fix(parquet/pqarrow): validate reader indexes

[fallintoplace]

Rationale for this change

Some Arrow Parquet reader paths index the field manifest or row-group metadata directly. Invalid input can panic even though the surrounding APIs return errors.

What changes are included in this PR?

Validate field indexes in GetFieldReader and row-group indexes in GetFieldReader and ReadColumn before using them.

Are these changes tested?

• go test ./parquet/pqarrow -run TestFileReaderIndexValidation

Are there any user-facing changes?

Invalid field and row-group indexes now return errors instead of panicking.

[zeroshade]

The bounds checks are correctly placed before the direct field and row-group index accesses, converting previously possible panics into typed errors while preserving valid reads.

The focused regression and complete parquet/pqarrow package suite pass locally on head d9988c9afc85. One small, non-blocking consistency observation is included inline. RC and Test CI are still running.

---

This review was drafted by an AI-assisted tool and confirmed by an Apache Arrow Go maintainer. The findings below are observations, not blockers; an Apache Arrow Go maintainer — a real person — will take the next look at the PR. If you think a finding is mis-applied, please reply on the PR and a maintainer will weigh in.

More on how Apache Arrow Go handles maintainer review: CONTRIBUTING.md.

[zeroshade]

The follow-up wraps the remaining column-reader bounds error with arrow.ErrIndex and adds matching regression coverage, resolving the prior consistency observation.

Re-review at 10ddd13aa2d7: the focused test passed 50 repetitions, the complete parquet/pqarrow package passed, and race plus 32-bit runs passed. All required CI checks are green.

---

This review was drafted by an AI-assisted tool and confirmed by an Apache Arrow Go maintainer.