feat: include benchmark for compatibility parser

[sunng87]

We are increasing complexity for our compatibility parser by introducing partial replacement in #259 . This benchmark is to measure the performance loose.

According to my local test:

master branch

parse_simple_queries    time:   [23.109 µs 23.121 µs 23.139 µs]
                        change: [−10.475% −10.262% −10.064%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 3 outliers among 100 measurements (3.00%)
  3 (3.00%) high severe

parse_complex_queries   time:   [67.685 µs 67.811 µs 67.942 µs]
                        change: [−25.996% −25.830% −25.661%] (p = 0.00 < 0.05)
                        Performance has improved.

parser_creation         time:   [199.03 µs 199.46 µs 199.88 µs]
                        change: [+11.408% +11.651% +11.892%] (p = 0.00 < 0.05)
                        Performance has regressed.

#259 change

parse_simple_queries    time:   [25.735 µs 25.798 µs 25.871 µs]
                        change: [−2.9878% −2.3813% −1.9585%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 2 outliers among 100 measurements (2.00%)
  1 (1.00%) high mild
  1 (1.00%) high severe

parse_complex_queries   time:   [91.307 µs 91.483 µs 91.680 µs]
                        change: [−14.165% −13.959% −13.767%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 2 outliers among 100 measurements (2.00%)
  2 (2.00%) high mild

parser_creation         time:   [178.38 µs 178.74 µs 179.14 µs]
                        change: [−3.0597% −2.8627% −2.6649%] (p = 0.00 < 0.05)
                        Performance has improved.

For most simple queries, there is a 10% lose which I think is acceptable.