[MINOR][ML] Simplify SummarizerBuffer branching

[zhengruifeng]

What changes were proposed in this pull request?

This PR simplifies SummarizerBuffer by precomputing requested metric and compute metric flags once, moving one-time array initialization into a helper, and splitting merge handling into empty-summary guard clauses plus focused non-empty merge/copy helpers.

The change preserves the existing aggregate state layout and public behavior.

Why are the changes needed?

SummarizerBuffer has several hot paths with repeated Seq.contains checks and nested null/empty branches. The refactor makes the control flow easier to read and removes repeated metric membership checks while keeping the same metric computations.

Does this PR introduce any user-facing change?

No.

How was this patch tested?

Ran:

git diff --check
awk 'length>100 && $0 !~ /^[[:space:]]*(import|package) / && $0 !~ /https?:\\/\\// {print FILENAME":"FNR": "length" chars"}' mllib/src/main/scala/org/apache/spark/ml/stat/Summarizer.scala\ngrep -rn -P "[^\\x00-\\x7F]" mllib/src/main/scala/org/apache/spark/ml/stat/Summarizer.scala\n```\n\nDid not run the full SBT test suite.\n\n### Was this patch authored or co-authored using generative AI tooling?\n\nGenerated-by: OpenAI Codex (GPT-5)\n