FastAPI admin backend

The problem

You need an internal admin API: list/search endpoints with pagination, create/update/delete, per-role permissions, a few computed display columns, and a soft-delete archive for auditing. A full ORM means model classes, migrations and a lot of boilerplate; a bare driver means hand-writing all of the access control and pagination yourself.

Why py-store

Schema-driven CRUD: register a JSON schema, then insert / update / remove / query with no model classes.
Read-time computed columns (fn / asyncFn / relation agg) give you display values without denormalizing.
Built-in RBAC: schema-level and field-level read / write whitelists, plus the creator pseudo-role for ownership (doc.createdBy == ctx.userId).
Soft delete is automatic — every schema gets a <Model>Deleted archive collection/table, and remove() archives before deleting.
query_with_count returns items plus total and pagination metadata in one call, with pageSize capped at 5000.

Walkthrough

from fastapi import Depends, FastAPI, Request
from py_store import init, store

app = FastAPI()

@app.on_event("startup")
async def startup():
    await init({"default": mongo_db})
    store.register({
        "name": "Article",
        "collection": "articles",
        "idPrefix": "AR",
        "timestamps": True,
        "fields": {
            "_id": "string",
            "title": {"type": "string", "default": ""},
            "body": {"type": "string", "default": ""},
            "status": {"type": "string", "default": "draft"},
            "authorId": {"type": "string", "default": ""},
        },
        "computes": {
            "titleUpper": {"type": "string", "depends": ["title"],
                           "fn": lambda d: d["title"].upper()},
        },
        # schema-level role whitelists; `creator` is resolved per document
        "read": ["editor", "viewer", "creator"],
        "write": ["editor", "creator"],
        "indexes": [{"keys": {"status": 1, "createdAt": -1}}],
    })

# One dependency sets the permission context for the whole request flow.
async def current_user(request: Request) -> dict:
    user = {"userId": request.headers["x-user-id"],
            "roles": request.headers.getlist("x-role")}
    store.set_context(user)            # roles drive schema/field whitelist checks
    return user

@app.get("/articles")
async def list_articles(page: int = 0, page_size: int = 50,
                        user: dict = Depends(current_user)):
    gql = "Article($condition:@c0,$sort:@s1) { _id, title, status, titleUpper }"
    params = {"c0": {}, "s1": {"createdAt": -1}, "page": page, "pageSize": page_size}
    return await store.query_with_count(gql, params)
    # -> {"items", "total", "hasMore", "page", "pageSize"}   (pageSize capped at 5000)

@app.post("/articles")
async def create_article(payload: dict, user: dict = Depends(current_user)):
    return await store.insert("Article", {**payload, "authorId": user["userId"]})

@app.delete("/articles/{article_id}")
async def delete_article(article_id: str, user: dict = Depends(current_user)):
    # archives to <collection>_deleted first, then deletes the original document
    return await store.remove("Article", {"_id": article_id})

Soft delete is transparent: the archive collection <collection>_deleted is auto-registered per schema, and re-creating the same _id does not collide because the archive write is upsert-by-_id.

Pitfalls

Set the context on the same async path as the call. The permission context lives in a ContextVar, so set it in an async dependency (or a helper wrapping the store call) — a sync dependency running in a thread pool will not propagate it to the endpoint.
Empty-condition writes are rejected. update_many / remove with {}, None or {"$and": []} is refused outright; it never falls through to a full-table write. Require an explicit filter from the request.
Without a context, checks are disabled. By default, a missing context means fail-open (no permission check). Enable store.set_require_context(True) at startup for fail-secure behaviour, and wrap internal/cron work in store.run_as_internal(...).
createdAt / updatedAt are framework-maintained. Do not set them from the payload; the unit follows the schema’s timestamps setting ("s" for seconds).
pageSize is capped at 5000. Larger values are clamped, so build your UI pagination around that ceiling.

FastAPI admin backend

The problem

Why py-store

Walkthrough

Pitfalls

See also