Design Changes for Dir Struct, Tagged Iterations, Metrics

[stanbrub]

• Added user-supplied metrics for queries and made metrics simpler
  • Added data.file.size user metric to Parquet tests
  • Reworked queries to allow joining metrics that are not present for every benchmark
• Got rid of collection of before and after bean metrics (They didn't really provide insight)
• Added iterations that run only on benchmark tests that have tag annotations (ex @Tag("Iterate")
• Unified the directory structure for all categories in GCloud (ex. category/owner/set-label/run-label/)
  • Updated all workflows and query snippets to this effect
  • Wrote converters for each category; nightly, release, compare
• Added default set labels for each category
  • adhoc blank: set-23YGkqB
  • nightly: 2024-07-11 (ex.
  • release and compare: 00.035.05
• Default input properties are now specified at the top level workflow level regardless of category
• SVG templates now generated to the 'set' directory and latest to the owner directory
  • Updated generation and README.md links
• Reworked generator file names and added file-name hashes for looking up reusable data
• Updated benchmark_tables.py and adhoc_tables.py in GCloud for use with the other snippets