refactor: Consolidate async UDF handling in physical planner

[GaneshPatil7517]

Which issue does this PR close?

Closes #18150
Fixes #18149

Rationale for this change

The previous implementation called try_plan_async_exprs() at three separate locations (projection, filter, aggregation), each with its own match arms handling PlanAsyncExpr::Sync vs PlanAsyncExpr::Async. This caused:

1. Duplication - The same pattern repeated three times
2. Cognitive overhead - Understanding async UDF handling required reading all three locations
3. Bug Async UDF rewrite input to Aggregate breaks UDAFs #18149 - Aggregation passed the wrong schema to AggregateExec, causing expression/schema mismatch

What changes are included in this PR?

Introduces two helper methods that consolidate async UDF handling:

• maybe_wrap_async_exec() — for plain expressions (filter, aggregation)
• maybe_wrap_async_exec_with_names() — for projection expressions with aliases

Both methods:

1. Scan expressions for async UDFs
2. If found, wrap the input with AsyncFuncExec
3. Return the (possibly wrapped) input and rewritten expressions

This makes each call site a simple function call instead of a multi-arm match.

Removed code

• try_plan_async_exprs() method
• PlannedExprResult enum
• PlanAsyncExpr enum

How this fixes #18149

The bug occurred because aggregation expressions were rewritten to reference __async_fn_0@N, but AggregateExec was created with the original schema (which didn't include async columns).

The fix uses input_exec.schema() after wrapping with AsyncFuncExec, ensuring the aggregate sees a schema that includes the async columns.

Are these changes tested?

Existing async_udf.slt tests cover async UDF usage in projections, filters, and aggregations. Physical plan outputs remain unchanged.

Are there any user-facing changes?

No. This is an internal refactor with no public API changes.

[Jefffrey]

Thanks for volunteering to pick up these issues. I'm not quite sure of the refactorings presented here as they don't seem to target the root cause of either issue 🤔