Skip to content

Core API Mode

GraphQL mode auto-generates a full query language. But many APIs are simpler — they just need well-shaped REST responses with related data loaded efficiently.

Core API gives you the same DataLoader batch loading and N+1 prevention, but as building blocks you control: define your DTOs, declare which fields need related data, and let the Resolver assemble the tree.

Step 1: Define DTOs with DefineSubset

DefineSubset picks fields from your SQLModel entities and generates Pydantic models. Relationship fields declared in the class body are auto-loaded by name:

from nexusx import DefineSubset

class UserDTO(DefineSubset):
    __subset__ = (User, ("id", "name"))

class TaskDTO(DefineSubset):
    __subset__ = (Task, ("id", "title", "owner_id"))
    owner: UserDTO | None = None   # Name matches Task.owner → auto-loaded

class SprintDTO(DefineSubset):
    __subset__ = (Sprint, ("id", "name"))
    tasks: list[TaskDTO] = []      # Name matches Sprint.tasks → auto-loaded

What DefineSubset does

  • Picks fields from the source entity — only the ones you list appear in the output
  • Hides FK fields automatically — owner_id won't appear in the JSON response, but remains accessible internally in resolve_*
  • Relationship fields go in the class body, not in __subset__ — their types must be DTO types

Warning

Relationship field types must be DTO types (DefineSubset or BaseModel subclasses). Using SQLModel entity types directly will raise a TypeError.

# Wrong — SQLModel entity type
class TaskDTO(DefineSubset):
    __subset__ = (Task, ("id", "title"))
    owner: User | None = None  # TypeError!

# Correct — DTO type
class TaskDTO(DefineSubset):
    __subset__ = (Task, ("id", "title"))
    owner: UserDTO | None = None  # OK

Step 2: Initialize ErManager and Resolver

At application startup — run once:

from nexusx import ErManager

er = ErManager(base=SQLModel, session_factory=async_session)
Resolver = er.create_resolver()
  • ErManager discovers all SQLModel entities and their relationships
  • create_resolver() returns a Resolver class bound to the entity graph

Warning

base and entities parameters are mutually exclusive — you can't pass both.

Step 3: Resolve in Your Route

from fastapi import FastAPI
from sqlmodel import select

app = FastAPI()

@app.get("/api/sprints")
async def get_sprints():
    async with async_session() as session:
        sprints = (await session.exec(select(Sprint))).all()
    dtos = [SprintDTO(id=s.id, name=s.name) for s in sprints]
    return await Resolver().resolve(dtos)

The Resolver traverses the DTO tree and fills in all relationship fields. You write the query, build the DTOs, and call resolve().

How Auto-Loading Works

When the Resolver encounters SprintDTO, it resolves the tree layer by layer:

SprintDTO(id=1, name="Sprint 1")
  → Auto-load: tasks → [TaskDTO(...), TaskDTO(...)]
    → Each TaskDTO: Auto-load: owner → UserDTO(...)
  → Result: complete nested response

Each relationship executes only one DataLoader query, regardless of how many Sprints or Tasks exist.

Four conditions for implicit auto-loading

The Resolver auto-loads a field when all four conditions are met:

  1. The field has no corresponding resolve_* method
  2. The field is an extra field (not in __subset__)
  3. The field name matches a registered ORM/custom relationship
  4. The field type is a BaseModel DTO compatible with the relationship target

When any condition fails, you can use resolve_* to load the field manually — see Core API Advanced.

Recap

  • DefineSubset generates DTOs from SQLModel entities — pick fields, hide FKs, declare relationship fields
  • Relationship fields are auto-loaded when the field name matches a registered relationship
  • ErManager discovers entities; Resolver traverses and assembles the DTO tree
  • Relationship field types must be DTO types, not SQLModel entities

Next Steps