Improving Data Bundle Management

[samuelfangjw]

I’ve been reflecting on how we currently handle data bundles, and it feels like the existing approach has some serious limitations:

• High maintenance for entity changes: Even small changes (e.g. [#13605] Normalize email addresses to lowercase #13630) require a surprisingly large number of edits, which makes the developer experience quite painful.
• Hard-to-edit test fixtures: The files are difficult to work with. Doesn't help that associations rely on opaque IDs, making it hard to reason about or update the data.
• Lack of validation: There’s no straightforward way to verify data bundles, so it’s easy to end up with invalid or inconsistent states.

Overall, it seems clear that the current data bundle system isn’t very scalable or maintainable. Long term, we need a solution that works cleanly across Java test files, scripts, and API-based seeding.

I’d welcome any thoughts or ideas on how we might tackle this.