As tech lead, I want a database seed implemented for testing

[andrew-jameson]

Description: Given our occasional setbacks resulting from database inconsistencies that result in deployment failures that back up our release schedules, we should mitigate this with more thorough testing.

Acceptance Criteria:

• [ ] We have a maintainable and extensible toolset
• [ ] Seeding a database is performant ( sub hour )
• [ ] Toolset is version-aware and sensitive
• [ ] Testing Checklist has been run and all tests pass
• [ ] README is updated, if necessary

Tasks:

• [ ] Create a tool/mechanism for generating "random" and internally consistent data
• [ ] Ensure tool is version/schema aware (we should be able to use django's migration table)
• [ ] Tool will "fuzz" or generate out of range values to intentionally create issues [ stretch ]
• [ ] Create django command such that tool can be pointed at deployed environments [ stretch ]
• [ ] Run Testing Checklist and confirm all tests pass

Notes:

Supporting Documentation:

Open Questions:

• Can we work with Alex to scrub PII from the prod postgres dataset?
• How can this be integrated into unit/integration tests? Pipelines?
• What drift exists between postgres migrations and ES documents? Is that prone to human error?
• Can we leverage co-pilot or other AI for generation?

[jtimpe]

A basic implementation of a random record generator is included here - https://github.com/raft-tech/TANF-app/pull/2881/files#diff-ce520f958d18ac5092cb7ab4baa305d5cf3d72a5ee1e00f93020334fb83fa92b