[Fix] Override data parallel size for accelerator configs

[YouNeedCryDear]

What this PR does

Adds accelerator-specific runtime argument overrides for data parallel size, covering --dp-size, --dp, --data-parallel-size, and -dp. It also extends tensor and pipeline parallel overrides to handle -tp and -pp aliases.

Why we need it

tensorParallelismOverride.dataParallelSize is part of the ServingRuntime accelerator config API, but the runtime argument override logic only applied tensor and pipeline parallel sizes. As a result, accelerator configs could set dataParallelSize without changing the generated runtime command. This makes DP override behavior match TP and PP override behavior.

Fixes # (not filed)

How to test

env GOCACHE=/private/tmp/ome-go-build-cache go test ./pkg/controller/v1beta1/inferenceservice/components -count=1

Checklist

• Tests added/updated (if applicable)
• Docs updated (if applicable) - N/A
• make test passes locally

[YouNeedCryDear]

@slin1237 Let me know when this can be merged.

[heymrbox]

Reviewed the TP/PP/DP override patch. I did not find a blocking issue.

Verification run on the PR head:

go test ./pkg/controller/v1beta1/inferenceservice/components -run TestMergeRuntimeArgumentsOverride -count=1

The DP override follows the existing overrideParam behavior used for TP/PP, and the added tests cover long flags, SGLang size flags, command-based shorthand flags, and single-dash aliases. PR checks are green as of this review.