fastberry 0.1.0

Performance helpers for strawberry-django: skip resolver overhead on sync Django.

fastberry

Performance helpers for strawberry-django GraphQL schemas running under synchronous Django.

strawberry-django wraps every field resolver in django_resolver, which calls in_async_context() on each resolution. Under sync Django that call raises and catches a RuntimeError internally (~0.15 ms per field). On large result sets this dominates response time — pure CPU overhead with zero extra database queries.

fastberry lets you opt specific "hot" types out of that machinery.

Install

pip install fastberry

Requires Python 3.10+, Django 4.2+, and strawberry-graphql.

Usage

Mark hot types with @fast_path (outermost, after @strawberry_django.type) and register FastPathExtension on the schema:

import strawberry
import strawberry_django
from fastberry import fast_path, FastPathExtension

from myapp.models import Stock


@fast_path
@strawberry_django.type(Stock, disable_optimization=True)
class StockType:
    id: int
    title: str


@strawberry.type
class Query:
    stocks: list[StockType] = strawberry_django.field()


schema = strawberry.Schema(
    query=Query,
    extensions=[FastPathExtension],
)

Only types decorated with @fast_path take the fast path; everything else resolves normally. To turn the optimization off, remove the extension (global) or the decorator (per type).

How it works

• Plain fields resolve via a direct getattr(root, attr). Related managers (*-to-many) are materialized with .all().
• Custom resolvers are called directly, bypassing the django_resolver wrapper.

The field→resolver map is built once, at class-definition time, and keyed by the GraphQL type name (info.parent_type.name) — not the Python class — because at resolve time root is the Django model instance, not the Strawberry type.

Benchmarks

Indicative numbers from a real schema (~3000 objects, ~27 fields each):

Scenario	Without fast-path	With fast-path	Speedup
stocks_by_house	4.4 s	1.8 s	~2.4×

Your mileage will vary with field count and result-set size.

Caveats

• Built for sync Django execution. Async schemas don't hit the overhead this targets.
• @fast_path does its own field resolution, so it skips any custom logic that django_resolver would otherwise apply. Use it on read-heavy, plain types; keep complex types on the default path.

Links

• Source: https://github.com/davidsarosi92/fastberry
• Issues: https://github.com/davidsarosi92/fastberry/issues

License

MIT