[SPARK-56120][PYTHON][TEST] Add ASV micro-benchmarks for SQL_WINDOW_AGG_ARROW_UDF

[Yicong-Huang]

What changes were proposed in this pull request?

Add ASV micro-benchmarks for SQL_WINDOW_AGG_ARROW_UDF.

Also adds optional runner_conf parameter to MockProtocolWriter.write_worker_input for passing configuration like window_bound_types.

Why are the changes needed?

Part of SPARK-55724. Establishes performance baselines before refactoring this eval type.

Does this PR introduce any user-facing change?

No.

How was this patch tested?

./python/asv run --python=same -b "WindowAggArrow" -a repeat=3 --show-stderr

··· ================ ============ ================
    --                            udf             
    ---------------- -----------------------------
        scenario       sum_udf     mean_multi_udf 
    ================ ============ ================
     few_groups_sm    11.3±0.3ms     9.85±0.1ms   
     few_groups_lg    45.7±0.1ms     44.4±0.1ms   
     many_groups_sm    317±4ms        273±5ms     
     many_groups_lg    177±2ms        158±1ms     
       wide_cols       85.1±1ms      83.2±0.7ms   
    ================ ============ ================

··· ================ ========= ================
    --                          udf            
    ---------------- --------------------------
        scenario      sum_udf   mean_multi_udf 
    ================ ========= ================
     few_groups_sm      468M         467M      
     few_groups_lg      492M         490M      
     many_groups_sm     474M         473M      
     many_groups_lg     505M         503M      
       wide_cols        478M         477M      
    ================ ========= ================

Was this patch authored or co-authored using generative AI tooling?

No.

[Yicong-Huang]

wait for fix in #55059

[zhengruifeng]

merged to master