Add an issue bucketing script for query execution tests

[kurapov-peter]

This is a basic version that collects test statuses in presence of failures. It captures the return codes of standalone tests and creates a yaml report. The report is coarse-grained since each test may include many queries, but it gives some coverage picture. The intended use is to simplify status checking for non-standard build/run options (like exec.heterogeneous.allow_cpu_retry = false). For example, I've collected the stats for ArrowBasedExecuteTest on Intel GPU (we have 211 tests in total there, each includes multiple queries):

status	number of tests
passes	116
assertion/driver failure	82
incorrect result	13

Bucketing granularity can be easily improved by capturing the stderr. The help should be self-explanatory. Let me know if anything is unclear.

[leshikus]

LGTM