129: Refactoring models

[JaeYeonLee0621]

1. Split the aqueduct/management/models.py into a management/models/ package

• models/__ init __.py — re-exports every model/helper
• models/mixins.py — LimitSet, LimitMixin, ModelExclusionMixin, MCPServerExclusionMixin
• models/accounts.py — Org, Team, UserGroup, UserProfile, TeamMembership, ServiceAccount, Token
• models/snippets.py — SnippetType, Snippet
• models/usage.py — Usage, Request
• models/files.py — FileObject
• models/batches.py — default_request_counts, BatchStatus, Batch
• models/vector_stores.py — VectorStore/VectorStoreFile/VectorStoreFileBatch + status enums

2. Deleted the old models.py

+) No DB changes → no migration needed.
+) Historical migrations still resolve via the package init.py

[meffmadd]

When I planned this I would have just created a single file for each models.Model. I think this is still the simplest solution. Some classes are long anyways and the naming is also straightforward then.

[meffmadd]

I generally dislike __all__ because it is always a hassle to update for little gain but if you have a case for it it is fine.

[JaeYeonLee0621]

I kept __all__ here because

1. Dropping these re-exports would force full paths everywhere and break historical migrations (e.g. management/migrations/0005_file_purpose_update.py).
2. The "as" re-export (from X import Y as Y) works , but __all__ is the pattern already used (e.g. gateway/views/__init__.py).

[meffmadd]

Ok got it! Keep it as is then.

[meffmadd]

Token should have its own file I think... this file is already very long again.

[JaeYeonLee0621]

I created a separate tokens.py file for this model.

[meffmadd]

To make things as simple as possible can you just create a new file for each model with the file having the same name as the class? I think that does make sense here and we do not have to worry about naming.

[JaeYeonLee0621]

I've applied your comments and moved each model into its own file, named after the class. :)

[meffmadd]

LGTM!