuffled/primitive/nonnull/chunks=16384/indices=16]	1.2 ms	1.3 ms	-10.51%

Tip

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

---

_{Comparing jackylee-ch:spark-partition-transform-tests (a70e7a0) with develop (17b190b)}

Footnotes

1. 85 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. ↩