SqliteReportBuilder: pull rng from test_utils instead of dependency injection

[matt-codecov]

i don't really like this approach so this is more of an RFC. now that it's written i'm kind of thinking about giving up on controlling the RNG to begin with though haha

currently parse_pyreport() takes an output file. i want to expose to Python a version that doesn't which will create a new db in std::env::temp_dir(). since parse_pyreport() and SqliteReportBuilder already have multiple entrypoints to allow controlling the RNG, this is making things messier

the other approach is something like

PyreportParser::new()
    .with_report_json(f1)
    .with_chunks(f2)
    .with_rng(StdRng::from_entropy())
    .with_outfile(out)
    .parse()

but then to use the same approach with SqliteReportBuilder you need a SqliteReportBuilderBuilder which is silly

---

Stack created with Sapling. Best reviewed with ReviewStack.

• -> SqliteReportBuilder: pull rng from test_utils instead of dependency injection #33
• create test_utils crate to house all fixtures and reusable helpers #32

[codecov]

❌ We are unable to process any of the uploaded JUnit XML files. Please ensure your files are in the right format.

[matt-codecov]

yeah this doesn't feel right, i am going to try again