GH-3820: Exercise explicit binary statistics truncation

[dhruv-15-03]

Rationale for this change

The binary-statistics truncation test calls a static default overload through a configured converter instance. The constructor's truncation length is ignored, and the existing ordering-only assertions let untruncated bounds pass.

What changes are included in this PR?

• Call the explicit toParquetStatistics(stats, truncateLen) overload.
• Use deterministic byte arrays and check exact truncated bytes and lengths, including the conservative increment of the maximum.
• Check that combined bounds of 4095 bytes are accepted and 4096 bytes are omitted, with both default and finite truncation.
• Preserve invalid-length constructor checks and current null-count omission behavior.

Only TestParquetMetadataConverter.java changes. The personal-fork validation workflow is not included.

Are these changes tested?

Yes. On the contribution commit a9940be5cfee6962dc0f3e064b4c619a07b2d3b7 (Temurin 17), the full TestParquetMetadataConverter class ran 78 tests with 0 failures, 0 errors and 0 skipped, including testBinaryStatsWithTruncation and testBinaryStatsWithTruncationAtSizeLimit. mvn -pl parquet-hadoop spotless:check reported 262 files clean. Run: https://github.com/dhruv-15-03/parquet-java/actions/runs/36249258685

Are there any user-facing changes?

None. Production code and output semantics are unchanged. #3819 addresses null-count serialization; this PR repairs a separate test-coverage gap and does not fix or close #3574.

Closes #3820

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.