Skip to content

Integrate With Django Computing API

Use this page after your computation already works locally.

The goal here is modest:

  • add a route
  • keep the handler thin
  • pass work into a callable computation boundary

The Files That Matter

Thin Computing API handler
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
@api_view(["POST"])
@schema(None)
def generate_example(request):
    state = request.data.get("state").lower()          # (1)!
    district = request.data.get("district").lower()
    block = request.data.get("block").lower()
    gee_account_id = request.data.get("gee_account_id")
    run_example_task.apply_async(                      # (2)!
        args=[state, district, block, gee_account_id],
        queue="nrm",
    )
    return Response({"Success": "Successfully initiated"})  # (3)!
  1. Normalize and validate only what the entry layer truly owns.
  2. Push real work into a task or callable boundary.
  3. Return a small response quickly.

This is why Local Pipeline First exists: the handler should not be where the science becomes debuggable for the first time.

Route Wiring Checklist

  1. add the route in computing/urls.py
  2. add the handler in computing/api.py
  3. call a stable computation boundary
  4. keep request parsing shallow
  5. make the next page decide whether Celery, auth, or retries need to become stricter

What Not To Do Here

  • do not bury analytical logic inside the request handler
  • do not make this the first place the computation is ever run
  • do not mix publication, validation, and route parsing into one large function

Bridge To The Computing Surface

Once a computation has a route, it becomes part of the broader Computing API surface.

That means it should appear clearly in:

Only after this step should you continue into Add Celery / Auth / Task Integration.