Add transaction information to summary

[systay]

Overview

Enhances the Markdown report generated by vt summarize by incorporating transaction pattern analysis from vt transactions JSON output.

This enhancement analyzes transaction patterns with a focus on identifying meaningful relationships for sharding key decisions. The analysis specifically targets patterns where queries share predicate values, filtering out patterns that don't contribute to sharding strategy insights.

Here follows an example of this section in the report:

Transaction Patterns

Pattern 1 (Observed 46 times)

Tables Involved: craftable_timesheets, pos_dailies, pos_tickets

Query Patterns

1. DELETE on craftable_timesheets
   Predicates: craftable_timesheets.store_id = 0 AND craftable_timesheets.rpt_date = 1 AND craftable_timesheets.intraday = 2

2. DELETE on pos_tickets
   Predicates: pos_tickets.store_id = 0 AND pos_tickets.rpt_date = 1 AND pos_tickets.intraday = 2

3. DELETE on pos_dailies
   Predicates: pos_dailies.store_id = 0 AND pos_dailies.rpt_date = 1 AND pos_dailies.intraday = 2

Shared Predicate Values

• Value 0 applied to:
  • craftable_timesheets.store_id
  • pos_tickets.store_id
  • pos_dailies.store_id
• Value 1 applied to:
  • craftable_timesheets.rpt_date
  • pos_tickets.rpt_date
  • pos_dailies.rpt_date
• Value 2 applied to:
  • craftable_timesheets.intraday
  • pos_tickets.intraday
  • pos_dailies.intraday

---

[rohit-nayak-ps]

The refactor looks so good!

[rohit-nayak-ps]

This might be inefficient for large number of queries. The usual pattern of using a map to check existence might be faster.

[rohit-nayak-ps]

We should think about using a template library like text/html to generate the markdown. That approach has worked well for me in the past.