GH-49385: [C++][Parquet] Clarify empty schema contract on stream_reader

[emkornfield]

Rationale for this change

StreamReader inherently does not support empty schemas. Guard this case with an exception.

What changes are included in this PR?

Added validation around the parquet reader passed in.

Are these changes tested?

Yes added unit tests.

Are there any user-facing changes?

A change that might be debatable is the constructor for this class can now throw, but it was never marked noexcept.

This PR contains a "Critical Fix".

• GitHub Issue: Parquet StreamReader should clarify its contract for parquet files without a schema. #49385

[emkornfield]

CI failures seem related to build infra?

[emkornfield]

@pitrou any objections to me merging?

[pitrou]

Perhaps rebase to try and fix CI (not sure that'll be sufficient though)? Apart from that, feel free to merge, and thanks for tackling this!

[emkornfield]

Perhaps rebase to try and fix CI (not sure that'll be sufficient though)? Apart from that, feel free to merge, and thanks for tackling this!

Unfortunately still seeing unrelated errors (i.e. for flight)

[conbench-apache-arrow]

After merging your PR, Conbench analyzed the 3 benchmarking runs that have been run so far on merge-commit fcf9dd6.

There was 1 benchmark result indicating a performance regression:

• Commit Run on amd64-m5-4xlarge-linux at 2026-03-04 20:47:23Z
  • file-write (Python) with compression=lz4, dataset=fanniemae_2016Q4, file_type=feather, input_type=table

The full Conbench report has more details. It also includes information about 41 possible false positives for unstable benchmarks that are known to sometimes produce them.