Random questions

Hey,

I have a few random questions while getting to know this setup.

1. The [three queries](https://github.com/edgedb/imdbench#queries) are the things actually being tested?
2. The `index.js` (or equivalent) inside the projects implements much more. The other queries are not benchmarked or not used?
3. Is there some test setup one can use to confirm that the returned data for a folder adheres to the expected data format and content during development? (Does the benchmark do that?)
4. Is there some additional information somewhere about the problems described under "Transactional queries 🏦" and how they would occur for the datamodel under test here?
5. Is there a commited version of the SQL DDL all the non EdgeDB tests run against?
6. Is connection creation part of the benchmark results or are queries done on brand new Clients?
7. Prisma specific: `TunedApp vs. `App`?
8. Is there a CI Setup to run the benchmarks or all manual?

Thanks!

Read the README sections linked by “three queries” and “Transactional queries,” then inspect the projects’ index.js files and any committed SQL DDL or CI configuration mentioned by the repository. Done means turning the eight questions into authoritative, scoped answers about benchmark inputs, validation, transactions, schema, connection setup, Prisma variants, and CI.

Written by the indexing model from the issue text.